Challenges in Applying Colonial Boundary Treaties to the Resolution of the Djibouti–Eritrea Border Dispute

Abstract

Djibouti and Eritrea have been in conflict since June 2008 when their troops fought along the Djibouti–Eritrea border. The conflict revolves around the location of the border and sovereignty over the strategically located Doumeira Islands and adjacent reefs. In 2010 Qatar brokered a mediation agreement and began to implement it, but withdrew in 2017 without notifying Eritrea and without providing reasons to either country. The dispute raises a number of international law issues. This article focuses on the validity and application of three relevant colonial treaties (from 1900, 1901 and 1935) that defined the boundary, one of which (the 1935 Treaty) did not enter into force. Issues relevant to the determination of the borderline and sovereignty over the disputed islands and the unique challenges that may arise are discussed in light of the colonial treaties, relevant International Court of Justice jurisprudence and other international law principles, particularly uti possidetis juris.

Introduction

This article is written in the context of resolving African border delimitation / demarcation issues with reference to colonial boundary treaties. It focuses on how colonial boundary treaties that were entered into by France and Italy in 1900, 1901 and 1935 – and whose validity, denunciation or termination status is, as it applies, contentious – can be interpreted and applied to resolve the border dispute between Djibouti and Eritrea. The principle of uti possidetis juris, generally understood as adopted through the July 1964 Cairo Declaration (Resolution AHG/RES 16(I) of the Organisation of African Unity (OAU)) and reaffirmed by the African Union (AU), will be examined as it applies to the dispute.

The article intends to contribute to the debate on (a) the status of ratified boundary treaties not followed by the exchange of instruments of ratification and their role in the resolution of border disputes, and (b) the exact meaning of the Cairo Declaration, ie whether it unequivocally confirms sanctity of colonial boundaries or accommodates possibility for re-negotiating boundary treaties. The main rationales for writing this are (a) to recommend alternatives for the peaceful and successful resolution of the border dispute, and (b) to particularly address the possibility of the strategic Doumeira Islands and the adjacent reefs remaining uncovered by any border treaty provision which attributes sovereignty to them.

Origin of the dispute and initial attempts at resolution

The following sub-sections briefly describe the history of the drawing of the border between Djibouti and Eritrea, the origin of the June 2008 border conflict and previous attempts to resolve the conflict.

The border treaties

Djibouti (formerly French Somaliland) and Eritrea inherited their common border from three treaties (protocols) signed between their respective former colonial powers France and Italy. The border runs for approximately 110 kilometres from the Red Sea to the tri-border of Eritrea, Djibouti and Ethiopia atop the highpoint of Musa Ali, a dormant volcano 2,020 metres high.¹ The following is a brief presentation of the series of agreements that delimited the border.

On 20 March 1897, France and Emperor Menelik II of Ethiopia entered into a treaty which delimited the Ethiopia–French Somaliland border as extending from the peak of Madaha Djalêlo to Ras Doumeira.² The north-eastern section of the border stretched from the northernmost tip of Ras Doumeira to Bissidiro. However, that particular section of the border was never demarcated on the ground.³ The 1900 and 1901 protocols signed between France and Italy, discussed below, generally followed the 1897 Ethiopia–French Somaliland route.

After a border dispute in 1898, France and Italy signed a Protocol Agreement on 24 July 1900. This protocol defined the north-eastern section of the border as running from the northernmost tip of Ras Doumeira on the Red Sea coast, through the watershed along the peninsula for approximately 18 kilometres and then heading along a straight line – delimited by the 1901 protocol but never marked – towards the southwest to Bissidiro on the banks of the We'ima (Oueima) River.⁴ Under this protocol, Ras Doumeira fell under the territory of what is now Djibouti. Article 3 of the 1900 protocol further provided that until France and Italy could resolve the issue of which state would hold sovereignty over the Doumeira Islands and the unnamed adjacent islands, both colonial powers would refrain from attempting to occupy them and would prevent any third party from occupying them. Article 2 of this protocol provided for a special commission to delimit the boundary inland from Ras Doumeira. The protocol of 10 July 1901 basically endorsed the 1900 protocol⁵ and stated that the Doumeira borderline “mentioned in Article I of the Franco-Italian Protocol of 24 January 1900 begins at the farthest tip of Cape Doumeira, and then follows the watershed of the promontory bearing the same name”.⁶

During the mid-1930s, France and Italy pledged to assist each other in matters relating to their colonial interests in Africa.⁷ Accordingly, on 7 January 1935 the then French prime minister Pierre Laval and the Italian head of government and foreign minister Benito Mussolini signed eight different agreements (together known as the Mussolini–Laval Accord), only four of which were published.⁸ One of the documents published was an agreement for the settlement of respective French and Italian interests in Africa, also known as the Treaty of Rome (the 1935 Treaty), which, inter alia, replaced the 1900 and 1901 protocols. The 1935 Treaty is partly an affirmation of an agreement between France and Italy in 1934, whereby France ceded a small strip of French Somaliland territory west of Der Eloua, including the islands of Ras Doumeira and Jazirat Sawabih.⁹ Under article 4 of the 1935 Treaty, the previous border of 1900 / 1901 was pushed south-eastwards parallel to the topography of the previous border, hence placing Ras Doumeira under Italian territory.¹⁰ Under article 6 of the 1935 Treaty, France recognized Italian sovereignty over the Doumeira Islands and the adjacent reefs.

