Government Unstuck

4. Measuring progress towards Government Unstuck

Abstract

This chapter considers how governments can assess whether reform efforts are producing the intended results. It examines the current landscape of internationally comparable governance data and identifies a “missing middle” in the evidence base: the intermediate outcomes connecting government reforms to their longer-term impacts. Using the four objectives of the direction of change for a Government Unstuck as its conceptual framework, the chapter proposes a phased measurement agenda for developing shared indicators with OECD member countries. These indicators would help governments understand where progress is occurring, where reforms are falling short and why.

4.1. Introduction

The preceding sections of this report set out four directions of change needed for governments to become unstuck, and the enabling capabilities on which that transformation depends. But ambition is not the same as progress. For reform to succeed, governments need to know whether their efforts are translating into the change they intend to produce, where they are falling short, and why. This requires evidence, and in particular evidence on the intermediate outcomes that sit between the reforms governments undertake and the long-term impacts they seek to achieve.

This section examines the current state of internationally comparable governance data, identifies the structural gap in the evidence base, what this report calls the "missing middle", and sets out how the four objectives of the direction of change for a Government Unstuck can provide the conceptual architecture for filling it. It then proposes a measurement agenda: a phased programme of work, developed in collaboration with Member countries, to build a shared set of intermediate outcome indicators that help governments monitor not only whether they are reforming, but whether their reforms are delivering results.

4.2. Measuring governments’ capacity to deliver: The state of play

Robust, internationally comparable data are essential to enable governments to monitor their performance, benchmark themselves against peers and demonstrate whether their action is supporting improvements for individuals and society as a whole, as well as strengthening the economy. In many sectoral policy domains, such as health, education, the environment, the labour market, the international statistical system already provides extensive and well-established outcome data, allowing countries to track whether people are healthier, better educated, more securely employed and living in a cleaner environment, and to compare those results across borders.

Yet, when it comes to the performance of government itself as an institution, the evidence base is far thinner. This is not to say that all governance evidence is missing. Internationally comparable data on the resources governments commit (inputs) and the processes they follow are reasonably comprehensive and well established—for example, every two years the OECD publishes Government at a Glance— while some measures of government outputs exist at the sectoral level. Recent OECD work has also made important progress at the other end of the chain: the OECD Trust Survey now provides nationally representative evidence across countries on the long-term impact of government performance on whether citizens trust public institutions, and on which specific dimensions of government competence and values are building or eroding that confidence. The OECD and other international institutions are also beginning to develop comparative measures of government productivity, capturing how efficiently public resources are converted into public value, thus offering new evidence on the impact of government activity on economic performance.

What remains largely unmeasured is the space between the two: comparable evidence on the intermediate outcomes that connect what governments do to what citizens, businesses and civil servants actually experience in their interactions with government remains scarce, fragmented and anchored in definitions that vary across countries (see Figure 4.1). Whether reform is actually making government more legible, more responsive, more fiscally sustainable and better equipped to anticipate change is, for the most part, not tracked in internationally comparable terms.

Figure 4.1. Theory of change of impact of government activity on change in society and the economy

Source: OECD authors’ elaboration.

This "missing middle" matters because without it, governments can observe their reform activity and monitor long-term impacts on citizens, society and the economy, but cannot understand and adjust how their efforts link one to the other. They are left unable to diagnose where reform is working, where it is stalling and why, and therefore unable to correct course. For governments seeking to get unstuck and unleash their capacity to deliver, this is the critical evidence gap.

This is in part because measuring how government performs as an institution is inherently harder than tracking outcomes in established policy domains, for several reasons. First, the methodological infrastructure underpinning governance statistics is still developing. Only a limited number of international statistical standards and technical guidelines on public governance exist, and the development across the different dimensions vary substantially (UN Praia City Group, 2020[1]). The absence of shared definitions of what should be measured, and how, means that governance statistics, unlike more established fields such as labour-market or national-accounts statistics, are not yet routinely or comprehensively produced at the national level by official statistical systems in most OECD countries (González, Fleischer and Mira d’Ercole, 2017[2]).

