Prospects for Judicial Settlement of the Danube Border Dispute Between Croatia and Serbia

Prospects for Judicial Settlement of the Danube Border Dispute Between Croatia and Serbia

[Thomas Bickl PhD researches dispute resolution issues between the EU and third countries, EU enlargement in the Western Balkans, and Law of the Sea issues. His book on the border dispute between Croatia and Slovenia and its implications for EU enlargement was published by Springer.]

This post aims at shedding some light on a territorial dispute along the Danube, Europe’s most international waterway, and scenarios for its judicial resolution. It will briefly outline the historical and international law context, the positions of Croatia and Serbia, and give some indication on tentative issues for consideration by the Court, should it be entrusted with the case in the foreseeable future.


The Danube border dispute between Croatia and Serbia has existed since the dissolution of Yugoslavia in 1991/92. The two countries share a joint section of the Danube River along 137 km (page 28) between the border with Hungary and Bačka Palanka/Ilok. Essentially, Croatia claims the border to run along the outer limits of its cadastres originating from a geodetic survey concluded in 1891 (under the Austrian-Hungarian Empire). Regulation works were carried out along the contemporaneously disputed section of the Danube shortly afterwards (1891-94) cutting channels to provide for improved navigation. The river has also considerably meandered since. All this has led to ‘pockets’ on the left bank of the river claimed by Croatia, and to a lesser degree also on the Danube’s right bank (which, according to Croatia, belong to Serbia; see fig. 1 below) and two river islands on the left-bank side of the main channel). Serbia itself claims the Danube’s navigable channel, known as Thalweg, as the common State border. The two parties entered into bilateral talks in the framework of an Inter-State Commission founded in 2002 holding meetings rather infrequently with no real progress over the last 20 years.

Worldwide, over one third of the total length of international land boundaries follows rivers (page 157). The Danube is an international waterway of 2,857 km (page 2), 1,071 km of which are State borders (page 106) linking Germany, Austria, Slovakia, Hungary, Croatia, Serbia, Romania, Bulgaria, Moldova, and the Ukraine. The Danube is a major component of the international waterway linking the Black Sea to the North Sea, the Trans-European (Rhine-Danube) Corridor VII from Sulina (Romania) to Rotterdam (The Netherlands). 

The land border with Hungary is fully delimitated and demarcated subject to a demarcation and maintenance agreement between Hungary and the SFR Yugoslavia (page 12) from 1983 to which both Croatia and Serbia are successor States. Serbia’s border with Romania along the Danube is fully delimitated according to the Thalweg rule. The delimitation is based originally on the Treaty of Sevres from 1920 that was technically amended by an agreement between Romania and the SFR Yugoslavia in 1975/77 (page 101, footnote 12) following the joint construction of the Đerdap I and II gates on the Danube with Serbia succeeding the SFRY regarding the above treaties. Croatia and Serbia signed a bilateral agreement on navigation and technical maintenance of international waterways (page 9) in 2009, and a corresponding Inter-State Commission was founded in 2010.

Fig. 1 Major pockets on the left (green) and right (red) bank of the Danube according to cadastral claim of Croatia with Serbia claiming the navigational channel of the Danube (blue). Source: Vukosav and Matijević, 2020: 195.

Prior to 1991/92, both Serbia and Croatia were constituent republics of the federal SFR Yugoslavia. In the spring of 1945, a domestic Boundary Commission proposed a provisional boundary line allocating the right-bank territory to Croatia and the left-bank area to Serbia. In July 1945, the Politburo of the Central Committee of the Communist Party of Yugoslavia (CPY), a quasi-federal body, endorsed that proposal. However, as Klemenčić and Schofield have noted (pages 14-16), no subsequent detailed delimitation has ever been carried out along the Croatia-Serbia boundary including the Danube. It is also important to note generally – as the Tribunal in Croatia/Slovenia (para 316), Radan (page 7), and myself (page 2) have – that no legal act has ever been adopted by any SFRY body which would establish and define the administrative boundaries between the Yugoslav republics.

Succession to State Borders

It is a universally recognised principle of international law that the former internal boundaries of a territorial entity become international borders after obtaining independence, known as uti possidetis juris, firmly established by the ICJ in e.g. Burkina Faso/Republic of Mali (paras 20 and 23) in 1986 and El Salvador/Honduras (para 44) in 1992. Originally applied in the context of decolonisation in Latin America in the 19th and later in Africa and Asia in the 20th century, uti possidetis was also used following the dissolution of Yugoslavia, as established in Opinion No. 3 of the Badinter Commission in 1991 and referred to by the Arbitral Tribunal in Croatia/Slovenia in 2017.