Article 7 provided that “[t]he present Treaty will be ratified and the ratifications will be exchanged within the shortest period possible. It will enter into force the day of the exchange of ratifications.”¹¹ The 1935 Treaty was separately ratified by the French and Italian parliaments.¹² The French Chamber of Deputies authorized ratification of the treaty by a vote of 555 to 9 and the Senate unanimously approved it in March 1935. The Italian Parliament also overwhelmingly approved the 1935 Treaty in May of that year and ratified it the following month.¹³ There is differing information that Italy did not ratify the 1935 Treaty while France ratified the entire Mussolini–Laval set of agreements.¹⁴ Other sources claim that the French Senate refused to ratify the 1935 Treaty.¹⁵ The International Court of Justice (ICJ) noted in 1994 in Libya v Chad, “[a]lthough ratification of the [1935] treaty was authorized by the parliaments of both parties, instruments of ratification were never exchanged, and the treaty never came into force”.¹⁶ The 2008 UN fact-finding report on the situation between Djibouti and Eritrea, noting this absence of exchange of instruments of ratification of the 1935 Treaty, stated that “[u]nder the 1935 agreement, Ras Doumeira and Doumeira Island, both of which were ‘seized’ by [Eritrean Defence Forces] in March 2008, form part of Eritrean territory, by Eritrean reasoning. Absent exchange of instrument of ratification, Djiboutians have all along assumed that the [1900 and 1901] Protocols still apply.”¹⁷ Only the eastern part of the border is disputed. Eritrea claims a total of 776 square kilometres south of the 1901 line, based on the 1935 Treaty, whereas Djibouti claims the same area based on the line of the 1901 treaty. The western half of the border was defined with precision and demarcated in 1954–55 by a French–Ethiopian protocol.¹⁸

In the build-up to the Second World War, Italy indicated that it would support Germany, and it became clear that this fundamental change of circumstances would not make it possible to honour the 1935 Treaty. In 1938, after Italy first denounced the 1935 Treaty, the entire set of agreements made in January 1935, including the 1935 Treaty, was repudiated by both countries.¹⁹

Eritrea continued to be an Italian colony until 1941, when Italy was defeated in Eritrea by Great Britain. On 12 December 1950 the UN General Assembly decided for Eritrea to be federated with Ethiopia. The federal arrangement was abrogated in November 1962.²⁰ An armed struggle for independence broke out in 1961; Eritrea obtained de facto independence in 1991 and de jure independence in 1993 following a referendum.²¹ Djibouti gained its independence from France on 27 June 1977. Djibouti and Eritrea inherited their respective territories, although without sufficient clarity on which one of the colonial treaties (the 1900 and 1901 protocols or the 1935 Treaty) defines their border.²²

The border conflict and the mediation process

Development of the dispute

From the mid-1990s, Djibouti and Eritrea were engaged in intermittent conflict regarding their respective territories along the border between them. The conflict mainly focused on the ownership of the Ras Doumeira area and the Doumeira Islands, which are adjacent to the extremely strategic strait of Bab-el-Mendeb along the southern tip of the Red Sea. In April 1996, for instance, the two countries nearly went to war when a Djiboutian official claimed that Eritrea had shelled the village of Ras Doumeira. The tension cooled off after Eritrean troops retreated from the disputed area, and Djibouti retracted the allegations in May 1996.²³ In June 1998 Djibouti deployed its military forces closer to the border to patrol the area and prevent any incursion.²⁴ Diplomatic relations between the two countries deteriorated towards the end of 1999, and both states continued to accuse each other. Eritrea accused Djibouti of supporting Ethiopia in the 1998–2000 border war with Ethiopia, while Djibouti charged Eritrea with supporting and abetting Djiboutian insurgent groups.²⁵ However, between 2000 and 2008 relations between the two countries generally improved in economic, political, social, cultural, diplomatic and security matters.²⁶

On 16 April 2008 Djibouti reported that Eritrean forces had entered Djiboutian territory and had started to dig trenches on both sides of the border.²⁷ Eritrea denied these accusations of penetrating into Djiboutian territory and claimed that the border was clear and indisputable.²⁸ Diplomatic efforts to resolve the dispute increased.²⁹ The tension could not subside, and Djiboutian accusations and Eritrean denials continued.³⁰ On 23 April 2008 Djiboutian president Ismaïl Omar Guelleh called Eritrean president Isaias Afwerki, and both leaders agreed to resolve the situation through negotiation.³¹ Military officials from both states met on 24 April to compare their respective versions of border maps.³² The tension could not, however, subside and Djibouti sent its troops to the border.³³

As part of the diplomatic effort to resolve the conflict, the African Union Peace and Security Council (AUPSC) conducted an exchange of views to address the situation and called on both sides to show restraint and resort to dialogue.³⁴ The Arab League also decided on 4 May 2008 to send a fact-finding mission to the disputed area. The Peace and Security Council of the Arab League requested Amr Mussa, the secretary general, to contact the AUPSC and the chairperson of the African Union Commission to launch an Arab-African initiative to settle the dispute.³⁵ These efforts could not ease the tension, and from 10 to 12 June 2008 the armies of the two states fought around Ras Doumeira mountain, with dozens reportedly killed and prisoners of war allegedly held.³⁶ While Djibouti continued to accuse Eritrea of penetrating into Djiboutian territory and launching the armed attack, Eritrea repeated its previous denials.³⁷