Most national-level and internationally comparable governance indicators are generated either from administrative records produced in the routine course of government operations, or from expert assessments provided by the OECD, and other researchers, practitioners or civil servants. Administrative data are particularly well suited to capturing governance inputs and processes on public revenues and expenditures, employment, procurement procedures, regulatory stocks, the existence and operation of integrity frameworks. Data from the 2026 OECD Government Unstuck Survey shows that an increasing number of countries draw on them to monitor the implementation of their public administration reform strategies. In turn, expert assessments are useful for capturing how government practices and procedures are implemented, and for synthesising more complex concepts that are difficult to measure directly, such as the rule of law, the quality of democracy or the integrity of public institutions. Both types of data are better suited to characterising the enabling conditions for government performance than to capturing what government reform actually delivers.

At a level above this, there are also multiple internationally comparable composite indices available to measure the performance of government as an institution. These composite indices provide valuable information for high-level benchmarking, helping to put governance quality on the international policy agenda, to compare broad patterns of performance across countries, and to flag where governance trajectories are improving or deteriorating over time. However, they provide limited disaggregated information to track the concrete effects of specific reforms (UN Praia City Group, 2020[1]) or to examine government delivery. In addition, most of these indices suffer from similar technical limitations, as they aggregate heterogeneous sources, as well as use definitions which sometimes vary across components and over time (UN Praia City Group, 2020[1]).

The OECD's own portfolio of governance data illustrates this missing middle. It is strong on the inputs and processes that go into government performance (in other words, the enabling conditions of government performance), increasingly rich on its long-term impacts, but sparse on the intermediate outcomes that connect the two.

The OECD’s portfolio of governance data, published in Government at a Glance, provides the most comprehensive international database tracking the inputs, institutional conditions and processes that underpin good public governance (covering regulatory quality, budgeting practices, human-resources management, digital government, public procurement, infrastructure governance, public-sector integrity, open government and citizen engagement). The portfolio has recently been further enlarged through the 2026 OECD Government Unstuck Survey (see Box 4.1). The OECD data on public governance is increasingly used by countries to benchmark their reform efforts.

Box 4.1. 2026 OECD Government Unstuck Survey: Improving government efficiency and effectiveness

The OECD Government Unstuck: Improving Government Efficiency and Effectiveness Survey was developed to generate new comparative evidence on governments’ efforts to improve the efficiency and effectiveness of public administrations. In particular, it aimed to address evidence gaps in areas such as simplification of internal government processes and organisational arrangements, and measurement of public sector productivity and efficiency. The survey also provides a comprehensive overview of current public administration reform efforts across countries.

It collects information on the administrative processes, governance arrangements and standards used to support public sector transformation and is structured around three pillars:

(i) public administration reform strategies

(ii) improving public administration efficiency

(iii) measuring public sector efficiency and productivity.

The survey was administered between February and June 2026 and was distributed to all 38 OECD Member countries and the 8 OECD accession candidate countries. Responses were co-ordinated by OECD Public Governance Committee delegates or designated officials from the relevant national institutions, primarily ministries and government bodies responsible for public administration, civil service management and public governance. Countries were asked to report on policies, practices and institutional arrangements in place at the central/federal government level as of 1 January 2026 and to submit a single consolidated response reflecting a whole-of-government perspective. Following data collection, the OECD conducted validation and verification rounds with respondents to ensure the accuracy and comparability of the information provided.

A total of 42 country responses were received, comprising 35 OECD Member countries and 7 accession candidate countries. Data for Brazil, Denmark, Germany, Lithuania and the European Union are not available.

To complement this data with evidence on government performance impacts, the OECD has invested in measuring levels of trust in public institutions. This work has been groundbreaking in proposing that trust is a major impact of government activity, of equal importance to classic concepts like rule of law and political stability. Methodologically, this work also moves beyond traditional indices of governance impacts. The data to measure this impact is collected directly from citizens and is based on an underlying conceptual framework of the drivers of trust (OECD, 2017[3]). The OECD Survey on Drivers of Trust in Public Institutions, known as the OECD Trust Survey, collects nationally representative evidence from respondents across countries and disentangles aggregate trust levels into the specific drivers through which government builds or erodes public confidence. By capturing how government performance is perceived and experienced by citizens, the Trust Survey provides a critical new resource for monitoring the longer-term impacts of government action and thus for informing the Government Unstuck agenda.