For delimitation purposes, we can draw on settled jurisprudence distinguishing between (i) legal title to territory (uti possidetis juris), and (ii) the effective control of an area (uti possidetis effectivités), with legal title carrying more weight than effectivités, as the ICJ noted in Nicaragua/Colombia in 2012 (para 66) and Benin/Niger in 2005 (paras 75-76), and the Tribunal in Croatia/Slovenia (para 340) confirmed in 2017.

In Croatia/Slovenia, the first-ever judicial resolution of a territorial dispute between successor States of Yugoslavia, the Tribunal further noted that, to establish effectivités, it requires evidence clearly pointing at the display of the exclusive power of the State, such as “the levying of taxes, the organisation of elections, conscription for military service, and law enforcement” which would take precedence over services such as the delivery of mail or the provision of telephone lines or electricity (paras 341; 343). Implementation is pending, however, as Croatia does not recognise the 2017 Final Award due to illegal communication between the Slovenian government and the arbitrator nominated by Slovenia, although the Tribunal subsequently reconstituted thereby remedying Slovenia’s violation of the Arbitration Agreement. It is important to note that the Final Award constitutes a binding settlement of the dispute under international law. Nonetheless, in 2020, the EU Court of Justice (CJEU) determined in Croatia/Slovenia (paras 102 and 106) that the award cannot be enforced through EU law, a rather formalistic view disregarding not least the pacta sunt servanda principle from the previous Arbitration Agreement between the two parties in the context of Croatia’s EU accession, as McGarry and myself have noted.

Claims of Croatia and Serbia

Croatia claims the cadastral limits of its districts and municipalities relating to uti possidetis juris. The cadastral data are based on a geodetic survey under the Austro-Hungarian Empire 1877-1891, i. e. from before the regulation works on the Danube. These cadastral limits were carried over to the Socialist Yugoslav Republic of Croatia after 1945 (author’s field notes at State Archive Zagreb, 21-09-2021).

Correspondingly, Croatia claims 115 km² of land on the Danube’s left bank (interview senior Croatian civil servant, Zagreb 21-09-2021). For the (scarcely populated) area of the northernmost pocket Karpanđa/Kenđija (see fig. 1 above; both the remainder of the left-bank pockets and the smaller right-bank pockets are natural reserves) and the two Danube islands further downstream, Croatia also claims effectivités in SFRY times, as Klemenčić and Schofield (pages 24-25) have demonstrated. The length of the disputed border, Croatia claims, stretches over 203 km (along some sections of the navigable channel of the Danube, but also along tributaries, and on land along the Croatian cadastral limits; interview senior Croatian civil servant, Zagreb 21-09-2021), whilst the length of the joint Croatian-Serbian stretch of the Danube is only 137 km. Croatia also points at Art. 7 of the Protocol between the Republic of Croatia and the Federal Republic of Yugoslavia (FRY) from 23 April 2002 (on file with the author) stipulating that any delimitation should be based on “cadastral and other relevant documentation”.

Serbia claims that there has never been any document succeeding the report of the Boundary Commission from 1945 (adopted by the federal CPY Politburo) fixing the Danube as the provisional boundary line between the Yugoslav Republics of Croatia and Serbia in a general way (see also Introduction). Therefore, it is the exact course of the river boundary which was now to be determined. International State practice in the context of customary international law and settled jurisprudence of international courts and tribunals clearly suggested that the navigable channel, known as Thalweg, was the appropriate delimitation line. Serbia was not able to accept the cadastral claims of Croatia as those neither reflected the substantial meandering of the river and regulation works from the late 19th century, nor were they in accordance with the existing laws in Serbia/Vojvodina. Therefore, the Thalweg line was to be newly measured and incorporated into a bilateral legal act between Serbia and Croatia (interview senior Serbian civil servant, Belgrade 20-09-2021).

The Thalweg principle surfaced for the first time in the Treaty of Luneville of 9 February 1801 delimitating the river boundary between between France and Germany, as Dimitrijević (page 14) has recalled. Generally, the Thalweg ‘line’ tends to be used for navigable boundary rivers, and, as Shah (page 366) posits, it may be said that the navigational freedom of riparian States is the paramount interest. It is worth noting, however, that the application of the Thalweg principle produces an area rather than a line as the navigable channel of the river (fairway) has a certain width (and depth) indispensable for navigation. According to the current Danube Region Strategy Fairway Rehabilitation and Maintenance Master Plan (page 7), the minimum fairway requirements for the Danube applicable to Croatia and Serbia (and Bulgaria and Romania) are 2.5 m of depth at low-water-level and 80 m in width.

