Provisional, but Not (Always) Pointless: Compliance with ICJ Provisional Measures

Written by

Provisional measures requests are keeping the International Court of Justice busy. In October alone, the Court heard requests for interim relief in two separate cases, Canada and the Netherlands v. Syria and Armenia v. Azerbaijan, and received a request in one more, Guyana v. Venezuela. This is in line with a broader trend. Parties are seeking—and obtaining—provisional measures more with increasing frequency: in the last decade, the Court has indicated provisional measures in eleven cases, versus ten in the first fifty years of the Court’s existence (1945-1995).

How are states responding? Surprisingly, although there have been some excellent theoretical discussions of the ICJ’s provisional measures orders (examples here and here), none appears, in recent years, to have systematically examined the practice of states with regard to compliance with such orders. This piece attempts to do so and concludes that compliance is modest but meaningful, hovering around 50% but decreasing in recent years as the Court has weighed in on more high-stakes, controversial cases.

The data: provisional measures since 2001

Under Article 41(1) the ICJ Statute, the “Court shall have the power to indicate, if it considers that circumstances so require, any provisional measures which ought to be taken to preserve the respective rights of either party.” Like interim injunctions issued by national courts, provisional measures seek to freeze the legal situation between parties to ensure the integrity of a final judgment. Questions persisted for a long time whether these measures were binding on the parties. But the Court put those doubts to rest in the LaGrand judgment, where it held that they were, given the Court’s “basic function of judicial settlement of international disputes.”

It makes sense, then, to examine state compliance with provisional measures after the LaGrand judgment in June 2001. Since then, the Court has issued 54 provisional measures in 15 cases. (This count considers the two Avena cases and the two cases between Armenia and Azerbaijan separately.)

To determine state compliance, I looked at documentary evidence of states’ behavior between the date of the order and the date of the ultimate judgment, either on preliminary objections or the merits. I examined both the Court’s own findings in its later judgments in each case as well as evidence provided by the states and third parties (e.g., newspapers, legal and political websites, and scholarly literature).

There is no space in a short piece of this nature to analyse or describe the facts regarding compliance in any detail or to set out why I consider compliance in particular cases to be full, partial or non-existent. However, my analysis is publicly accessible here and I encourage interested readers to check it for accuracy and provide feedback in the comments box below.

The evidence: mixed levels of compliance

As observers of the Court might expect, the results are mixed. Parties fully complied with provisional measures in one-third of cases (five out of fifteen), namely in:

  • Jadhav (India v. Pakistan) [measures ordered in 2017]
  • Immunities and Criminal Proceedings (Equatorial Guinea v. France) [2016]
  • Questions relating to the Seizure and Detention of Certain Documents and Data (Timor-Leste v. Australia) [2014 & 2015]
  • Request for Interpretation of the Judgment in Temple of Preah Vihear (Cambodia v. Thailand) [2011]
  • Avena and Other Mexican Nationals (Mexico v. United States of America) [2003]

There was partial compliance—the parties obeyed some measures, but not others—in one-fifth of cases (three out of fifteen):

  • Application of the CERD (Azerbaijan v. Armenia) [2021]
  • Application of the CERD (Qatar v. United Arab Emirates) [2018]
  • Certain Activities Carried Out by Nicaragua in the Border Area (Costa Rica v. Nicaragua) [2011 & 2013]

And there was no compliance at all in almost half of the cases (seven out of fifteen):

  • Allegations of Genocide (Ukraine v. Russia) [2022]
  • Application of the CERD (Armenia v. Azerbaijan) [2021-2023]
  • Application of the Genocide Convention (Gambia v. Myanmar) [2020]
  • Treaty of Amity (Iran v. United States) [2018]
  • Application of the ICSFT and CERD (Ukraine v. Russian Federation) [2017]
  • Application of the CERD (Georgia v. Russian Federation) [2008]
  • Request for Interpretation of the Judgment in Avena (Mexico v. United States of America) [2008]

So overall, parties meaningfully complied around 50% of the time. This is in line with the results at the individual measure level. Excluding procedural orders, where the Court instructs the parties to file compliance reports, states complied with 49% of ICJ provisional measures (22 out of 45 “substantive” orders).

Looking at trends over time, compliance has clearly decreased in recent years. Of the five cases where provisional measures have been ordered in the last five years, there has been partial compliance in one case (Azerbaijan v. Armenia) and no compliance in four. It is worth noting, though, that the four instances of defiance relate to some of the world’s major conflicts: the Ukraine-Russia war, the Nagorno-Karabakh conflict, the Rohingya genocide, and U.S. sanctions on Iran. This marks a departure from previous cases, which often revolved around more localized, non-violent disputes with somewhat lower stakes (see, for example, Equatorial Guinea v. France and Timor-Leste v. Australia).