Alongside trust, the OECD is developing new work on government productivity, which aims to measure how efficiently public resources are converted into public value and how government activity contributes to economic performance. This responds to long-standing calls in the public administration literature for more robust measurement of public-sector efficiency alongside legitimacy, and builds on pioneering national efforts, most notably the UK Office for National Statistics' implementation of the Atkinson Principles and the 2024-25 National Statistician's Independent Review of the Measurement of Public Services Productivity, which strengthened the measurement of public sector productivity as a whole (UK Office of National Statistics, 2025[4]). Together with the Trust Survey, this stream will complement citizen-experience evidence with objective, output-based measures of how government activity translates into societal and economic results.

Taken together, the international evidence base offers reasonably comprehensive measurement at both ends of the governance chain: on one side, the enabling conditions for government performance (the inputs, institutional arrangements and processes that allow governments to operate); on the other, the long-term impacts that governments ultimately seek to influence, such as citizen trust and economic performance.

What is missing is the middle: a concept and measurement framework for the intermediate outcomes that connect the two, capturing what governance reforms actually deliver and how that delivery is experienced by citizens, businesses and civil servants. This is not merely a data gap. It is a structural weakness in how governance reform is tracked and managed. Without comparable evidence on intermediate outcomes, governments can demonstrate reform activity without demonstrating reform results, and accountability defaults to process compliance rather than delivery of change. The consequences are visible: the indicators that governments themselves select to monitor reform are overwhelmingly focused on inputs, processes and outputs on one side, and long-term impacts on the other, rather than on the effects that reforms are intended to produce. Public administrations can record steady reform progress while citizens report little change in their experience of government (OECD, 2024[5]), and whether this gap reflects a genuine failure to deliver or simply a lag between improvement and public perception, governments currently lack the intermediate outcome data to tell the difference.

Building internationally comparable evidence on the intermediate outcomes of government action, anchored in shared definitions and robust sources, is therefore essential to enable countries to learn from one another, as well as to anchor the Government Unstuck agenda in a shared and verifiable language of progress. In this sense, understanding why public administrations and the public diverge in their reading of government performance is not a precondition for getting governments unstuck — it is part of the agenda itself.

4.3. Identifying the missing middle

The four objectives of the direction of change of Government Unstuck can collectively provide a framework for linking governance enabling conditions (inputs, institutions, processes) to societal and economic outcomes. These objectives – a renewed relationship between government and citizens; significantly simplified government and regulation; restored room for manoeuvre in public finances; and a shift from prepared to proactive government – are not a blueprint for a perfect government or an ideal to be scored against. They are the areas where governments need to focus their efforts in order to deliver the change that citizens expect and that society and the economy require.

Each objective describes both the symptoms of a "stuck" government and the characteristics of an unstuck one, from the perspective of those who experience and depend on government performance: stakeholders wanting to have a voice in government decision making, users of public services, businesses seeking to navigate regulatory environments, civil servants and elected officials trying to act on reform, and taxpayers who want to understand where their tax money goes. Embedded in each objective are therefore implicit intermediate outcome objectives that, once made explicit and translated into measurable indicators, would articulate the goals through which government inputs and enabling conditions are translated into the long-term impacts that matter: trust and prosperity. As such, the Government Unstuck pillars provide a strong conceptual architecture for the "missing middle".

Translating this conceptual architecture into a measurement framework will require, first, identifying for each objective what governments need to deliver in order to become unstuck. It will then require developing and combining several complementary types of evidence to track progress against those objectives, rather than relying on any single source (see Annex Table 4.A.1 for an example of intermediate outcome objectives and sample corresponding indicators for each direction of change of Government Unstuck). The intermediate outcomes that the four objectives describe are themselves heterogeneous: some are best captured through the lived experience of those who interact with government, others through factual administrative records of what government delivers, others through observed behaviours and revealed preferences of citizens, businesses and civil servants, and others again through structural and macroeconomic data on the conditions that government action produces. Different, complementary types of indicators could therefore be included:

• Experiential and perception-based measures, grounded in the perspectives of those most directly affected by governance performance, citizens as voters and taxpayers, service users, businesses and civil servants, are particularly relevant for identifying how to improve the quality of the interface between government and society: for example, how legible and navigable government information appears to a user seeking to access a public service, how responsive and predictable regulatory interactions feel to a business seeking a permit, how aware taxpayers are of fiscal trade-offs available to governments, or how equipped and empowered civil servants report feeling to propose and implement reform.

