Military peacekeepers in UN missions accused of sexual exploitation and abuse are left to their own countries to investigate and punish — Security Council resolution 2272 of 2016 lets the UN send units home but gives it no power to prosecute

What is happening?

When a member of a UN peacekeeping force is accused of sexual exploitation or abuse, the United Nations can investigate administratively, send the person or the whole unit home, and support the victim. It cannot put the accused on trial. Under the status-of-forces arrangements that govern peacekeeping, the host country where the conduct took place does not exercise criminal jurisdiction over military contingents; the country that sent the troops does.

The UN has declared a zero-tolerance policy on this conduct since a Secretary-General bulletin of 2003, and the Secretary-General has reported to the General Assembly on special measures against it in a series of reports that runs from A/58/777 to A/80/644 (2025). The report covering 2024, A/79/789 of 2025-02-17, recorded 102 allegations in 2024 across peacekeeping and political missions, according to wire reports of it. Those reports state that all claims are referred to the home country of the peacekeeper, and that the report found most of those countries had not taken meaningful action.

In 2016 the Security Council adopted resolution 2272, which allows the Secretary-General to repatriate a unit where there is credible evidence of widespread or systematic abuse, and to replace all units of a sending country that does not investigate, hold perpetrators accountable and report on progress. It moved the power to remove troops to the UN. It did not move the power to prosecute them.

Whose problem is this?

The body that records the allegations and the body that can punish them are not the same body. The UN collects, counts and publishes; the power to try the accused sits in the capitals that sent them, and no opened source describes an obligation to publish the outcome of each case.

Where does this problem end?

The line is drawn at the gap between removal and prosecution because that is the part the UN has itself documented every year and has not been able to close through its own instruments. Staff of UN agencies and programme partners are excluded because the sending-state jurisdiction arrangement does not apply to them in the same way.

What is the state now, and what should it be?

Now

What it should be, in the words of the UN itself

The 2003 Secretary-General bulletin defines sexual exploitation and abuse as prohibited conduct for UN personnel. Resolution 2272 of 2016 expects each sending country to investigate allegations, hold perpetrators accountable and inform the UN of progress, and ties the replacement of its units to failure on those three points. That is the standard the UN has set for itself and for the countries that supply its troops. No source opened here sets a numerical target for prosecutions, a deadline for sending-state action, or a timeframe within which a paternity claim should be settled.

How big is it?

102 allegations in 2024 against personnel of UN peacekeeping and political missions, involving 125 victims, of whom 27 were children. The report described 2024 as the third year in the past decade in which allegations exceeded 100. The number of victims fell from 145 in 2023.

Alongside those annual counts sits a stock of 747 paternity and child support claims against peacekeepers on a UN dashboard, with 29 counted as resolved as of reporting in January 2025. No opened source says from what year that stock begins, so it cannot be converted into a rate. For the country with the second-largest number of claims, two had been resolved.

Separate from mission personnel, the same annual report counted 190 allegations against staff of UN agencies, funds and programmes in 2024, down from 284 in 2023, and 382 against non-UN personnel implementing UN programmes. These are outside the scope of this document but show that mission figures are one part of a larger reporting system.

These numbers count what was reported and recorded. They do not estimate what was never reported, and the affected population could not be derived from them.

Under what conditions does it arise?

1. Jurisdiction follows the uniform, not the place. Status-of-forces arrangements keep military contingents outside the criminal jurisdiction of the host country and leave it with the sending country. A victim who lives beside a mission would have to rely on a prosecution in a country they have never been to, under laws they may not know.

2. The UN can act only on the unit, not on the person. The UN can repatriate individuals, remove whole contingents under resolution 2272, and bar personnel from future service. Each of these ends the presence of the accused in the mission area. None of them produces a trial or a verdict.

3. Outcomes travel back only if the sending country sends them. Since 2015 the UN has published the nationality of accused personnel but not the outcome of each national investigation. Whether a case ended in prosecution, discipline or nothing depends on what the sending country chooses to report.

4. Child support depends on the same channel. A paternity claim needs the father to be identified, tested and held liable, and each of those steps runs through the authorities of the sending country. The UN dashboard counts the claims; settling them is outside its reach.

What has been tried?

Every attempt in the table strengthened what the UN itself can do — define, count, disclose, remove, assist. None of them changed who can prosecute. The pattern is visible in the table itself: the reforms accumulate on the UN side of the line and stop at it.

