We Built Courts to Stop War Crimes. Then We Made Enforcement Optional.

There is a way in which the photograph tells you everything you need to know about what happened here. It is March 2023, and Vladimir Putin is landing at Johannesburg's O.R. Tambo International Airport for a BRICS summit. South Africa is a signatory to the Rome Statute, which means it is legally obligated to arrest Putin on sight—the ICC issued a warrant for his arrest six months earlier for the forced deportation of Ukrainian children. The photograph shows President Cyril Ramaphosa shaking Putin's hand on the tarmac. No arrest was made. No apology was offered. The warrant, in practical terms, was treated as advisory.

I thought about that photograph again this month when Omar al-Bashir—wanted since 2009 for genocide in Darfur—attended a state funeral in Cairo. Egypt is not an ICC member, so technically no law was broken. But fifteen African nations that are members have hosted al-Bashir at least forty-seven times since his indictment without arresting him. The court issued a warrant. The warrant was ignored. The system continued.

The Arrangement

We built the International Criminal Court on a particular theory: that law could substitute for power, that justice could be institutionalized, that the threat of prosecution would deter mass atrocity. The Rome Statute, adopted in 1998 and entering force in 2002, created a permanent tribunal with jurisdiction over genocide, crimes against humanity, war crimes, and—since 2018—aggression. One hundred twenty-three states are parties to the statute. The court sits in The Hague. It has a prosecutor, judges, a registry. It looks like a court.

But the Rome Statute contains a structural flaw so fundamental that it renders the entire architecture provisional: the court has no enforcement mechanism. Article 86 obliges state parties to cooperate with arrest warrants, but there is no penalty for refusal. The court cannot impose sanctions. It cannot freeze assets. It cannot deploy marshals. When a state refuses to arrest an indicted suspect on its territory, the court can refer the matter to the Assembly of States Parties, which can refer it to the UN Security Council, which can decline to act. This has happened thirty-one times.

The result is a system in which arrest is optional for the powerful and guaranteed for the weak. Joseph Kony, leader of the Lord's Resistance Army, was indicted in 2005 and has never been arrested; he operates in the border region between South Sudan and the Central African Republic, where no state has the will or capacity to apprehend him. Uhuru Kenyatta, indicted in 2011 for crimes against humanity during Kenya's 2007 post-election violence, became president of Kenya in 2013 and attended summits across Africa without incident. The charges were eventually dropped. The Kenyan government had refused to cooperate with evidence-gathering, and witnesses recanted under what prosecutors described as "intimidation."

The Pattern Emerges

There is a pattern to who gets arrested and who does not. Of the ten individuals currently in ICC custody, nine are African. The tenth is from the former Yugoslavia, transferred by a domestic court before the ICC had jurisdiction. Not one sitting head of state from a permanent UN Security Council member has been indicted. Not one senior NATO officer has faced charges. The selectivity is not accidental—it is structural.

The United States is not a party to the Rome Statute. Neither is Russia, China, India, or Israel. The ASPA—American Service-Members' Protection Act of 2002, nicknamed the "Hague Invasion Act"—authorizes the U.S. president to use "all means necessary and appropriate" to free any American detained by the ICC. In 2020, the Trump administration imposed sanctions on ICC Prosecutor Fatou Bensouda for investigating U.S. conduct in Afghanistan. The message was explicit: international law applies to others.

I am not sure what I expected when the court issued warrants for Israeli Prime Minister Benjamin Netanyahu and former Defense Minister Yoav Gallant in November 2024 for alleged war crimes in Gaza. Perhaps I thought the willingness to indict a Western-aligned leader represented a turning point. But Netanyahu visited Paris in January 2025 for a state visit. France is a founding member of the ICC. No arrest was made. The French foreign ministry issued a statement noting that sitting heads of state enjoy immunities under customary international law—a claim the ICC itself disputes. The warrant was, again, treated as symbolic.

