The drafters of the 1990 Dublin Convention knowingly and willfully disregarded divergences among Member States' protection systems. Like a boomerang, the fallacy of this approach has now reappeared in cases where asylum seekers have challenged transfers under the Convention, tasking judges to tackle a problem which its drafters avoided. The core issue of this article is whether and when such divergences force a Member State to abstain from removing an asylum seeker to another Member State, provided that such removal entails a downgrading of protection.

Which approaches can be taken by authorities and courts seized with the legality of removal under the Dublin Convention? On one end of the spectrum, the formal approach relies on the fact that all Member States are bound by the 1951 Refugee Convention, the 1967 Protocol and the ECHR, and its proponents affirm the legality of removal regardless of factual differences amongst Member States in offering protection. On the other end of the spectrum, we find an empirical approach. Its proponents ask whether a specific asylum seeker would actually be treated in accordance with international law by the responsible Member State.

In Adan and Aitseguer, the U.K. Court of Appeal and the U.K. House of Lords have explicitly rejected the formal approach and opted for the empirical approach. So did the ECtHR in T.I. vs. the U.K. In the Minority Bosnians Decision, the Swedish Government has paid lip-service to the formal approach, but de facto applied the empirical approach, which effectively determined its outcome. Our analysis shows that there are good reasons for these preferences: the formal approach is untenable mainly for reasons of its logical and practical inconsistency. As long as Member States interpret their international obligations towards protection seekers differently, the empirical approach must be regarded as inherent in the prohibitions of refoulement contained in international law and thus a mandatory part of Dublin procedures.