Policymakers and governments are facing daunting challenges such as climate change, aging demographics, income and wealth inequalities, low productivity, unstable geopolitics, and international conflicts. For some, this means governments need to develop new approaches and capacities to meet these challenges. For others, the solution is to limit the size of government by focusing on key priorities and reallocating resources.

Both groups should agree that measuring and improving government efficiency and effectiveness is a smart way to boost state capacity. Without accurate and ongoing measurement of government performance, it is not possible to know how well governments work or whether initiatives to make them work better are successful.

Assessing government results requires measuring how efficiently programs are delivered and how effectively they achieve their objectives. Efficiency improves when programs are delivered at lower cost or with fewer inputs over time through productivity enhancements such as technology investments, alternative delivery channels or reorganization. Using scarce public funds more efficiently frees up financial resources for other priorities.

Effectiveness measures the extent to which programs achieve their expected outcomes and provide value for money — that is, whether they provide benefits that are greater than their costs. Measuring effectiveness allows governments to make better-informed decisions by identifying which programs achieve desired outcomes most effectively. It also helps governments better understand the trade-offs involved when benefits are difficult to measure, such as when pursuing equity or meeting national defence imperatives.

To examine productivity in Canada’s federal public service and inform the government’s economic plan, the President of the Treasury Board appointed the Working Group on Public Service Productivity in fall 2024. In 2025, the working group (of which I was a member) provided 19 recommendations to measure and improve the productivity and effectiveness of the public service.

Implementing these recommendations would help the federal government both measure and improve public service efficiency and effectiveness.

Measuring Government Efficiency

If a government wants to improve public-sector productivity, the first thing it needs to do is measure it. This is easier said than done. Since governments mostly produce non-market outputs (i.e., they do not sell their services at market prices), measuring output indicators for public services is challenging.

Until recently, Statistics Canada has followed international standards that allow for two ways of measuring government output: the direct inputs approach, which treats the resources used to produce public services as a measure of output; and the direct output approach, which measures what the public sector actually produces.

While the output approach is recommended, the input approach is considered the fallback when measuring output is too challenging (e.g., defence spending). What is unique about Canada is that it uses the input approach across all government sectors, whereas many countries have used the output approach for sectors such as health and education for some time.

This makes for an unsatisfactory and often misunderstood measure of productivity for the public sector. In a review of Statistics Canada’s measure of public-sector productivity, Phillip Smith, a former assistant chief statistician at Statistics Canada, concluded that the measure was misleading and should no longer be published. In December 2025,  Statistics Canada terminated its national public-sector productivity series “due to the difficulty of adequately measuring production” in that sector and began to publish labour productivity indicators only for the business sector.

An alternative would be for Statistics Canada (with input from departments and agencies) to follow the example of the United Kingdom and the European Union and develop and test a direct output approach to estimate public service productivity. As with the business sector, the output of many public services can be measured using physical counts or quantities of services provided to Canadians, such as patients treated, students enrolled, inspections conducted or passports processed.

While the output of many key activities conducted by the federal government would inherently be difficult, if not impossible, to measure, programs delivering direct benefits to Canadians would be strong candidates for testing an output-based approach to measuring public-sector productivity.

The federal government should require departments and agencies that provide services directly to Canadians to develop and report on productivity metrics for those services. These metrics should be standardized by type of service to facilitate comparisons across programs and reported alongside how often service standards are respected. (Service standards are commitments made by governments on the level of service Canadians can reasonably expect, such as processing times for passport applications).

Productivity indicators and service standards should be consolidated on a single government website to facilitate public access. This information would not only help Canadians and government officials assess performance but also provide Statistics Canada with the data needed to develop output-based productivity measures for the federal public service.

The federal government’s Expenditure Management System is designed to ensure that all existing and new programs are focused on results. It requires public servants to identify, measure and report on an array of performance indicators and targets. The system is built  on three pillars: the Policy on Results, which provides the framework for measuring and evaluating program performance;  Departmental Plans, which set out targets; and Departmental Results Reports, which report on results achieved.

