A Hague return order sends a child back to a country, not to a parent — the Convention says so in a single sentence

by Karim Haddad · September 9, 2026

World Affairs Correspondent, The Globe Ledger

Almost every news report of an international child abduction case describes the outcome the same way: a court "awarded" or "refused" custody. Almost every one of those reports is describing something the court in question was forbidden to decide. The instrument they are reporting on, the Hague Convention of 25 October 1980 on the Civil Aspects of International Child Abduction, says so in a single sentence at Article 19: "A decision under this Convention concerning the return of the child shall not be taken to be a determination on the merits of any custody issue."

The Convention's purpose, set out in Article 1, is "to secure the prompt return of children wrongfully removed to or retained in any Contracting State" and to ensure that rights of custody and access recognised in one Contracting State are respected in the others. It is, in the technical sense, a jurisdictional instrument. It answers the question of which country's courts will hear the custody dispute, by putting the child back where the dispute properly belongs, and then it stops.

The machinery is built around that limit. Article 16 provides that once a court in the country the child was taken to has notice of a wrongful removal or retention, it "shall not decide on the merits of rights of custody until it has been determined that the child is not to be returned under this Convention." Article 17 closes the obvious workaround: the fact that a custody decision has already been given in, or is entitled to recognition in, the requested State "shall not be a ground for refusing to return a child under this Convention." A parent who flies to a friendly jurisdiction and obtains a local custody order has not obtained a defence.

What counts as wrongful is narrower than the everyday meaning of abduction. Under Article 3, a removal or retention is wrongful when it breaches rights of custody under the law of the child's habitual residence and those rights were actually being exercised, or would have been but for the removal. Article 5 defines rights of custody to include "the right to determine the child's place of residence," which is why a non-resident parent with a veto over relocation can invoke the Convention while a parent with generous contact but no such veto may not. Nothing in this turns on which parent is the better one. That question belongs to the court that will hear it later, somewhere else.

Two clocks govern the process. Article 12 requires the authority to order return "forthwith" where proceedings begin less than a year after the wrongful removal or retention; past that year, return may be refused if the child is now settled in the new environment. Article 11 sets the pace: authorities "shall act expeditiously in proceedings for the return of children," with a six-week benchmark before the applicant may demand an explanation for the delay.

The six weeks is where the design and the practice separate. The most recent global picture comes from the statistical study of applications made in 2021, prepared for the Hague Conference by Nigel Lowe and Victoria Stephens and issued as Preliminary Document No 19A in September 2023. It counted 2,180 return applications received by participating Central Authorities and found that reaching a final outcome took, on average, 207 days from the date the Central Authority received the application. The comparable figure was 164 days in 2015 and 188 days in 2008.

The same study found an overall return rate of 39 per cent — 23 per cent by judicial order and 16 per cent by voluntary return — with 13 per cent of applications ending in a judicial refusal. The authors describe that as a marked decrease on earlier surveys, listing the previous rates as "45% in 2015, 46% in 2008, 51% in 2003 and 50% in 1999." A treaty whose default is return now ends fewer than four applications in ten in a return, and the 2021 average of 207 days is longer than either of the two earlier averages the study sets beside it.

The refusals cluster on one provision. Article 13(1)(b) permits a court to decline return where there is "a grave risk that his or her return would expose the child to physical or psychological harm or otherwise place the child in an intolerable situation." The 2021 study records that this ground featured in 46 per cent of judicial refusals when sole and multiple reasons are combined, which it calls the greatest proportion ever recorded. Article 13 also allows a court to refuse where the child objects and has attained an age and degree of maturity at which it is appropriate to take account of those views — and Article 20 permits refusal where return would be incompatible with the requested State's fundamental principles on human rights and freedoms.

Two further limits are routinely missed. Article 4 provides that the Convention "shall cease to apply when the child attains the age of 16 years" — not when proceedings started, but on the birthday, which can end a live case. And Article 35 confines the Convention, as between Contracting States, to wrongful removals or retentions occurring after it entered into force between them. A removal that predates the treaty relationship is outside it entirely.

That phrase — between them — carries the Convention's least understood feature. Its status table, published by the Hague Conference on Private International Law and accessed on September 9, 2026, lists 103 Contracting Parties, with entry into force on 1 December 1983 and the most recent entries, Cabo Verde and Botswana, taking effect at the start of 2023. It is tempting to read that as a network of 103 countries all bound to one another. It is not.

Article 38 provides that any State may accede, but that "the accession will have effect only as regards the relations between the acceding State and such Contracting States as will have declared their acceptance of the accession." Accession is therefore bilateralised: each existing party decides, one by one, whether the Convention will operate between itself and the newcomer. The Hague Conference maintains a separate acceptances table for exactly this reason. Morocco's accession, for instance, has its own long list of accepting States with individual dates, running from the Netherlands on 24 March 2010 to Bosnia and Herzegovina on 22 June 2023 — thirteen years between two entries on the same line of the same treaty.

The practical consequence for a parent is severe and rarely explained in advance. Both countries can be on the list of Contracting Parties and the Convention can still be unavailable between them, because the acceptance step was never taken. There is no Hague remedy in that gap, only diplomacy and whatever the local courts of the destination country choose to do under their own law.

None of this makes the Convention a weak instrument. It makes it a specific one. It moves a child back across a border so that the country the child actually lived in can decide, under its own family law, with its own evidence, who that child should live with. The Convention's preamble is candid that this is a proxy for the real question, stating the drafters' conviction "that the interests of children are of paramount importance in matters relating to their custody" — and then building a mechanism that deliberately refuses to answer it. Article 19 is not fine print. It is the whole architecture, compressed into one line.