What was found?

Why is it still unsolved?

Enforcement absent — the rule exists and the UN enforces what it can reach, but the only authority able to impose a criminal sanction is a separate state that the UN cannot compel.

The UN has built an increasingly detailed system for recording misconduct and removing those accused of it. That system ends at a border. Once a peacekeeper is repatriated, what happens next is decided by the authorities of the sending country, and the UN can only ask. Resolution 2272 raised the cost of not answering by making replacement of all units possible, but the sanction falls on the contingent and the mission, not on the individual case, and it depends on the Secretary-General and the Council choosing to use it.

The second part is visibility. Because the outcome of each national case is not published in full, there is no public record that would show whether a given allegation led to a trial. An annual report can say that most sending countries have not taken meaningful action, but it cannot say which cases were closed and why, and so the finding recurs from one year to the next without a mechanism that turns it into a consequence.

The third part concerns victims who are owed something other than punishment. A paternity claim can be verified by DNA and settled by a payment, yet the stock of 747 claims shows 29 counted as resolved. The steps between identification and payment pass through the same sending-country channel as prosecution.

What observation would mean it is solved?

Candidates — (a) the annual special measures report records, for each referred allegation, whether the sending country investigated, prosecuted or closed it, and the share with a reported outcome approaches all cases (b) the share of paternity and child support claims on the UN dashboard counted as resolved rises substantially year on year (c) an arrangement is adopted under which a body other than the sending country can prosecute when the sending country does not act.

(a) alone measures reporting, not justice. A sending country can report that it investigated and closed a case with no finding. Full reporting would make the gap visible; it would not by itself close it.

(b) alone can move for reasons unrelated to accountability. Claims can drop off a dashboard because claimants give up or lose contact, and the sources do not say whether resolved means acknowledged paternity, payment, or both.

(c) alone is a text. A new jurisdictional arrangement would change what is possible; whether anyone is tried under it is a separate observation. The three have to be read together, and alongside the annual allegation counts, which can fall because conduct changed or because fewer victims come forward.

What is it connected to?

Fills with researchthe reporting of conflict-related sexual violence by parties to conflict, the accountability of staff in UN agencies and humanitarian organizations, the use of status-of-forces agreements in other multinational deployments, and victim assistance funding. Relation type and evidence grade were not confirmed in this round.

What these sources do not say

• When the stock of 747 paternity claims begins. The press report gives the total and the number resolved but not the starting year of the count; the dashboard itself returned an error and could not be checked.
• What resolved means for each of the 29 claims. The press report describes resolution as either legal responsibility for the child or some form of child support, without separating the two.
• The outcome of each national case. The UN publishes the nationality of accused personnel, not the result of each investigation or prosecution, and no opened source gives a count of prosecutions or convictions.
• How criminal jurisdiction works for police officers and civilian staff. The opened sources describe sending-country jurisdiction over military contingents only, although the 2024 count of 102 allegations covers mission personnel of all kinds.
• Which countries have not acted. The wire reports relay the finding that most sending countries took no meaningful action, without naming them; this document does not name them either.
• The position of sending countries. No statement by any troop- or police-contributing country explaining its handling of referred cases was found in the sources opened here.
• The distribution of the remaining 18 percent of 2024 allegations. Beyond the two missions that account for 82 percent, the per-mission counts were not in the opened sources.
• How much goes unreported. Sources describe conditions that discourage reporting, such as dependence on the mission, but none offers a numerical estimate of underreporting.
• The full text of the 2024 annual report. The report PDF, A/79/789, could not be parsed, so its figures are taken from wire reports of it rather than from the report itself.

See the evidence

No primary UN report was read in full. The resolution index and the UN list of Secretary-General reports were opened directly and carry the terms of resolution 2272 and the sequence of reports. The 2024 figures come from two copies of the same wire report of A/79/789, and the paternity figures from press reporting of a UN dashboard that could not itself be opened. The jurisdiction structure is described consistently by two independent secondary sources published two years apart. The UN position is represented by its own instruments — the 2003 bulletin, resolution 2272, the 2017 victims rights advocate and the annual reports; no public statement from a sending country was found. This is a Path A output (research-based definition), so observation_refs is empty and provenance_mode: press-derived.

People affected

Estimated range Not derivable

The reason and what is missing are listed under “What is missing” below