Few political messages travel further than a promise to clean up government. Candidates who campaign on transparency, accountability, and a break from the past often win support precisely because voters want the distance between public office and private interest closed. Yet once in government, the same commitments can prove difficult to translate into durable change.
That gap is not automatically evidence of bad faith. It is, in large part, a story about institutions: what happens when a political promise meets the incentives, capacities, and constraints of governing. Researchers who study governance tend to describe it less as a personal failing than as a structural one – and understanding why helps explain both the difficulty and what can make reform last.
Why clean-government promises are politically attractive
Anti-corruption messaging tends to be popular across very different electorates. Perceptions of corruption are consistently associated with lower trust in public institutions, which is one reason the theme crosses ideological lines. For opposition movements and political outsiders, a commitment to transparency can serve as a unifying idea when agreement on other policies is thin.
The incentive to promise reform is therefore strong. The incentive to implement every element of it once in office can be weaker, because implementation touches the same systems a government relies on to function – budgets, appointments, procurement, and the informal relationships that hold coalitions together.
From promise to institution: the implementation gap
Most modern anti-corruption efforts are institutional. Since the United Nations Convention against Corruption (UNCAC) was adopted in 2003 and entered into force in 2005, it has become one of the most widely ratified treaties in the world. According to the UN Office on Drugs and Crime’s ratification status page, the convention had 192 parties as of September 2025.
The convention asks states to maintain preventive anti-corruption bodies with the independence needed to carry out their work. In 2012, practitioners and experts meeting in Jakarta set out the Jakarta Principles, a non-binding statement of what that independence generally requires: a clear legal mandate, secure appointments, protection from removal for political reasons, and adequate, reliable resources. The principles also note that independence does not mean isolation, and that agencies are expected to work with courts, legislatures, the private sector, and civil society.
The distance between adopting those standards and realising them is the subject of a substantial body of research. A recurring finding is that creating an anti-corruption body is the straightforward part. Whether it can investigate powerful actors, retain trained staff, and absorb budget pressure is what tends to determine its practical effect.
Corruption as a collective action problem
One influential explanation for reform failure comes from a 2012 study published in the journal Governance, based on interview research in Kenya and Uganda. Its central argument is that anti-corruption reforms are often designed around a principal–agent model: the idea that a well-informed principal – the public, or a reforming leader – can discipline a limited number of corrupt agents through monitoring and punishment.
Where corruption is systemic, the study argues, that model can misread the problem. If informal payments and patronage are widespread and widely expected, no single actor can unilaterally stop participating without bearing a cost. Corruption then behaves less like a few individuals breaking the rules and more like a collective action problem, in which the expectations everyone follows are themselves part of the difficulty. Reforms that assume a ready supply of honest, capable “principals” can struggle to gain traction.
| Dimension | Principal–agent framing | Collective action framing |
|---|---|---|
| Core assumption | A limited number of agents deviate from the rules; the principal is understood to be honest | Many actors are locked into a shared pattern of expectations |
| Typical remedy | Monitoring, sanctions, and enforcement aimed at individuals | Coordination, credible signals, and changes to shared expectations |
| Main limitation | Depends on honest, capable, well-resourced principals | Hard to shift without broad buy-in across institutions |
Source: summary of the argument in a 2012 study published in the journal Governance. The two framings are analytical models and are often used together rather than as strict alternatives.
Incentives, credible commitment, and why promises can stall
Political incentives matter as much as institutional design. Credible commitment theory, borrowed from economics, suggests that actors may publicly support a policy while retaining the practical ability to intervene in it later. Where oversight bodies are created by ordinary legislation rather than constitutional provisions, their powers can be narrowed, or their leadership replaced, through routine political and legal processes – sometimes framed as administrative reform.
This dynamic is best studied at the level of systems rather than personalities. The point is not that any particular government acts in bad faith. It is that the incentives facing a governing coalition can pull against the very oversight it has promised, because governing often depends on informal networks, allied institutions, and established administrative practices.