Issues for the Court

As mentioned above, the two parties have been in touch bilaterally over the dispute infrequently and without any progress for 20 years. It would seem, therefore, that the dispute, if it is indeed to be resolved, is likely to be submitted to the ICJ at some point (arbitration does not appear to be a suitable means; see Succession to State Borders), tentatively involving a mediator to arrive at the Special Agreement.

As Thirlway observes, if one looks at the jurisprudence of International Courts and Tribunals, it appears that there is a strong emphasis on legal title/uti possidetis juris, and that in the absence of legal title, the judicial body will (have to) ground its deliberations in  effectivités (see also Succession to State Borders). This is apparent also with regard to navigable river boundary delimitation where Thalweg has develped into the default mode, as established by the ICJ notably in Benin/Niger in 2005 and previously in Botswana/Namibia in 1999.

The paramount issue, it seems, will be the deliberations of the Court as to the identification of acts or documents sufficiently establishing legal title, and to what degree effectivités will have to be considered. Three matters in particular come in here: (i) pertinent legislation that can be considered to be (still) in force, (ii) the scope of the Court’s deliberations with regard to a critical date, and, in the absence of the latter, (iii) the role of the effective management of the Danube as an international waterway.

(i) It is useful to note what the Tribunal in Croatia/Slovenia (para 348) determined about the status of cadastral records: 

“The cadastres evidence title to land and the location of the boundary: but as a matter of international law they do not definitively constitute either title or the boundary […] The cadastral limits do not have any inherent special status that entitles them to prevail automatically over any evidence that indicates that the administrative or ‘political’ boundary of a republic is different from that of a cadastral district.”

With regard to legislation proper, legal acts of Croatia and Serbia on the scope of their territorial units (regions and municipalities) naturally matter. It will, however, also be crucial to see how the Court would weigh the findings of the 1945 Boundary Commission approved by a quasi-federal body, and whether those could be regarded as still being in force – which would attach the river as the boundary a strong bearing and very likely give rise to subsequent delimitation by the Thalweg rule;

(ii) As to both legislation in force and effectivités, a critical date in the Special Agreement would obviously be crucial and a major point of contention: if there were a critical date and if it related to the disintegration of Yugoslavia, no factual post-independence and present-day legislation or control of an area would obviously be subject to consideration by the Court. Seen from both ends of the spectrum: a critical date around 1991 would seem to favour Croatia (invoking effectivités during SFRY times in the northernmost pocket and to some degree also on the two Danube islands); no critical date at all would seem to favour Serbia who may not least invoke the contemporaneous joint management of the river;

(iii) Relevant not only to a critical date, yet in a broader sense, is the role of the Danube as an international waterway. The Court already took note of that fact generally in Hungary/Slovakia (para 54). In addition, it may also want to look at the various international conventions governing the use of the Danube and obligations relating to e.g. the safety of navigation and the fairway maintenance. Relevant conventions in this regard include the Danube Convention from 1948 (signed in Belgrade with Yugoslavia as a signatory, and present-day Croatia and Serbia being successor States not only to Yugoslavia, but thus also to the Convention), and, more recently, the EU Danube Strategy from 2011 de facto positing, inter alia, a joint responsibility of both countries along the joint Danube section. The bilateral agreement between Croatia and Serbia from 2009 on the collaboration on navigation and maintenance is also likely to play a role with regard to the joint management of the river in terms of effectivités.


If the single unresolved river boundary left along the Danube were brought before the Court, we would see an interesting yet barely predictable interplay of considerations ranging from the identification of applicable legislation in terms of legal title to the weighing in of effective control of the area. Both uti possidetis juris and effectivités would be prone to contemplation over a considerable time period including the pre-independence and also the pre-1945 period. This scenario would be on the condition that the submission to the Court were based entirely on Art. 83(1) of the ICJ’s Statute with neither any supplementing bilateral treaty nor a critical date. Given the protracted nature of the conflict, it would seem that the parties would need a third-party mediator to arrive at the Special Agreement for submission to the ICJ (even when presupposing the above minimum scope) in the first place.

Print Friendly, PDF & Email
Courts & Tribunals, EU Law, Europe, Featured, General, Law of the Sea
No Comments

Sorry, the comment form is closed at this time.