Public financial management reform in post-transition environments: lessons from Somalia

From effectively no functioning central PFM apparatus in 2012, Somalia progressed to a single treasury account, functioning appropriation processes, published budgets, biometric payroll for civil servants and security forces, and eventual qualification for HIPC debt relief, reaching Completion Point in 2023. That sequence was neither linear nor donor-designed in any tidy sense, but its broad shape is instructive: the reforms that stuck were those tied to flows of money the government needed and could not access otherwise.

Three patterns recur, in Somalia and in the comparative literature. First, reforms anchored to conditional flows, budget support tranches, debt relief milestones, and revenue-sharing agreements, outlast reforms anchored to technical assistance outputs. The incentive persists after the advisers leave. Second, reforms that automate a control, payroll biometrics, treasury single account sweeps, and e-procurement thresholds, outlast reforms that depend on discretionary compliance, because reversal requires a visible act rather than quiet neglect. Third, reforms survive political transition when their beneficiaries include actors outside the executive: parliaments that receive budget documents, suppliers paid faster, and civil servants paid reliably, which build constituencies that object to backsliding.

Comprehensive PFM reform strategies fail at a remarkable rate. The pattern is over-scoping: fifty-action reform matrices that exceed the implementation bandwidth of a ministry of finance operating with a few dozen professional staff. The counsel of the evidence is sequencing discipline: fewer reforms, chosen for their incentive structure rather than their position in a standard PFM maturity model, implemented to the point of irreversibility before the next tranche begins.

For Somalia's next phase, the frontier issues are intergovernmental: revenue assignment between federal and member state governments, transfer predictability, and harmonised customs. These are political settlements wearing technical clothing, and programming them as technical exercises repeats the known failure mode.

• 01Anchor reforms to money flows the government needs, not TA outputs.
• 02Prefer automated controls; reversal should require a visible act.
• 03Build constituencies beyond the executive into each reform.
• 04Sequence narrowly: implement to irreversibility before scaling scope.