What Enforcement Would Require

There is a reason the Nuremberg Tribunal succeeded. It was convened by the victors of World War II, and it had military enforcement behind it. The defendants were already in custody. The tribunal did not depend on voluntary compliance—it depended on occupation. The same was true, eventually, for the International Criminal Tribunal for the former Yugoslavia. Ratko Mladić, who commanded Bosnian Serb forces during the Srebrenica genocide, evaded arrest for sixteen years. He was finally captured in 2011 by Serbian security forces under intense pressure from the European Union, which had made his arrest a condition for Serbia's accession negotiations. The arrest was not voluntary—it was coerced.

If the ICC were serious about enforcement, it would require three structural reforms. First, automatic sanctions for non-compliance. Any state party that refuses to execute an arrest warrant would face suspension of voting rights in the Assembly of States Parties, exclusion from ICC-related aid programs, and referral to the UN Security Council with a presumption of enforcement action. Second, an independent judicial police force. The court currently depends on national police agencies to execute warrants. A standing ICC enforcement unit—modeled on Interpol but with arrest authority in signatory states—would eliminate the political buffer that allows non-compliance. Third, bilateral extradition treaties with non-member states, backed by trade conditionality. The EU could make extradition cooperation a requirement for preferential trade access. So could the United States.

None of these reforms will happen. They would require powerful states to surrender discretion, and powerful states do not join international institutions to constrain themselves—they join to constrain others.

Who Benefits

The current arrangement serves a purpose. It allows Western governments to claim moral leadership on international justice while ensuring their own officials remain beyond reach. It allows middle-power states to demonstrate commitment to human rights without risking diplomatic crises. It allows the court to issue warrants that generate headlines and grant symbolic legitimacy to victims without requiring anyone to pay the political cost of enforcement.

The losers are those who believed the promise. I think of the survivors of the 2008 Kenya violence who testified before the ICC, expecting prosecutions that never came. I think of the Syrian refugees who provided evidence of war crimes to ICC investigators in 2019, only to be told the court lacks jurisdiction because Syria is not a member state—and Russia would veto any Security Council referral. I think of the relatives of disappeared Ukrainians who watched Putin land in South Africa and understood that the warrant issued in their name carried no weight.

The Alternative We Refuse to Name

There is an alternative, but it requires acknowledging what enforcement actually means. It means sanctions that inflict economic pain. It means military interdiction of indicted suspects attempting to cross international borders. It means NATO or AU peacekeepers authorized to arrest individuals on foreign soil if host governments refuse. It means accepting that international law without enforcement is just a recommendation.

The objection is predictable: such measures would constitute imperialism, violate sovereignty, trigger diplomatic crises. But these objections assume that the current system is preferable—that symbolic justice is better than no justice, that issuing warrants no one will execute preserves the legitimacy of international law. I do not believe this anymore. I believed once.

The ICC's failure is not administrative—it is philosophical. We built a court to prosecute war crimes but refused to grant it the power to enforce its judgments, because enforcement would require powerful states to cede control. We wanted justice without coercion, accountability without consequences, law without force. We told ourselves this was possible. The evidence suggests otherwise.

The Reckoning

I return to that photograph of Putin shaking hands with Ramaphosa. What the photograph shows is not South Africa's hypocrisy or the ICC's weakness, though both are present. What it shows is the bargain at the center of the international legal order: we will build institutions that proclaim universal justice, and we will ensure they cannot enforce it. We will condemn atrocities and issue warrants and hold conferences in The Hague, and we will allow the perpetrators to move freely across borders because enforcement would require choosing justice over stability, law over diplomacy, consequences over relations.

There are people who still believe the system can be reformed, that stronger political will or better funding or more public pressure will transform the ICC into an effective tribunal. I am not among them. The flaw is not operational—it is definitional. A court without enforcement is not a court. It is a stage for the performance of justice, a venue where victims are invited to testify and warrants are issued and nothing happens. The performance serves a purpose. It reassures us that we have not abandoned the project of international law, that there are limits, that someone is keeping score.

But the score does not matter if the game has no referees. And the referees, it turns out, have no whistles. They have gavels and legal briefs and elegantly worded judgments, and the accused go home to state dinners while the victims wait in refugee camps for justice that will not arrive. This is the system we built. This is the system we maintain. We know what it is. We keep pretending otherwise.