• Administrative data on how government actually delivers are helpful for identifying how to improve the scale, speed or efficiency of delivery. They can include the uptake of public services, processing times for permits and licences, civil-service mobility, or the share of public consultations followed by published government responses, which anchor assessments of intermediate outcomes in observable government activity.

• Behavioural indicators capture what citizens, businesses and civil servants actually do in their interactions with government, as distinct from what they report or what administrative systems record. Examples include the uptake of digital public services when alternatives are available, the rate at which businesses choose to invest in jurisdictions after regulatory simplification, civil-servant turnover and mobility between agencies, or the frequency with which citizens voluntarily engage in participatory processes. Behavioural indicators are particularly valuable because they reveal preferences through action rather than self-reporting, providing a check on whether perception-based improvements correspond to real changes in how people interact with government.

• These, in turn, are complemented by structural and macroeconomic data on the conditions that government action produces, such as compliance with pre-set fiscal targets, budget execution rates, the rigidity of public expenditure, or the alignment of public investment with stated priorities, which are particularly relevant for objective 3 on fiscal sustainability.

The choice to draw on this full spectrum of evidence is deliberate. While perception-based measures are central to capturing how government action is experienced, they are neither sufficient on their own nor universally appropriate: perception data alone risk capturing reputation rather than performance, can lag actual change, can be affected by a range of factors other than actual government performance, and provide limited insight into structural outcomes such as fiscal sustainability; administrative data alone can record activity that bears little relation to the experience of those affected; behavioural indicators, while powerful as a check on self-reported perceptions, can be influenced by factors outside government control and require careful interpretation; and macroeconomic indicators, while essential for fiscal and economic dimensions, do not capture the everyday quality of the interface between state and society. Beyond these conceptual limitations, each data type also presents practical challenges: administrative data are rarely standardised across countries and often require significant harmonisation work before they can support international comparison; survey-based measures, whether of perceptions or behaviours, impose a collection burden on respondents and on the institutions that field them; and linking behavioural data across administrative systems raises questions of data governance and privacy that countries handle differently. Building a measurement framework for the missing middle will therefore need to be realistic about what can be collected, at what cost and at what frequency, and to sequence its ambitions accordingly.

Despite these challenges, the case for building this evidence base is strong. What unites this spectrum of indicators is their shared function as intermediate-level evidence of progress towards Government Unstuck. This shift from process compliance to experience and intermediate outcome mirrors broader developments in public-sector measurement methodologies, including the adoption of life-event-based service assessment frameworks and co-produced performance indicators, and reflects a wider recognition that governance quality is ultimately determined at the interface between state and society. Over the past five years, the OECD has already begun building elements of this evidence base:

• Through the OECD Drivers of Trust Survey, it collects nationally representative, perception-based data on how citizens view government competence and values.

• Through its Workforce Insights from Central Governments survey, it gathered first-hand evidence on how civil servants themselves assess their working environment, their capacity to act on reform, and the internal obstacles they face.

• Through its recent Simplifying for Success initiative, it surveyed business associations to capture how businesses experience regulatory interactions, administrative burden and the legibility of government requirements.

Taken together, these three instruments already span several of the perspectives (citizens, civil servants, businesses) from which the intermediate outcomes of the Government Unstuck framework need to be assessed, and they provide a foundation on which the measurement agenda set out below can build.

4.4. Government Unstuck: A measurement agenda

As part of the Government Unstuck report, the OECD introduces an initial measurement framework for the "missing middle" connecting the enabling conditions that governments put in place (institutions, processes, resources) and the long-term impacts they seek to achieve (trust, productivity, public value). For each of the four objectives of the direction of change of Government Unstuck, it identifies a select number of the intermediate outcomes that reform should produce and sets out the direction in which those outcomes are expected to move when the enabling conditions are working as intended.