Entrenched interests and the limits of formal rules
Reforms redistribute something: discretion, access, and influence. That gives actors who benefit from existing arrangements reasons to support reform in principle while resisting its specifics. Research on policy implementation describes this as resistance that rarely appears as open opposition. More often it shows up as delay, narrow mandates, underfunding, or uneven enforcement.
Design alone does not determine outcomes. Anti-corruption bodies operate inside a wider political system, and their room to act can depend on how courts, legislatures, audit offices, and the civil service function. The relationship between a government and its anti-corruption bodies also varies considerably from one country to the next, and international news organisations provide further reporting on how those relationships evolve in particular national contexts.
Why measuring progress is genuinely difficult
Knowing whether anti-corruption reform is working is harder than it sounds. The most widely used cross-country measure, the World Bank’s Worldwide Governance Indicators, does not count corruption directly. Its “Control of Corruption” dimension captures perceptions, drawn from 35 data sources that include household surveys, firm surveys, and expert assessments, covering more than 200 economies from 1996 to 2024.
Because perception-based measures carry uncertainty, the WGI publishes margins of error alongside its estimates and cautions against treating small year-to-year movements as decisive. A score can shift because something real changed, or because the mix of underlying sources changed. For reformers, that creates a paradox: the indicators used to judge progress are noisy, while the political cost of appearing to fail is immediate.
What tends to make anti-corruption institutions more resilient
No single design guarantees results, but research and international guidance converge on a handful of features that appear to help. These include a legal mandate that is difficult to change on short notice, transparent appointment and removal procedures for leadership, budgets that are not easily redirected, internal integrity systems, and external oversight by bodies independent of the executive.
Institutional structure involves trade-offs. A dedicated anti-corruption agency can develop focused expertise and clear accountability, but it can also become a single point of failure if it is weakened. Specialised units embedded in larger police or prosecution services may be more resilient to concentrated pressure, though their resources can be redirected toward other priorities. There is no global consensus on which arrangement performs best, and outcomes appear to depend heavily on context.
International frameworks treat reform as a long-term process rather than a single announcement. Peer review, periodic assessment, and sustained engagement with civil society are typically presented as ways to keep commitments on the agenda after the initial political moment has passed.
Frequently asked questions
Do anti-corruption agencies reduce corruption on their own?
Generally, not by themselves. Studies and international guidance associate their effectiveness with independence, stable funding, investigative capacity, and support from the broader justice system. An agency with a strong mandate but limited resources or political insulation often struggles to produce results.
Why do anti-corruption reforms often fail?
Common explanations include design that does not match the scale of the problem, weak implementation capacity, conflicting political incentives, resistance from actors who benefit from existing arrangements, and the difficulty of changing expectations when corruption is systemic.
What is the principal–agent model of corruption?
It treats corruption as the behaviour of a limited number of agents who deviate from the rules while an honest, well-informed principal monitors and sanctions them. It is a useful model in many settings but is widely discussed as incomplete where corruption is systemic.
What are the Jakarta Principles?
A non-binding set of principles for anti-corruption agencies, developed in 2012 at a conference convened with support from the United Nations Development Programme and the UN Office on Drugs and Crime. They cover mandate, permanence, appointments, removal, resources, accountability, and public engagement.
How is corruption measured?
Cross-country measures such as the World Bank’s Worldwide Governance Indicators are composite, perception-based estimates that combine surveys of households, firms, and experts. They are useful for comparison but carry margins of error, so small changes should be interpreted cautiously.
Can anti-corruption reform be made to last?
Reforms tend to be more durable when they are embedded in law rather than dependent on a single administration, when oversight bodies are adequately resourced, and when external actors such as courts, auditors, and civil society can hold them accountable over time. No approach eliminates the underlying tension entirely.
What the pattern actually tells us
The puzzle that frustrates observers – reform platforms that are difficult to deliver in full – looks less mysterious when the focus moves from personalities to systems. Anti-corruption pledges are made in an environment that rewards them; they are implemented in one where full delivery can create friction with the very machinery of governing. The most durable reforms tend to be those designed with that tension in mind: written into law rather than resting on individual reputation, funded on a predictable basis, and answerable to institutions beyond the government of the day.
That is a less dramatic story than a campaign promise. It is also closer to how governance actually changes.