The system should not be replaced, but there is an opportunity to simplify and improve it. Replacing a plethora of program-level performance indicators, objectives and self-determined targets with a unified framework for measuring efficiency and effectiveness would reduce the reporting burden on departments and be more useful for decision-makers.

Improving Public-Sector Productivity: The AI Example

The federal government can start making changes now, without waiting for an overarching measurement framework. The Working Group on Public Service Productivity made many recommendations for the federal government on how it could boost productivity, including accelerating investments in artificial intelligence (AI) technologies.

In its response to the Working Group on Public Service Productivity, the federal government indicated that improving public-sector productivity and measuring the impact of its actions are priorities. It went on to emphasize that it intends to accelerate the adoption of AI to enhance public-sector productivity and develop a “consistent approach to measuring the productivity impacts of AI use cases across the Government of Canada”.

The Treasury Board Secretariat recently published the federal government’s first public AI register, providing information on where and how AI is being used across the public service. The register includes over 400 projects from 42 federal organizations, including most departments, where a range of AI tools are being developed or deployed. The mere fact that so many AI projects are being developed and implemented within the federal government is impressive.

Less impressive is the absence of quantitative estimates of the expected or realized net gains from any of the projects, such as reductions in operating costs net of investment costs or specific improvements in services provided to Canadians. This oversight suggests either a lack of capacity or, possibly, a lack of willingness within the federal public sector to adopt rigorous measurement.

Assessing Value for Money in Government Programs and Activities

While measuring public-sector productivity is important for determining whether its capacity is growing or shrinking, there is a question that is even more important to consider: are government programs and activities providing a net benefit to society?

To answer this question, we need to consider the costs and benefits provided by a program. For example, a training program would pass a cost-benefit test if the benefits to participants — such as increased employability and earnings — exceed the program’s costs, including the cost of raising taxes to finance it.

As mentioned above, this cost-benefit assessment is, in principle, handled by the federal government’s Expenditure Management System. Under this framework, program performance is assessed in terms of relevance, efficiency — which encompasses both operational efficiency (e.g., productivity) and allocative efficiency (e.g., the net benefit experienced by program participants or beneficiaries) — and effectiveness, which concerns whether programs achieved their intended outcomes.

Deputy ministers are responsible for establishing, implementing and maintaining departmental results frameworks. Within departments, the head of performance measurement and the head of evaluation, together with program managers, are responsible for ensuring the appropriate data collection and other systems are in place.

In theory, the Expenditure Management System should provide the federal government with the information needed both to measure public-sector productivity and to determine whether programs provide value for money through cost-benefit analysis.

Unfortunately, a recent C. D. Howe Institute report indicates that when evaluations are performed, they rarely assess productivity or value for money. Instead, they tend to focus on implementation issues, qualitative assessments (e.g., recipient satisfaction) or how program beneficiaries are affected, without assessing whether the outcomes can be attributed to the intervention itself. There are, however, a few federal programs that have been evaluated with state-of-the-art cost-benefit analyses. For example, the federal-provincial Labour Market Development Agreements and Pathways to Education program invested in systems that linked administrative and economic data. The federal government should build on these exemplary evaluations.

But these evaluations are the exceptions rather than the rule. Excessive flexibility in the federal evaluation framework makes it difficult to produce useful assessments of productivity and value for money.  Departments have wide latitude in how they assess these core economic concepts, including through approaches that are only loosely connected to them. As a result, evaluations often fail to report on these basic performance indicators that would allow decision-makers to compare programs and assess whether they deliver value for money at the lowest possible cost.

Given the diversity of government programming, evaluation approaches should retain some flexibility. However, that flexibility should apply only once departments have met a set of mandatory requirements. A recent C.D. Howe Institute conference report suggests the evaluation framework should be revised to require departments to assess programs’ value for money, alongside relevance, efficiency and effectiveness, using clearly specified methodologies. Standard efficiency metrics, such as output-to-input and unit-cost ratios, should be mandatory, as should appropriate value-for-money tools, such as cost-benefit analysis.