The UN Security Council (UNSC) met on 24 June 2008 to discuss the situation.³⁸ A UN fact-finding mission was sent to the region and reported that the standoff could have “a major negative impact on the entire region and the wider international community”.³⁹ Following the passing of Resolution 1862 (2009) that urged both states to settle the matter peacefully and withdraw their troops to the status quo ante,⁴⁰ the UNSC passed Resolution 1907 (2009) on 23 December 2009 to impose various targeted sanctions on Eritrea for failing to comply with Resolution 1862 (2009), as well as for its destabilizing role in the Somalia conflict. The part of Resolution 1862 (2009) pointing to the conflict with Djibouti condemned Eritrea's failure to withdraw its troops from the disputed territory and for not providing information regarding any Djiboutian prisoners of war that it may have held.⁴¹

The Qatari mediation process

On 6 June 2010 Djibouti and Eritrea agreed in Doha to solve the dispute through mediation led by the Emir of Qatar.⁴² In 2017 Qatar, without notifying Eritrea or providing reasons to either country, terminated the mediation process and pulled its peacekeepers out of the disputed territory, presumably because Djibouti and Eritrea appeared to side with the countries of the “quartet” (Saudi Arabia, the United Arab Emirates, Egypt and Bahrain) which were engaged in a multidimensional conflict with Qatar.⁴³ On 17 September 2018 President Guelleh and President Afwerki briefly met in Jeddah and promised to restore relations between their countries, although without signing a formal agreement.⁴⁴ On 14 November 2018 the UNSC, encouraged in part by Eritrea's warming relations with Somalia and Djibouti, unanimously lifted the sanctions on Eritrea through Resolution 2444 (2018), under the condition, among others, of a six-monthly review of progress on the normalization of relations between Djibouti and Eritrea being submitted to the UNSC.⁴⁵ Despite the lifting of the sanctions, however, the disputed territory between Djibouti and Eritrea has not yet been officially demarcated by any process.

Key issues in relation to the colonial boundary agreements

Legal analysis regarding the Djibouti–Eritrea border conflict relies on the application and interpretation of the three colonial treaties signed between France and Italy. As mentioned above, Eritrea relies on the 1935 Treaty because it believes that the 1900 and 1901 protocols were replaced by this treaty, while Djibouti relies on the 1900 and 1901 protocols as the 1935 Treaty was not ratified and nor were its instruments of ratification exchanged.⁴⁶ The contention is therefore which one of the two possibilities (the 1900 and 1901 protocols or the 1935 Treaty) applies to determine the border. The following section discusses the respective status, relevance and application of these colonial treaties based on relevant principles and practices of international law. These issues will be addressed by discussing the following points: (a) state succession and continuity of colonial borders (uti possidetis); (b) which agreement determines the border inherited by Djibouti and Eritrea; and (c) the retroactive application of the principle embodied in article 62(2)(a) of the 1969 Vienna Convention on the Law of Treaties (VCLT).

State succession and continuity of colonial borders (uti possidetis)

The present-day territories of Djibouti and Eritrea are creations of colonial treaties. Under international law, newly created states operate under what is known as the “clean slate” or “non-transmissibility” rule. Article 16 of the 1978 Vienna Convention on State Succession in Respect of Treaties (VCSS) provides that “a newly independent State is not bound to maintain in force, or become a party to, any treaty by reason only of the fact that at the date of the succession of states the treaty was in force in respect of the territory to which the succession of states relates”. One of the exceptions to this rule – or, according to Hillier, the only exception – is the succession of treaties establishing boundaries or concerning other territorial matters.⁴⁷ In other words, while newly independent states are free to choose whether to be bound by treaties entered into by governments that previously governed or occupied the territory, they are nevertheless bound by treaties establishing the borders of that territory. Article 11 of the VCSS provides that “A succession of States does not as such affect (a) a boundary established by a treaty, or (b) obligations and rights established by a treaty and relating to the regime of a boundary.”

In Burkina Faso v Mali, the ICJ stated that “there is no doubt that the obligation to respect pre-existing international frontiers in the event of a State succession derives from a general rule of international law, whether or not the rule is expressed in the formula of uti possidetis”.⁴⁸ This statement reflects general international practice on the sanctity of boundaries and is in line with article 62(2)(a) of the VCLT which provides that “a fundamental change of circumstances may not be invoked as a ground for terminating or withdrawing from a treaty if the treaty establishes a boundary”. A state that acquires a given territory automatically inherits the boundaries of that territory, whether the boundaries were fixed by a treaty or by the application of rules of customary law concerning title and acquisition of territory. The boundary so fixed stays permanent, despite the termination or nullification of the treaty that had established it.⁴⁹ If a treaty that contains provisions establishing boundaries and other provisions has to be terminated or withdrawn from due to a fundamental change of circumstances, the change may only be invoked in respect of the provisions that are not related to the boundary establishment, since the purpose of article 62(2)(a) is to provide for the stability of boundaries.⁵⁰

