Building Institutions That Outlast the Project
Most institutional strengthening work leaves behind documents rather than capability. The difference is decided at procurement, long before anyone discusses handover.
There is a familiar shape to advisory work in the public sector. A programme is funded, a team is engaged, systems are designed, reports are produced, a final workshop is held, and the team leaves. Two years later the ministry is running the same process manually, or not at all, and a new programme is funded to address the same problem. Everyone involved is competent and well intentioned. The failure is structural, and it is decided at the point the work is bought, not at the point it ends.
The parallel institution problem
The most common failure is the parallel unit. A programme needs delivery capacity that the ministry does not have, so it creates one: an implementation unit, staffed by contracted specialists at salaries the ministry's own scale could not sustain, sitting beside the department it exists to support. The unit performs. It produces the reports, runs the system, meets the milestones. It also becomes the only place in the institution where the work is understood.
When funding ends, the capability leaves with the payroll. What remains is a department that watched the work being done, which is not the same as a department that can do it. Nobody chose this outcome; it follows from a procurement decision made early, when the fastest route to the first milestone was to hire around the constraint rather than through it.
Capability that lives in a contracted unit leaves when the contract does. Where it sits is decided at procurement, not at handover.
Documents are not capability
The second failure is subtler. An engagement produces a manual, a set of templates, a costed plan and a training pack, all technically sound, and the assumption is that capability has been transferred because it has been written down. It has not. A manual describes a process; it does not create the judgement required to run the process when something unexpected happens, and something unexpected always happens. A finance officer who has produced one quarterly report under supervision knows more than one who has read the manual twice.
The test worth applying is simple and uncomfortable: can the institution produce the next cycle of this output without us, using its own staff, on its own systems, at its own pay scale? If the honest answer is no, the engagement has produced a deliverable and not a capability, whatever the completion report says.
Designing for the second cycle
Work that lasts is designed around the second cycle rather than the first. That changes concrete things. The first cycle is run jointly and deliberately slowly, with ministry staff holding the pen and the adviser reviewing, which produces a worse first output and a better second one. Systems are built to the capacity that actually exists, on tools the institution already licenses and can administer, rather than to the capacity a well-resourced unit would have. Named counterparts are identified at the start and are accountable for the output from the beginning, not introduced at handover.
It also means accepting constraints that make the work harder. If a ministry cannot retain a specialist at its own pay scale, the system has to be designed so that it does not require one. A design that assumes staffing the institution cannot fund is a design that has scheduled its own failure.
The incentive problem is ours
Advisory firms are not rewarded for this. Contracts are written around deliverables, evaluated on completion and timeliness, and renewed on the strength of a relationship that a genuinely self-sufficient client no longer needs. A firm that builds itself out of a role is optimising against its own pipeline. That is worth stating plainly rather than pretending otherwise.
The way through is contractual. Capability can be specified as a deliverable: a defined output produced by named institutional staff without adviser input, verified before final payment. Handover can be scheduled as a phase with its own budget rather than a final week. Commissioners can ask, at procurement, where the capability will sit at the end and how that will be tested. Those are questions a buyer has to ask, because the supplier has every reason not to volunteer them.
The measure
The right measure of a governance engagement is what the institution can do afterwards, not what was delivered during it. That standard is inconvenient, and it is the only one that separates institutional strengthening from institutional substitution.