The results are rosier when looking only at procedural measures. In the eight cases where the Court ordered states to report on the measures they took in response to the Court’s orders, the parties appear to have universally complied. This is not surprising given the low costs of compliance with such measures, especially since compliance reports are typically not public and thus carry few reputational consequences. Meanwhile, these reports are an opportunity for states to craft a favorable narrative to the Court and signal their respect for its authority.

One tentative explanation: the importance of compliance costs

What explains this track record of mixed and declining state compliance? While a detailed theoretical analysis is beyond the scope of this piece, one tentative explanation stands out based on the evidence above: compliance with provisional measures appears to be driven by the relative costs and benefits of obeying the Court.

On the benefit side, compliance can help resolve—or at least mitigate—the existing dispute. And more broadly, obedience to the Court can shore up a state’s reputation on the world stage, which is especially important for states that value their standing as law-abiding actors. But compliance is also costly: it often requires states to change behavior that they perceive as politically or economically advantageous. Even where changing that behavior is ultimately beneficial, states may have difficulty justifying the change to their citizens or interested stakeholders.

The most recent provisional measures orders provide good examples of cases where the perceived costs of compliance far outweigh the benefits. In the Ukraine v. Russia and Armenia v. Azerbaijan contexts, the leadership in both Russia and Azerbaijan perceive the two conflicts as critical to their regimes. And the Myanmar military regime has for decades used national security concerns to justify its atrocities against the Rohingya—behavior that it is unlikely to change given its perceived benefits to the government. A similar logic applies to the other Ukraine v. Russia case and Georgia v. Russia, both situations of conflict where de-escalation was likely perceived as costly. As the Court continues to address high-stakes human rights cases, it is likely to keep confronting this same calculus by the parties—and the ultimate outcome of non-compliance.

Conclusion

More than twenty years ago, Judge Oda warned that “the repeated disregard of the judgments or orders of the Court by the parties will inevitably impair the dignity of the Court and raise doubt as to the judicial role to be played by the Court in the international community.” It’s not clear that this has happened yet, nor that the Court should live in fear of this threat. Even where compliance is very unlikely, the Court may want to continue issuing provisional measures orders for other reasons: to signal its commitment to and help consolidate the law, to outcast violators of international rules, or to enable other institutions to enforce those rules. But recent experience counsels the Court not to expect compliance in most cases, especially where the costs to the parties outweigh the benefits.

Compliance with Provisional Measures Since 2001

Source: author’s analysis (available here)

Case

Year of Order(s)

Substantive compliance

Procedural compliance

Allegations of Genocide (Ukraine v. Russia)

2022

No

Application of the CERD (Azerbaijan v. Armenia)

2021

Partial

Application of the CERD (Armenia v. Azerbaijan)

2021-23

No

Application of the Genocide Convention (Gambia v. Myanmar)

2020

No

Yes

Treaty of Amity (Iran v. United States)

2018

No

Application of the CERD (Qatar v. United Arab Emirates)

2018

Partial

Jadhav (India v. Pakistan)

2017

Yes

Yes

Application of the ICSFT and CERD (Ukraine v. Russian Federation)

2017

No

Immunities and Criminal Proceedings (Equatorial Guinea v. France)

2016

Yes

Questions relating to the Seizure and Detention of Certain Documents and Data (Timor-Leste v. Australia)

2014-15

Yes

Yes

Request for Interpretation of the Judgment in Temple of Preah Vihear (Cambodia v. Thailand)

2011

Yes

Yes

Certain Activities Carried Out by Nicaragua in the Border Area (Costa Rica v. Nicaragua)

2011, 13

Partial

Yes

Application of the CERD (Georgia v. Russian Federation)

2008

No

Yes

Request for Interpretation of the Judgment in Avena (Mexico v. United States of America)

2008

No

Yes

Avena and Other Mexican Nationals (Mexico v. United States of America)

2003

Yes

Yes

Total Compliance Rate

 

53%

100%

 

Print Friendly, PDF & Email

Leave a Comment

Comments for this post are closed

Comments

L. Aubain says

November 3, 2023

Thank you for this very interesting post and for making the data available.

Just a minor point: in October the Court "heard" the parties in two cases, namely Canada and the Netherlands v. Syrian Arab Republic and Armenia v. Azerbaijan. The Court has not yet heard the parties in the Guyana v. Venezuela case. Hearings have yet to be scheduled in that case.

Matei Alexianu says

November 4, 2023

Many thanks for pointing this out. This has now been rectified.