The issue of succession of border treaties becomes more significant in the case of colonial border treaties. The breakup of Spanish and Portuguese empires and the consequent change of their former Latin American administrative regions into independent republics brought with it the need to negotiate the boundaries of the new republics. To avoid the cumbersome process of negotiating borders anew and to maintain stability, the new republics which started to emerge in the 1820s adopted the old Roman principle of uti possidetis to accept colonial administrative boundaries as their presumptive international boundaries.⁵¹ Acquisition of title to territory through uti possidetis is presumptive because the principle requires that colonial boundary lines be unambiguous. If there is ambiguity, proof may be brought by either effectivités (ie activities on the ground tending to show the exercise of sovereign authority by the party engaging in that activity or its predecessor) or, failing effectivités, an adjudicating organ may refer to equity to establish the line of the border.⁵² However, Shaw states, “this does not mean that an ambiguous provision cannot be settled through the subsequent practice of the parties, nor that the parties themselves may not subsequently decide to modify that provision”.⁵³ The Arbitration Commission of the European Conference on Yugoslavia emphasized that except where otherwise agreed, former boundaries become frontiers protected by international law, and added that “it is well established that, whatever the circumstances, the right to self-determination must not involve changes to existing frontiers at the time of independence (uti possidetis juris) except where the states concerned agree otherwise”.⁵⁴

Succeeding states are free to redraw their border by express or implied agreement.⁵⁵ African leaders likewise “deliberately defined and stressed the principle of uti possidetis”⁵⁶ through Resolution AHG/RES 16(I), entitled “Border disputes among African states”, which was adopted during the First Ordinary Session of the Assembly of Heads of State and Government of the OAU held in Cairo on 17–21 July 1964.⁵⁷ The resolution declared that “all member states solemnly declare and pledge themselves to respect borders existing on their achievement of national independence”. The Chamber of the ICJ in El Salvador v Honduras observed that in the case of inherited colonial boundaries, the principle of uti possidetis was first recognized in Spanish America, where former colonial boundaries between administrative units became international boundaries of the new independent states.⁵⁸ Burkina Faso v Mali was the same as El Salvador v Honduras in the sense that it involved two former administrative units of the French colony in West Africa.⁵⁹ The Burkina Faso v Mali Chamber added that the previous colonial status of the two independent states in dispute does not matter as far as application of uti possidetis is concerned, because it is a general rule of international law:

“The territorial boundaries which have to be respected may also derive from international frontiers which [as in the case of Djibouti and Eritrea and the borders of most African countries] previously divided a colony of one State from a colony of another, or indeed a colonial territory from the territory of an independent State, or one which was under protectorate, but had retained its international personality.⁶⁰ There is no doubt that the obligation to respect pre-existing international frontiers in the event of a State succession derives from a general rule of international law, whether or not the rule is expressed in the formula uti possidetis. Hence the numerous solemn affirmations of the intangibility of the frontiers existing at the time of the independence of African States, whether made by senior African statesmen or by organs of the Organisation of African Unity itself, are evidently declaratory rather than constitutive: they recognise and confirm an existing principle, and do not seek to consecrate a new principle or the extension to Africa of a rule previously applied only in another continent.”⁶¹

Moreover, article 4(b) of the Constitutive Act of the AU recognizes the principle of “respect of borders existing on achievement of independence”.⁶² The arbitral tribunal in Guinea v Guinea-Bissau noted that the 1886 land boundary convention between colonial powers France and Portugal

“[r]emained in force between France and Portugal until the end of the colonial period, and became binding between the successor States by virtue of the principle of uti possidetis. This principle [conforms] … also with the Vienna Convention of 23 August 1978 on the Succession of States in Respect of Treaties. The relevant provisions of this … convention, which is not yet in force … are nonetheless held to reflect customary rules of international law.”⁶³

In Burkina Faso v Mali, the parties submitted a compromis to the Chamber of the ICJ, where they specified that the settlement of the dispute should be based upon respect for the principle of the “intangibility of frontiers inherited from colonisation”, ie uti possidetis. The Chamber, like the arbitral tribunal in Guinea v Guinea-Bissau, noted that the principle had in fact developed into a general concept of contemporary customary international law.⁶⁴ The Chamber reasoned that African countries adopted the principle of uti possidetis to “prevent the independence and stability of new states being endangered by fratricidal struggles provoked by the challenging of frontiers following the withdrawal of the administering power”.⁶⁵ Thus, although neither Djibouti nor Eritrea are state parties to the VCSS, the general or customary international status of its key provisions, such as article 11 cited above, can make it relevant in settling the dispute between the two countries.⁶⁶ The two countries succeeded the border as fixed by previous boundary treaties and as it existed at the respective dates of their accessions to independence.

The main issue in addressing the Djibouti–Eritrea dispute is not whether uti possidetis will (or should) apply, since the two parties do not disagree on the clarity or completeness of the line indicating the entire border; the 1900, 1901 and 1935 borderlines are clear and complete. The point of contention remains whether the 1900 and 1901 protocols or the 1935 Treaty should apply to demarcate the border. As there is no bilateral border agreement between Djibouti and Eritrea that alters the 1900, 1901 or 1935 line, therefore, there may be little need to resort to the principle of effectivités or equity to identify and demarcate the Djibouti–Eritrea border. It may be concluded that uti possidetis will apply to the dispute between these two members of the AU, which has maintained the OAU's adoption of uti possidetis.

