Reconstruction programs are good at measuring what has been rebuilt. Evidence from Mostar, read alongside Nicosia and post-war Lebanon, suggests recovery should also track what happens to authority, coordination and public capacity while rebuilding is under way.
After war, rebuilding a city often begins with a straightforward question: what was destroyed? Homes, schools, bridges, water networks and public buildings can be surveyed, costed and reconstructed. A second question deserves equal attention: what kind of institutions are being built while reconstruction takes place?
Mostar, in Bosnia and Herzegovina, makes the distinction unusually visible. Between 1994 and 1996, the European Union Administration of Mostar (EUAM) spent about 170 million Deutsche Marks repairing 6,000 homes, 30 public buildings, 25 schools, 18 health facilities, 70 water projects and five bridges, while also establishing an urban planning system. Physical delivery was substantial. However, political and administrative reunification was much slower.
By 2003, the Commission for Reforming the City of Mostar described parallel systems as redundant layers of government and responsibility spread so widely that authorities could evade accountability. The 2004 reform abolished the six city-municipalities and moved toward a single city administration, a single budget and power-sharing safeguards. Subsequent reporting recorded progress in joining divided institutions. The sequence matters: institutional consolidation came after much of the visible reconstruction had already been delivered.
John Yarwood, EUAM’s Director of Reconstruction, later described the lesson in practical terms: once the humanitarian emergency had passed, repairing the institutions and processes of urban management could be even more basic than repairing damaged physical objects. Roads and buildings can be contracted. Shared authority, budgets and routines have to be built among actors who may only recently have been in conflict.
From lesson to policy
Nicosia shows why continuity matters. Since 1981, professionals from the city’s two communities have worked through the Nicosia Master Plan on shared urban problems. The mechanism did not resolve the island’s political division. It did, however, preserve a channel for repeated joint planning rather than treating cooperation as a temporary feature of one project. The United Nations Development Programme (UNDP) still identifies the plan as part of a longer bicommunal peacebuilding architecture.
Lebanon exposes the opposite risk. Kathrin Höckel’s study of post-war reconstruction links the concentration of reconstruction in Beirut’s central district, combined with weak integration of peripheral areas, to weakened government legitimacy and state control in marginalized regions. The lesson is not that every city needs the same administrative model. It is that the institutional destination of reconstruction cannot be left unspecified.
A practical response would be to require an institutional transition plan for major externally supported reconstruction programs, alongside the physical delivery plan. Before funding is committed, the program would identify which public body is expected to own each reconstructed service or function after exceptional arrangements end; which temporary bodies, budgets or coordination structures are being created; what staff, information and budget authority must move into ordinary institutions; and the milestones that trigger that transfer.
Rather than simply being a governance checklist, the plan would give donors and local authorities a point at which to test whether a project is strengthening the institution expected to carry the service later, bypassing it, or creating a parallel structure with no clear exit. It would also make transition reviewable before project closure, when staffing, procurement, data systems and budget lines can still be changed.
There is precedent for the underlying logic even if the instrument proposed here is not a standard post-conflict template. OECD guidance on post-conflict transition warns against parallel systems and stresses moving external delivery toward country systems where conditions allow. Its statebuilding guidance similarly cautions that aid can weaken state decision-making capacity even when its immediate intentions are constructive. More recently, the OECD’s 2026 assessment of Ukraine’s recovery architecture calls for time-bound measures to reduce coordination gaps, strengthen subnational capacity and align external partners. These approaches point in the same direction: delivery arrangements are themselves part of institutional recovery.
The Mostar case also suggests why the transition plan needs triggers rather than a vague promise of later integration. Temporary arrangements can harden into the structure of government. The 2003 reform commission was dealing with parallel institutions that had survived well beyond the immediate emergency. A transition plan agreed at the start would not guarantee integration, but it would force the question earlier: what is temporary, what is meant to endure, who is responsible for consolidation, and what event starts it?
Measuring reconstruction
Physical indicators would remain essential. But major programs should report a small set of institutional transition measures beside them: the share of reconstructed services transferred to the intended public authority; whether temporary units have an agreed closure or transfer date; whether operating budgets and staff are in place; and whether shared planning, data and coordination routines continue after exceptional funding ends.
That changes the policy question. A rebuilt water network is more than just a completed asset. It is also a future responsibility: someone must budget for it, maintain it, hold its data and answer when it fails. If those functions remain fragmented or external, physical completion can conceal an unfinished recovery.
The cases do not supply one institutional model for every post-conflict city, and they should not. They do show why reconstruction programs need an explicit route from exceptional delivery to durable public responsibility. Otherwise, governments and donors may discover the institutional consequences only after the contracts are closed and the structures created during recovery have become difficult to change.
Reconstruction lasts longer when it leaves behind more than repaired assets. It should also leave someone clearly able, authorized and resourced to govern what has been rebuilt.
