I. Outsourcing Violence: The UK's Border Externalisation Architecture in the Western Balkans - Domestic Generator

Externalisation is often presented as something that happens at the edges of the state: at ports, in offshore camps, or through bilateral deals with third countries. But the report Outsourcing Violence: The UK's Border Externalisation Architecture in the Western Balkans argues that the process begins much earlier, inside the domestic asylum system itself. Its central claim is that externalisation should be understood as a three-stage architecture. First, the UK asylum system generates a growing population of people who are refused asylum but not removed. Second, those domestic pressures are translated into a wider apparatus of transnational enforcement, funding, and diplomatic negotiation. Third, the consequences land in the Western Balkans, where return hubs and externalised controls would be layered onto border regimes already marked by abuse, detention, and precarity. Rather than treating these as separate developments, the report shows how they form a connected chain in which domestic dysfunction becomes the justification for outsourced violence.

The first stage of that chain is what the report calls the domestic generator. This section is anchored in one of the report’s key findings: the UK asylum system is producing more refusals than it can turn into removals. That gap matters politically because it generates a visible population of people who are refused asylum but remain in the country, often for long periods, and this population is then used to justify calls for harder, faster, and more punitive migration control. In other words, externalisation does not start with a deal in the Balkans. It starts with the way the asylum system itself creates the conditions for those deals to look necessary.

The evidence presented in the report points to a system under sustained structural strain. As of 2024, the UK’s work-in-progress asylum caseload stood at around 224,700 people. More than half of applicants were waiting longer than six months for an initial decision, and a substantial share were waiting between one year and eighteen months. The report also notes that around 39 percent of the caseload was awaiting an initial decision, while 61 percent was awaiting further action after an initial decision had already been made. That distinction is important because it shows that the problem is not simply delay at the front end. A large proportion of the system is jammed after decisions are issued, producing a growing post-decision population that the state cannot effectively resolve.

This is where the refusal-removal gap comes into focus. Drawing on recent figures, the report notes that approximately 33,700 people were refused asylum in the relevant period, while only around 10,600 asylum-related returns took place. That means only about 31.6 percent of refused cases resulted in removal, leaving a gap of more than 23,000 people who were refused but not removed. These are not marginal numbers. They point to a system that produces refusal at a rate far beyond its capacity to carry out return. The report treats this gap not as a temporary policy failure but as a central contradiction of the UK asylum system: refusal decisions are generated at scale, but the legal, political, and operational conditions needed to enforce them do not exist at the same scale.

The evidence also shows why refusal cannot simply be treated as a stable endpoint. According to the report, around 47 percent of appealed refusals are overturned. That is a striking figure because it suggests that nearly half of challenged refusal decisions do not hold up when reviewed. Once that is taken seriously, the political discourse around “failed asylum seekers” starts to look much less straightforward. The category is not the product of a clean, reliable, and neutral administrative process. It is produced by a system in which decisions are often delayed, contested, and later reversed. The report therefore raises a deeper challenge: if refusal itself is unstable, then externalising people on the basis of refusal becomes even harder to justify.

The report then connects these figures to the broader political narrative around crisis. Backlogs, unresolved cases, asylum hotels, and high administrative costs are repeatedly presented in public debate as signs that the system is broken because it is too slow or too soft. But the report argues that this framing obscures the structural drivers of the crisis. The issue is not simply weak enforcement. It is that the asylum system is designed in ways that produce delay, contestation, and prolonged limbo. Instead of addressing those causes, political discourse reframes the outcome as a problem of insufficient control. That move is crucial because it turns a state-produced contradiction into a justification for tougher measures.

Financial evidence in the report reinforces this point. It notes that removal flights and escorting can cost about £22,000 per person, while the abandoned Rwanda scheme carried projected per-person costs of around £169,000. It also highlights reporting on substantial UK investment in externalisation-linked initiatives, including approximately £150 million for a Border Security Command and an additional £24 million directed toward prosecutors and operational partners in the Western Balkans. These figures are significant not because they prove externalisation is efficient, but because they show the scale of state willingness to spend on moving coercion outward rather than fixing internal dysfunction. The costs do not disappear. They are simply redistributed into new forms of enforcement infrastructure.

What emerges from this first stage is a different understanding of how externalisation works. It is not a bolt-on response to irregular migration, and it is not simply about deterrence at the border. It is a political project built on a domestic system that generates unresolved cases and then uses those unresolved cases to legitimise offshore solutions. The report’s evidence suggests that the refusal-removal gap is not an accidental side effect. It is the engine that allows policymakers to say that the only remaining option is to move people, procedures, and responsibility elsewhere.

Seen in that light, the domestic generator is more than a background condition. It is the first mechanism of externalisation itself. Before return hubs are proposed, before bilateral agreements are signed, and before people are transferred to other jurisdictions, the groundwork has already been laid through delay, refusal, and the political manufacture of crisis. The question is no longer simply whether externalisation works. It is what kind of asylum system is required to make externalisation appear both normal and necessary.