However, while some call the OAU / AU resolution an “African uti possidetis juris”,⁶⁷ others disagree. In his separate opinion in Burkina Faso v Niger, where the Court held that the OAU resolution embodies uti possidetis juris, Judge Abdulqawi Yusuf argued that uti possidetis and the OAU resolution should not be considered the same.⁶⁸ He stated that these two principles are different “with regard to their origin and purpose, their legal scope and content and their legal nature”.⁶⁹ He added that the official documents of the OAU and its successor, the AU, do not mention uti possidetis juris in relation to African conflicts or territorial or boundary disputes.⁷⁰ Such lack of mention or reference, Judge Yusuf added, is not because of lack of cognizance of the existence of uti possidetis juris by OAU member states but that compared to the Latin American situation mentioned above, “rather, different situations, and historical circumstances, dictated the adoption of different legal rules and principles”.⁷¹ According to Judge Yusuf, the 1964 resolution, which called for respect of boundaries inherited from colonial treaties, should be interpreted as prohibiting the use of force and the forceful occupation of territory to settle territorial disputes. He concluded:

“Consequently, it may be said that the principle of respect for boundaries in the Cairo Resolution places the boundaries existing at the time of independence in a ‘holding pattern’, particularly to avoid armed conflict over territorial claims, until a satisfactory and peaceful solution is found by the Parties to a territorial dispute in conformity with international law, or until such time as closer integration and unity is achieved among African States in general, or between the neighbouring countries in particular, in keeping with the Pan-African vision. As such, it implies a prohibition of the use of force in the settlement of boundary disputes and an obligation to refrain from acts of seizure of a portion of the territory of another African State.”⁷²

Surely, over and above limiting itself to uti possidetis – which depends on contrasting title and effectivités ⁷³– the OAU resolution reflects a wider principle, as Judge Yusuf also acknowledged,⁷⁴ aimed at guaranteeing peace and stability in Africa through respect for inherited borders and the peaceful resolution of disputes. The OAU and AU resolutions should not be read as isolating the prohibition of use of force from uti possidetis and recognizing only the former or the latter, but as embracing both notions toward the higher objective of maintaining peace and avoiding conflict between African countries.

In Burkina Faso v Niger, Judge Yusuf noted that both parties had already signed an agreement in 1987 to delimit their boundaries and that the ICJ should have interpreted and applied this agreement rather than applying uti possidetis,⁷⁵ because “the principle of uti possidetis juris had become redundant in this case as a result of the conclusion of the 1987 delimitation agreement between the two independent States”.⁷⁶ Without challenging the plausibility of Judge Yusuf's arguments, they are nevertheless not directly relevant to the Djibouti–Eritrea case. Firstly, the prevailing view (including the ICJ's interpretation) is that uti possidetis is well accepted and practised by African states, as reflected in the OAU resolution and the AU's Constitutive Act. Secondly, Djibouti and Eritrea have not signed bilateral agreements to replace or modify the colonial boundary treaties. Thirdly, whether to peacefully settle the dispute or to sign further bilateral agreements regarding the boundary, Djibouti and Eritrea will have to refer to, or depend on, the three colonial boundary treaties.

Which border was inherited?

As stated above, Djibouti and Eritrea have different views as to which treaty applies to demarcate their border. The UN fact-finding mission on the Djibouti–Eritrea crisis had rightly concluded that resolution of this border dispute depends primarily on “which of the colonial treaties and protocols should be accepted as the basis for defining [the] common border (1897 Abyssinia–France treaty, 1900 / 1901 France–Italy Protocols, 1935 France–Italy Treaty)”.⁷⁷ It is true that the 1935 Treaty explicitly replaced the 1900 and 1901 protocols, but there are three factors that need to be discussed to examine the validity, or applicability in any case, of the later treaty.

Validity of the 1935 Treaty

There is an issue with the ratification of the 1935 Treaty: the treaty (in article 7) expressly required exchange of instruments of ratification, which did not happen. As mentioned above, the ICJ noted that the 1935 Treaty never came into force because of this failure to exchange instruments of ratification, and therefore the ICJ referred to the treaty as “the non-ratified Treaty of 1935”.⁷⁸ It may be debated whether the failure to exchange instruments of ratification nullifies the 1935 Treaty, which was ratified by the French and Italian parliaments. Failure to exchange instruments of ratification as required by the treaty itself may render the treaty null and void. Article 16(a) of the VCLT, which has been held to reflect “traditional procedures”,⁷⁹ provides that “[u]nless the treaty otherwise provides, instruments of ratification … establish the consent of a State to be bound by a treaty upon their exchange between the contracting States” (emphasis added). This provision has been interpreted to mean that without the exchange of instruments of ratification which is expressly provided for in the treaty (as in the case of the 1935 Treaty), “the treaty has no effect qua contractual obligation and has no binding force”.⁸⁰

It may be counter-argued whether the exchange of instruments of ratification was, or should be considered, a “mere formality” and should not affect the validity of the 1935 Treaty, which was overwhelmingly endorsed by the Italian and French parliaments. There is at least one source which states that state practice in 1935 was moving away from the requirement to exchange instruments of ratification.⁸¹ However, this practice needs to be considered in light of the express requirement in the 1935 Treaty for the exchange of instruments of ratification, hence rendering the treaty of no effect.⁸²

