Performance-Based Budgeting

Rather than allocating money based on how many staff a department employs or how much office space it occupies, performance-based budgeting ties funding to measurable results—the number of pothole repairs completed, the percentage of benefit applications processed on time, the test-score improvements in schools. It promises to make government spending more efficient and politicians more honest about what their money actually buys.

The contrast with traditional input-based budgeting

The conventional government budget is built on inputs: salaries for staff, purchase of equipment, rent for office space. A transportation agency receives funding based on the number of engineers and maintenance crews it employs. A welfare office receives money based on the number of case workers on staff. This approach has the virtue of simplicity—you can easily count employees and verify expenses—but it creates a perverse incentive: the department is funded for effort, not for results.

Performance-based budgeting inverts this. The same welfare office would receive a budget that rises as the percentage of eligible applicants receiving benefits increases, or falls if processing times lengthen. The transportation agency would be rewarded for reducing pothole complaints per kilometre of road maintained, not for how many trucks it owns. The shift is subtle but profound: departments must now prove they are spending money wisely rather than merely accounting for where it went.

This is not a novel idea. Businesses have long managed budgets partly by outputs: sales departments are budgeted based on revenue targets, manufacturing plants based on units produced. Government’s adoption of performance budgeting represents an attempt to import market discipline into the public sector, at least in measurement.

How performance metrics are chosen

The critical weakness in performance budgeting lies in metric selection. Not everything that matters can be easily measured, and choosing the wrong metrics can distort behaviour catastrophically. A school might optimise for standardised test scores while neglecting critical thinking. A healthcare provider might maximise patient throughput while sacrificing quality of diagnosis. A police department might target visible crime statistics while ignoring organised crime that occurs away from public view.

Governments attempting performance budgeting therefore face a choice: measure what is easy or measure what is important. Most choose a mix, and the results are mixed. The UK’s Health Service reforms in the 2000s, for example, introduced waiting-time targets that were relatively easy to measure. Hospitals responded by meeting them—but sometimes by parking patients in recovery wards rather than discharging them, so the official wait time improved even as throughput declined. The metric was gamed because it was divorced from the underlying outcome (actual patient improvement).

Selecting metrics requires political judgment. Should a job-training programme be measured by the number of people trained, the number who find employment, the wage gains they achieve, or their long-term employment stability? Each is defensible; each will produce different budget allocations. The choice is not technical but ideological: what does the government believe the programme should achieve?

The risk of metric cascades

Once a metric is enshrined in a budget formula, organisations optimise for it relentlessly. If a public utility’s budget is determined by the percentage of customer complaints resolved, the utility will inevitably invest in how complaints are recorded and categorised rather than in the underlying service quality. Agencies learn to measure more strategically (some might say strategically measure more).

This creates a cascade of unintended consequences. A police force paid by arrest metrics will arrest more people, even if that does not reduce crime. An employment service paid by job-placement metrics will place people in short-term jobs that end after the metric-counting period, making the placement appear successful when it is merely temporary. An environmental regulator paid by number of inspections conducted will conduct many inspections but may issue fewer enforcement actions, since enforcement is costly and slow.

The best performance-based budgets account for this by tracking multiple metrics, including some leading and some lagging indicators. Some also build in qualitative review or randomised auditing to catch gaming. But this adds complexity and cost, reducing the efficiency gains that performance budgeting was meant to deliver.

Why politicians favour it (and why they should be cautious)

Performance-based budgeting appeals to elected officials because it offers a straightforward justification for budget discipline. If a programme’s performance drops, cutting its budget becomes a data-driven decision rather than a political preference. This is politically useful: the minister can point to the metrics and say “We did not cut this; the performance did.” It shifts accountability from political choice to mechanical outcome.

This appeal is powerful enough that many governments have adopted some form of performance budgeting, even when implementation remains partial or symbolic. Australia introduced performance-based elements in the 1980s and has expanded them steadily. The UK, Nordic countries, and New Zealand have all embedded performance metrics into budget-setting. The US, by contrast, has been more skeptical, with Congress preferring to make budget decisions on an ad hoc basis rather than automating them through formulas.

Yet the danger of over-relying on performance metrics is real. If budget reductions are determined mechanistically by metrics, elected officials can escape accountability for policy choices that are actually their own. A decision to accept longer waiting times or fewer inspections becomes a data point, not a deliberate trade-off. Democracy arguably requires that elected officials take responsibility for such choices explicitly, not delegate them to formulas.

Partial and hybrid implementations

Most countries have adopted performance budgeting partially, reserving it for some departments or programmes while leaving others on traditional input-based budgets. Mandatory government services (healthcare, policing, defence) are often partially performance-based, since measuring outcome is genuinely difficult. Transfer programmes (unemployment benefits, pension administration) are easier to measure and more frequently link budget to performance.

The Australian government uses a “Programme Budget” structure that incorporates both input (staff, equipment) and performance (outcome targets) elements. Departments must report not only how much they spent but what they achieved. Yet the budget allocation itself is not purely mechanical; it still involves ministerial discretion. This hybrid approach accepts that some human judgment is necessary even in a performance-oriented system.

The UK similarly uses performance targets alongside traditional departmental budgets, with most significant programmes having a “key performance indicator” (KPI). But these are monitored rather than deterministic—they inform political decisions rather than dictate them.

The evidence on effectiveness

Research on performance-based budgeting yields cautious results. When metrics are well-chosen and programmes are monitored carefully, performance-based approaches can improve efficiency and reduce waste. A study of Denmark’s adoption of performance budgeting in the 1990s found modest improvements in service timeliness and cost control. Australian agencies have reported improvements in efficiency and transparency.

However, studies also document gaming, metric misalignment, and unintended consequences when performance budgets are implemented carelessly. The risk is highest when a single metric dominates, when there is no auditing for gaming, or when the metric is easily manipulated but distorts the underlying outcome.

The consensus among public-finance researchers is that performance budgeting works best as one input to budget decisions, not as the sole basis. It forces explicit conversation about outcomes and helps identify persistently underperforming programmes. But it cannot replace political judgment about priorities and trade-offs.

The persistence of input-based budgeting

Despite decades of advocacy for performance budgeting, many governments still allocate funds primarily on input bases—staff allocations, historical spending, negotiated rates. This persistence suggests either that input-based budgeting is more robust than its critics allow, or that the transition to performance budgeting is politically harder than it appears.

Part of the explanation is path dependence: civil services have budgeting processes embedded in decades of practice. The staff structure and salaries are contractual obligations; changing them takes time. Part is that performance budgeting requires political consensus about what outcomes matter, and such consensus is rare. Part is that not all public goods have easily measurable outcomes, and trying to force measurement can do more harm than good.

See also