For a century, budget reformers have tried to shift budgeting from controlling what money buys to managing what it achieves. This module traces that effort — programme budgeting and performance budgeting — and is candid about its limits, which are the multitasking and measurement problems you have already met in the Governance course, applied to the budget.
From line-item to programme to performance
- Line-item budgeting — the traditional form: money is allocated and controlled by input (so many shillings for salaries, so many for travel, per ministry). Excellent for control (you can check every shilling was spent on what it was authorised for) but says nothing about what was achieved — you can run a perfectly-controlled line-item budget that accomplishes nothing.
- Programme budgeting — organises the budget by programme/objective (a 'primary education' programme) rather than only by input, so the budget shows what the money is for, and allocation can be made between competing objectives. This is the structural reform underneath performance budgeting.
- Performance budgeting — goes further: it links the funding to results (outputs and outcomes), bringing performance information into the allocation decision so that, in principle, money follows results.
Three kinds of performance budgeting
How tightly is money linked to results?
Robinson and Last (IMF) distinguish three forms by how directly performance drives funding: • Presentational — performance information is reported alongside the budget but does not affect allocation. The numbers are shown, not used. (Most 'performance budgeting' in practice is this.) • Performance-informed — performance information is one input among many into allocation decisions; budgets are influenced by, but not mechanically tied to, results. (The realistic target for most governments.) • Direct/formula — funding is mechanically tied to output (e.g., a hospital paid per patient treated, a school per pupil). Powerful but dangerous, and viable only where output is well-measured and hard to game. The progression from presentational to direct is one of increasing power and increasing risk. Most failures come from attempting more linkage than the measurement system can support.
The gaming problem
The deeper the link between a measure and money, the stronger the incentive to game the measure — the multitasking problem from the Governance course, now inside the budget. Fund schools by exam pass rates and you may get teaching to the test, exclusion of weak pupils, or outright cheating; fund hospitals by patients treated and you may get inflated numbers and avoidance of complex cases; fund anything by a target and the target tends to be hit while the underlying goal is missed (Goodhart's law: when a measure becomes a target, it ceases to be a good measure). Performance budgeting works only where the chosen measures are hard to game and genuinely aligned with the outcome — a stringent condition that rules out mechanical funding for most multidimensional public services.
The attribution problem
Even honest measurement faces attribution: outcomes (children who can read, lower mortality) depend on far more than the programme's spending — on households, the economy, other services, and luck. So you often cannot tell whether a good or bad outcome was caused by the programme or by something else, which makes it unfair and unwise to fund the programme on the outcome alone. This is why performance budgeting tends to retreat to outputs (which the programme does control) even though outcomes are what we care about — and why rigorous evaluation (the Impact Evaluation course) is needed to establish the output-to-outcome link that performance budgeting assumes.
The low-capacity reality
Performance budgeting and the capability trap
Many developing countries adopt the apparatus of programme/performance budgeting — programme structures, indicator frameworks, performance reports — under donor encouragement, and end up with presentational performance budgeting at best: elaborate indicator documents that no one uses for allocation, layered on a system that still really runs on line-item control. This is the isomorphic mimicry of the Governance course: the form of a modern budget without the function. The honest guidance (Schick, Robinson): match the ambition to the capacity. Get the basics right first (a credible budget, a working classification, reliable execution), aim for performance-informed rather than direct/formula budgeting, use a few robust, hard-to-game indicators rather than hundreds of soft ones, and treat performance information as an input to judgement, not a machine that allocates money. Performance budgeting is a tool for a system that already has the basics, not a substitute for them.
Exercise
A government introduces performance budgeting in education, proposing to allocate each school's funding directly in proportion to its national-exam pass rate, to 'reward results'. (1) Classify which of the three forms of performance budgeting this is, and the risk that classification carries. (2) Apply the gaming problem specifically to the pass-rate measure. (3) Apply the attribution problem — why is the pass rate an unfair basis for funding? (4) Redesign the reform to be useful and feasible in a low-capacity setting, and justify each change.