The issue of the ratification and validity of the 1935 Treaty was brought before the ICJ in Libya v Chad. In this case, Libya relied on the 1935 Treaty to base its claim to the Aouzou Strip; the ICJ held that the 1935 Treaty never came into force because of the failure to exchange instruments of ratification as required.⁸³ However, the ICJ made it clear that it disregarded the treaty not because it was unratified (owing to the failure to exchange instruments of ratification) and inapplicable to determine the status of the territory, but because both Libya and France (succeeded by Chad) negotiated and concluded a new agreement in 1955 to define the two countries’ territories.⁸⁴ The 1955 Treaty recognized and annexed the texts of all colonial treaties that defined the territory, but did not include or annex the 1935 Treaty.⁸⁵ The ICJ stated that “[o]btaining Libyan acceptance of [the Anglo-French Agreements of 1898, 1899 and 1919], which entailed recognition of the inapplicability of the non-ratified Treaty of 1935, was important to the French”.⁸⁶ Libya did not question the validity of the 1955 Treaty.⁸⁷ Article 3 of the 1955 Treaty and its Annex I excluded any other treaty except the treaties specified therein. The ICJ added: “[the] drafting of Article 3 and Annex I excludes any other international instrument en vigueur, not included in the Annex, which might have concerned the territory of Libya. A fortiori is this the case of the non-ratified Treaty of 1935, which was never en vigueur and is not mentioned in the Annex.”⁸⁸ Nevertheless, the ICJ admitted, after mentioning the details of the 1935 Treaty, that “[o]f the treaties prior to 1955 bearing upon a boundary line in this region, the non-ratified treaty of 1935 was … the most detailed. Yet it was not mentioned in Annex I.”⁸⁹ The ICJ did not clarify, or did not want to comment on, whether it would have accepted the application of the 1935 Treaty to the dispute if that treaty were not excluded by the 1955 Treaty and its Annex I. It also did not explain whether the term it used to refer to the 1935 Treaty – the “non-ratified” 1935 Treaty – meant that the Court would not recognize the 1935 Treaty as applicable to the dispute.

As stated above, the emphasis of the ICJ in resolving the dispute was the 1955 Treaty, which expressly excluded the application of the 1935 Treaty to determine the Chad–Libya border. However, in the Djibouti–Eritrea case, there is no agreement signed and ratified after 1935 between the governments that have hitherto administered the respective territories of the two states to authoritatively determine the boundary and / or accord title to Doumeira and the adjacent islands. Thus, if the ICJ's line of analysis in Libya v Chad is to be applied to the Djibouti–Eritrea dispute, it is in the sense that the dispositive value of the 1935 Treaty should not be simply ruled out because of the failure of France and Italy to exchange instruments of ratification, as required by the treaty, despite ratifications by their respective parliaments.

Effect of the denunciation and repudiation of the 1935 Treaty in 1938

The effect of the denunciation and repudiation of the 1935 Treaty in 1938 should also be examined. This denunciation and repudiation may be taken to mean that the 1900 and 1901 protocols will determine the Djibouti–Eritrea border, because the non-enforceability of the 1935 Treaty means the continued applicability of the earlier protocols. But why would Italy and France officially denounce and repudiate a treaty that had already become null and void? Why should the 1900 and 1901 protocols be counted as applicable if both France and Italy had indicated, by ratifying the 1935 Treaty, that they wanted to replace those protocols? Unless both countries believed that they were bound by the 1935 Treaty, they would not officially denounce and repudiate it. It may be argued, therefore, that France and Italy, by their subsequent acts of denunciation and repudiation, rendered the 1935 Treaty valid even if would otherwise have been null and void owing to failure to exchange instruments of ratification.

If the 1935 Treaty is, as argued above, held to be valid, what is the effect of its denunciation and repudiation in 1938? Does this mean that the 1935 Treaty remained valid only until 1938, or that it was nullified as a result of its denunciation and repudiation in 1938? It cannot be denied that the “fundamental change of circumstances” (ie Italy's siding with Germany) necessitated the termination of the 1935 Treaty by both parties. The customary international law rule of rebus sic stantibus,⁹⁰ now embodied in article 62 of the VCLT, states that if a fundamental change of circumstances occurs after an agreement has been concluded, a party to that agreement may withdraw from or terminate it.⁹¹ However, under paragraph 2(a) of article 62, the principle of rebus sic stantibus does not to apply to treaties that establish boundaries. Moreover, it is an accepted rule of international law that if a treaty that establishes a boundary ceases to be in force because the parties terminated it for whatever valid reason, the boundary established by the treaty continues unless the parties expressly modify it by a subsequent treaty.⁹² Thus, if it can be shown that the 1935 Treaty, which modified the 1900 and 1901 protocols, is valid despite the failure to exchange instruments of ratification, the boundary established by it cannot be nullified by the subsequent acts of denunciation and repudiation which terminated it in 1938. In Libya v Chad, the ICJ stated that:

“A boundary established by treaty … achieves a permanence which the treaty itself does not necessarily enjoy. The treaty can cease to be in force without in any way affecting the continuance of the boundary … When a boundary has been the subject of agreement, the continued existence of that boundary is not dependent upon the continuing life of the treaty under which the boundary is agreed.”⁹³

The Eritrea–Yemen arbitral tribunal also held that “[b]oundary and territorial treaties made between two parties are res inter alios acta vis-à-vis third parties. But this special category of treaties also represents a legal reality which necessarily impinges upon third states, because they have effect erga omnes.”⁹⁴