This framework is thus not a composite index designed to rank countries or define what "good governance" looks like. It is a set of intermediate outcome indicators intended to help countries monitor whether they are focusing on the right things and moving in the right direction to become "unstuck",that is, to deliver the change that citizens expect, as well as that society and the economy need. Some of the data that feed these indicators already exist in internationally comparable form; others will need to be collected, whether by the OECD directly or by national authorities working within shared definitions and standards; and for some intermediate outcomes, well-chosen proxies will be needed where direct measurement is not feasible. The framework's value lies not in producing a single score but in providing a common diagnostic language, anchored in a clear theory of change, through which countries can assess their own progress, identify where reform is and is not translating into results, and learn from the experience of peers.

Translating this framework into a functioning set of internationally comparable indicators will require a phased programme of work, developed in close collaboration with OECD Member countries, through the working parties and committees:

• The first step is to reach agreement on what to measure: to validate and, where necessary, refine the intermediate outcome objectives identified across the four objectives of Government Unstuck, and to confirm that they capture the dimensions of government performance that countries themselves recognise as central to becoming "unstuck".

• The second step is to map the existing indicator landscape against those agreed objectives, identifying where internationally comparable data are already available, whether from OECD sources, national statistical systems or other international providers, and where gaps remain.

• The third step is to fill those gaps by developing harmonised definitions and robust collection methodologies for new measures, focusing on where collection is practically feasible and where new indicators would add the most diagnostic value relative to what already exists. For some intermediate outcomes, direct measurement may not be possible or cost-effective, and well-chosen proxy indicators will need to be used instead.

This is a long term endeavour, that will require investment by working parties and committees. The OECD's role in this third step may vary depending on the nature of each indicator. In some cases, the OECD may be best placed to collect data directly, building on its existing survey infrastructure or through new purpose-built instruments. In other cases, the OECD's contribution will be to convene countries around the development of joint methodologies and shared definitions, so that they can collect the data through their own national systems in a way that ensures cross-country comparability. The balance between these approaches will need to be determined objective by objective, indicator by indicator, in dialogue with countries, as the work progresses.

Annex 4.A. Example outcome objectives and indicators for Government Unstuck

Annex Table 4.A.1. Examples of intermediate outcome objectives and corresponding indicators for measuring Government Unstuck