Strengthening evaluation practices would better equip ministers and other decision-makers to make difficult choices about whether to create, maintain, scale up, scale down or eliminate programs. The objective is not simply to identify and cut “wasteful” spending, but to give decision-makers comparable information about the costs and results of different programs when allocating scarce public resources.

Some may argue that it is difficult to apply cost-benefit analyses across all federal programs and that they may not be feasible for initiatives such as social or redistributive programs. We think the evidence suggests otherwise. For example, the U.S.-based Policy Impacts initiative created a library with over 200 programs, including social and income-transfer programs, with unified cost-benefit analyses.

The federal government’s experience mandating and applying cost-benefit analyses across departments under the Cabinet Directive on Regulation demonstrates what can be achieved when a framework clearly requires a common analytical approach. Under the directive, departments and agencies developing or amending regulations must provide a regulatory impact analysis that identifies and assesses the costs and benefits of the proposed regulation. This approach has  been successfully applied to regulations governing activities that, at first glance, might not seem well suited to cost-benefit analysis, including energy efficiency, food nutrition labelling and foreign influence transparency and accountability. This suggests that a mandatory requirement to assess costs and benefits need not prevent departments from applying judgment or adapting their analysis to the characteristics of individual programs.

This is not to suggest that cost-benefit analysis will always provide sufficient information for decision-makers to assess a program. Programs that pursue fairness objectives — such as income redistribution, reconciliation or strengthening social trust — involve outcomes that are inherently difficult to quantify. Evaluations should therefore complement the value-for-money perspective provided by cost-benefit analysis with a thorough assessment of a program’s broader impacts, including its effects on fairness and equity. Blending quantitative and qualitative analysis would give policymakers a more complete basis for weighing trade-offs between economic performance, fairness and other policy objectives.

Where the Federal Government Stands

Is the federal government interested in measuring and enhancing public service efficiency and effectiveness? The evidence suggests that it genuinely is. What is less clear is whether the public service can walk the walk.

In the 2025 Speech from the Throne and the prime minister’s May 2025 mandate letter, the government recognized the importance of improving the productivity of the federal public service to achieve better results for Canadians. In its response to the Working Group on Public Service Productivity’s recommendations on measuring and enhancing productivity, the government indicated that most of the recommendations either align with existing government actions, including those outlined in its 2025 budget, or could inform future work by the Cabinet Committee on Government Transformation.

One notable exception is the government’s decision not to pursue the working group’s recommendation to measure productivity across Canada’s public sector. This is unfortunate and a missed opportunity but should not prevent Statistics Canada from continuing its exploratory work on measuring output in the education and health sectors.

But does the public sector even have the capacity to improve how it measures productivity? Evidence from program evaluations and the AI register suggests that the public sector lacks capacity and, possibly, willingness, to establish a rigorous measurement framework. If the government is serious about its commitment to measuring and improving public-sector productivity, it will need to provide clear direction and invest in the capacity required to deliver it. In addition to rewriting the evaluation framework, as discussed above, and investing in data systems and people skills through training and recruitment, the government could take some relatively simple steps. For example, it could return to departments and agencies included in the AI register and require them to provide an assessment of the expected or realized returns on their AI projects.

What about ensuring that programs focus on results and provide value for taxpayers’ money? The public service does not have a good record of conducting cost-benefit analyses of government programs, and the evaluation policy framework needs to be updated to clarify what exactly is expected from program evaluation. The federal government should update the framework to make systematic assessments of programs’ value for money through rigorous cost-benefit analysis a mandatory part of program evaluation. This would be the most important long-term contribution the federal government could make to achieving better results for Canadians.

Conclusion

The troubled times in which we live may require governments to increase their capacity to deliver on new and ambitious policies and missions. But any added capacity must be effective and efficient. Likewise, current programs and initiatives should also be able to deliver results.

In both cases, the federal government will need a measurement framework that systematically tracks the efficiency and effectiveness of government programs and activities. Such a framework would ensure appropriate results are achieved for current programs and activities, as well as new initiatives. The federal government appears genuinely interested in improving public-sector efficiency and effectiveness.  But without the right framework to achieve its mission, it risks overestimating its capacity to deliver for Canadians.