Use of unratified treaties in resolving boundary or territorial disputes

The ICJ has indicated that unratified boundary treaties may be relevant for settling international boundary or territorial disputes. The 1935 Treaty may be analysed in this regard. In Cameroon v Nigeria, the ICJ noted that a delimitation agreement signed in 1913 between Great Britain and Germany had recognized the Bakassi peninsula as part of German territory (Cameroonian by succession).⁹⁵ However, Nigeria challenged the application of the 1913 agreement as invalid because the agreement, at least the provision concerning Bakassi, was not approved by the German parliament in accordance with German law. Cameroon responded that “the German Government took the view that in the case of Bakassi the issue was one of simple boundary rectification, because Bakassi had already been treated previously as belonging de facto to Germany and thus parliamentary approval was not required”.⁹⁶ The Court responded:

“The Court notes that Germany itself considered that the procedures prescribed by its domestic law had been complied with; nor did Great Britain ever raise any question in relation thereto. The Agreement had, moreover, been officially published in both countries. It is therefore irrelevant that the Anglo-German Agreement of 11 March 1913 was not approved by the German Parliament. Nigeria's argument on this point accordingly cannot be upheld.”⁹⁷

Alvarez-Jimenez, after examining the view of the ICJ in the first decade of the new millennium regarding boundary or territorial agreements, observed that that Court “[f]irst was strict in finding the existence of a boundary agreement between the parties relating to a particular territory. Secondly, once the Court decided that a boundary agreement existed, it was reluctant to declare its unlawfulness” and that “[s]tates that have subsequently invoked the nullity of boundary treaties have not found a receptive Court to uphold such claims”.⁹⁸ Regarding the issue of the validity of the 1913 Anglo-German Agreements raised in Cameroon v Nigeria, Alvarez-Jimenez, observing the Court's reasoning cited above, stated that “instead of delving into German law to assess whether the agreement was valid, the Court looked at the parties’ external behavior regarding it”.⁹⁹

In Qatar v Bahrain, the ICJ stated that although:

“[b]oth Parties agree that the 1913 Anglo-Ottoman Convention [relating to the Persian Gulf and surrounding territories] was never ratified … signed but unratified treaties may constitute an accurate expression of the understanding of the parties at the time of signature. In the circumstances of this case the Court has come to the conclusion that the Anglo-Ottoman Convention does represent evidence of the views of Great Britain and the Ottoman Empire.”¹⁰⁰

In fact, Reisman has argued that in Qatar v Bahrain, the ICJ “validated” the unratified convention of 1913 owing to the clarity of the text and the reference in a subsequently ratified treaty of 1914 to the part of the 1913 treaty that had determined the boundary line between Qatar and Bahrain.¹⁰¹ The understanding or intention of states over the subject matter of an unratified treaty can, therefore, be inferred from the text of the treaty or the circumstances surrounding its making, even if the treaty is not ratified or instruments of its ratification are not exchanged. Similarly, in El Salvador v Honduras, the Chamber of the ICJ, “being satisfied that the line of the uti possidetis juris in this area is impossible to determine … considers it right to fall back on equity infra legem, in conjunction with the unratified delimitation of 1869” and that, taking inspiration from the Chamber of the Court in Burkina Faso v Mali, the Chamber “considers that it can in this case resort to the line proposed in the 1869 negotiations … as a reasonable and fair solution in all the circumstances”.¹⁰²

It is admitted that the three disputes before the ICJ referred to in this section (Libya v Chad, Cameroon v Nigeria and Qatar v Bahrain) refer to the views of the ICJ on unratified treaties, not, as in the case of the 1935 Treaty, ratified treaties whose instruments of ratification were not exchanged. Nevertheless, if the ICJ has shown readiness to consider unratified treaties for determining boundary disputes, the ICJ's line of thinking would make it even more convenient to rely on the substance of, if not consider valid, treaties (such as the 1935 Treaty) duly ratified but whose instruments of ratification were not exchanged. Thus it can be argued that the text of the 1935 Treaty and the circumstance of its making, as well as the absence of discernible subsequent practice regarding the now-disputed territory, can be relied upon to ascertain the intention of the ratifying parties (and now their succeeding states) regarding the boundary line as well as title to the Doumeira Islands and their adjacent islands. It needs to be noted here that the 1900 and 1901 protocols had left the determination of title to the (sovereignty over) Doumeira Islands and their adjacent islands for future agreement between Italy and France and that the only future agreement signed thereafter is the 1935 Treaty.

Retroactive application of the principle embodied in article 62(2)(a) of the VCLT

We may ask whether article 62(2)(a) of the VCLT (ie the non-applicability of rebus sic stantibus to treaties establishing boundaries) governs acts of denunciation and repudiation of a treaty establishing a boundary carried out before the entry into force of the VCLT. In other words, does the principle of non-retroactivity, contained in article 4 of the VCLT, prohibit the use of article 62(2)(a) against the denunciation and repudiation of the 1935 Treaty in 1938, both the 1935 and 1938 instruments being signed before the VCLT?¹⁰³ The response to this question depends on whether article 62(2)(a) codifies customary international law and how far the customary nature of article 62(2)(a) of the VCLT goes back. In both Fisheries Jurisdiction and Gabčíkovo-Nagymaros Project, the ICJ held that article 62 of the VCLT, including the conditions and exceptions to which it is subject, is considered as declaratory of customary international law, although it does not constitute jus cogens.¹⁰⁴ Thus, should it be held that the 1935 Treaty can be relied upon, if not considered valid, to determine the border between Djibouti and Eritrea, article 62(2)(a) of the VCLT can be applied to bar the effect of the 1938 denunciation and repudiation.