Direction of change	Characteristics of an Unstuck Government	Underlying intermediate outcome objective	Example indicators	Type of indicator
OBJECTIVE 1 A renewed relationship between government and citizens Government listens, communicates and engages in ways that strengthen legitimacy and voice	Public services that satisfy user needs	Citizens and businesses experience public services as accessible, timely and responsive to their needs	[Perception] Share of users who report being satisfied with the quality of public services they used in the past 12 months [Perception] Share of users who found it easy to access the public service they needed [Administrative] Average time between initial request and service delivery for key life-event services (e.g. birth registration, permit issuance)	[Perception] [Administrative]
	A government that listens and responds to society	Government uses citizen insight, feedback and real-time evidence to shape its decisions and communication	[Perception] Share of citizens who believe people like them have a say in what government does [Perception] Share of citizens who believe government communicates openly about the reasons behind its decisions [Administrative] Share of public consultations for which a government response was published and communicated	[Perception] [Administrative]
	Meaningful citizen voice in public decisions	Citizen input visibly shapes outcomes; engagement goes beyond procedural compliance	[Perception] Share of citizens who are confident in their ability to participate in politics [Behavioural] Number of national-level deliberative processes whose recommendations were formally considered by government [Administrative] Share of regulatory proposals for which stakeholder contributions were received and publicly addressed	[Perception] [Behavioural] [Administrative]
OBJECTIVE 2 Significantly simplified government and regulations Rules, structures and services are understandable, coherent and responsive to changing needs	A legible government	Citizens and businesses can navigate government without decoding its internal structure	[Perception] Share of users who find it easy to find and understand the information needed to access a public service [Perception] Share of businesses that consider regulatory requirements clear and predictable [Administrative] Average number of steps and contacts required to complete a life-event transaction (e.g. starting a business, registering a birth)	[Perception] [Administrative]
	An agile government	Government adapts rules, services and structures to changing evidence and user needs	[Perception] Share of users who believe government would improve a poorly performing service if many people complained [Behavioural] Change over time in the average cost to businesses, measured in time and money, of complying with a set of key regulatory obligations	[Perception] [Behavioural]
	A coherent government	Government institutions work together; authority is aligned with responsibility	[Perception] Share of civil servants who report that inter-agency co-ordination works well on cross-cutting priorities [Administrative] Share of permits requiring interaction with more than one authority that are handled through a single window [Administrative] Number of public bodies subject to a periodic mandate review or institutional sunset clause	[Perception] [Administrative]
OBJECTIVE 3 Restoring governments' room for manoeuvre in public finances Public finances are sustainable, strategically allocated and democratically anchored	Sustainable public finances	Governments can have the fiscal space and tools to allocate funds to deliver on their priorities	[Macro/structural] Structural primary balance as a share of GDP and trajectory over the medium term [Administrative] Share of total expenditure classified as non-discretionary (mandatory entitlements, debt service, wage bill) [Administrative] Share of expenditure identified for reallocation, reduction or discontinuation through spending reviews that was effectively acted upon in subsequent budgets	[Macro/structural] [Administrative]
	Budgets that finance the future	Public resources are directed to where they generate the greatest public value; investment is aligned with stated priorities	[Macro/structural] Public investment as a share of total government expenditure and change over time [Administrative] Budget execution rate: actual expenditure vs. appropriations for priority spending categories [Administrative] Share of public procurement conducted through strategic frameworks (e.g. green, innovation-oriented, SME-targeted)	[Macro/structural] [Administrative]
	A transparent and accountable fiscal contract with citizens	Citizens understand where their tax money goes and trust that it is spent for the public good	[Perception] Share of respondents who believe tax revenues are spent for the public good [Perception] Share of taxpayers who report understanding how public money is spent [Administrative] Publication and scope of tax-expenditure reporting (number of tax expenditures quantified and published)	[Perception] [Administrative]
OBJECTIVE 4 From prepared to proactive government Government anticipates, absorbs shocks, learns from experience and acts before needs are expressed	A government that anticipates emerging risks	Foresight, anticipation and preparedness are embedded in planning, budgeting and legislative processes as a routine capability	[Administrative] Share of major policy proposals accompanied by a published long-term risk or foresight assessment [Perception] Share of civil servants who report that their organisation uses evidence from foresight exercises to inform decisions [Behavioural] Number of cross-government scenario exercises conducted and connected to budget or legislative cycles per year	[Administrative] [Perception] [Behavioural]
	A government that learns systematically from experience	Lessons from past successes and failures are captured, shared and used to improve future action	[Administrative] Share of programmes and regulations subject to mandatory post-implementation review within a defined period [Perception] Share of civil servants who report that lessons from past failures are used to improve current practices [Behavioural] Number of policy areas with published evaluations whose findings were formally considered in subsequent policy design	[Administrative] [Perception] [Behavioural]
	Responsible and trusted use of AI in government	Citizens are confident that government uses AI fairly; accountability and human oversight are embedded in deployment	[Perception] Share of citizens who are confident that government uses AI responsibly and fairly [Administrative] Share of high-risk public-sector AI applications subject to a published impact assessment and human-oversight protocol [Perception] Share of civil servants who report having the digital skills and authority to use new technologies effectively	[Perception] [Administrative]

Source: OECD authors’ elaboration.

References

[2] González, S., L. Fleischer and M. Mira d’Ercole (2017), “Governance statistics in OECD countries and beyond: What exists, and what would be required to assess their quality?”, OECD Statistics Working Papers, No. 2017/3, OECD Publishing, Paris, https://doi.org/10.1787/c0d45b5e-en.

[5] OECD (2024), OECD Survey on Drivers of Trust in Public Institutions – 2024 Results: Building Trust in a Complex Policy Environment, OECD Publishing, Paris, https://doi.org/10.1787/9a20554b-en.

[3] OECD (2017), OECD Guidelines on Measuring Trust, OECD Publishing, Paris, https://doi.org/10.1787/9789264278219-en.

[4] UK Office of National Statistics (2025), National Statistician’s independent review of the Measurement of Public Services Productivity, https://uksa.statisticsauthority.gov.uk/wp-content/uploads/2025/03/National-Statisticians-Independent-Review-of-the-Measurement-of-Public-Services-Productivity.pdf.

[1] UN Praia City Group (2020), Handbook on Governance Statistics, https://paris21.org/sites/default/files/2021-12/PRAIA%20Handbook%20final%20WEB-REVISED2021.pdf.