Conclusions and recommendations

This article has highlighted that resolution of the Djibouti–Eritrea boundary dispute primarily centres on the application and interpretation of the colonial treaties concluded between Italy and France: the 1900 and 1901 protocols and the 1935 Treaty. Both parties have not disputed the clarity or completeness of these protocols or treaty; they only disagree on which treaty must apply to define the boundary and determine sovereignty.

The contention raises some fundamental issues relating to boundary disputes under international law. At the outset, the parties have not yet disagreed on the application of the inherited colonial boundary treaties to settle the dispute; therefore, the principle of respect for the borders existing at independence (uti possidetis) will be applied in resolving the dispute. This article has examined Eritrea's claim based on the 1935 Treaty and Djibouti's claim based on the 1900 and 1901 protocols in light of the perennial arguments made in boundary / territorial disputes by reference to relevant ICJ jurisprudence and the writings of scholars. It has been observed that the 1935 Treaty, being a border treaty, survives the denunciation and repudiation by France and Italy, but that its validity remains debatable given that the instruments of its ratification were not exchanged between the two countries as required. Although the ICJ observed in Libya v Chad that the 1935 Treaty never came into force because of the lack of this formality, the Court nevertheless never disregarded its relevance to resolve the dispute. In Qatar v Bahrain and Cameroon v Nigeria, the ICJ expressly stated that unratified treaties may, at least, constitute accurate expression of the understanding of the parties at the time of signature. In fact, Reisman has observed that in Qatar v Bahrain the Court validated the unratified treaty debated before it. The jurisprudence thus far developed before the ICJ is that unless the parties expressly exclude the application of an unratified border treaty before the Court, or in the absence of a subsequent treaty that explicitly modifies or repeals it, that treaty will be relied upon, if not considered valid, to determine a disputed border.

Through the 1935 Treaty both France and Italy made it clear that the Doumeira Islands and the adjacent reefs would fall under Italian sovereignty. Moreover, the failed mediation agreement brokered by Qatar appears to have been limited to demarcating the land border between the two countries; the ownership of the Doumeira Islands and the reefs would not be decided by it. Thus if the 1935 Treaty is held to be invalid and not relevant to determine the dispute between the two states, the situation will create a difficult scenario whereby no treaty or agreement will have governed the sovereignty of the strategically crucial Doumeira Islands and the reefs (because the 1900 and 1901 protocols left the status of the island and the reefs to be decided in the future).

We agree with the UN fact-finding mission on the Djibouti–Eritrea crisis, which observed that the existence of several colonial treaties, among others, makes the status of the borderline contentious, and that if a political process or the reactivation of existing bilateral mechanisms fail to resolve the dispute, “both parties could seek recourse in an arbitration process that would culminate in a border demarcation ruling”.¹⁰⁵ To date, the chances for resuscitating the now-frozen mediation process or initiating a new mediation process under Qatar or other mediators appear to have become slim. On 18 February 2019 Djibouti sent a letter to the UNSC acknowledging the positive development in the Horn of Africa after Eritrea and Ethiopia signed a peace agreement and stated that it wants a binding international arbitration to determine its border dispute with Eritrea.¹⁰⁶ This indicates that Djibouti no longer believes in political methods for settling the dispute, and Eritrea would be advised to follow suit in seeking binding arbitration (or recourse to the ICJ) to solve it. If only the 1900 and 1901 protocols are held to govern the border between the two countries, the arbitrators will likely be challenged by the issue of determining the title of the Doumeira Islands and the adjacent reefs, which will be left with no treaty or, in the absence of a treaty, another principle of international law to ascertain their title. The arbitrators may have to rely on an agreement between Djibouti and Eritrea either to make a special joint agreement regarding the determination of title to the Doumeira Islands and the reefs or to allow the arbitrators to use equity to determine the title – as in the previous positive Eritrean experience in the maritime delimitation award in the Eritrea–Yemen territorial dispute, where the tribunal took into consideration the Islamic tradition of territorial sovereignty in determining sovereignty over, and allowing traditional fishing rights in, the disputed islands.¹⁰⁷ The trend of international jurisprudence on the determination of borders, however, has shown that unratified treaties or ratified treaties which have not entered into force because of the absence of exchange of instruments of ratification, as in the case of Djibouti and Eritrea, remain highly relevant to determine boundary lines if they clearly show the intent of the parties to such treaties regarding the boundary line and / or territorial sovereignty – hence the significance of the 1935 Treaty.

The above, however, does not disregard the well-advised recommendation that Djibouti and Eritrea could better resort to border dispute settlement processes that consider the unique nature of African border problems, because African borders are “generally considered to have been arbitrary acts, imposed by the European powers without reference to local conditions”.¹⁰⁸ Oduntan, for instance, argues that the predominant use of foreign-based adjudicatory mechanisms in trying to solve African border disputes has had the tendency of alienating those mechanisms from the traditional norms and values of African people. While not totally rejecting the use of conventional government-to-government dispute settlement mechanisms (negotiation, conciliation, mediation, arbitration, adjudication, etc), he strongly favours an understanding and application of multidisciplinary dispute resolution mechanisms and strategies, which can allow for a more holistic and effective treatment of African boundary disputes.¹⁰⁹ In any case, whichever dispute settlement mechanism they choose, Djibouti and Eritrea need to devise a comprehensive framework to effectively and definitively solve this dispute which is not limited to delimiting and demarcating the contentious border.¹¹⁰