1. Introduction
Over the past three decades, anti-corruption reform has become a central component of domestic governance agendas and international development policy. Governments have enacted specialised legislation, established anti-corruption agencies, strengthened audit institutions, introduced asset-declaration regimes, adopted access-to-information laws and joined international conventions. The United Nations Convention against Corruption (UNCAC), adopted in 20031, represents the most comprehensive global framework in this field. Many states have signed and ratified the Convention and subsequently developed broader legal and institutional architectures directed at corruption prevention, investigation, prosecution, public integrity and asset recovery.
The expansion of formal institutions, however, has not produced a uniform transformation in the incidence or political consequences of corruption. In many countries, specialised agencies coexist with weak accountability; laws coexist with selective enforcement; public officials are required to declare assets while verification remains limited; audit findings are produced without systematic sanctions; and international commitments are accompanied by domestic implementation gaps2. The persistence of these contradictions raises a fundamental question: why do anti-corruption institutions so often exist without acquiring the capacity to constrain powerful actors?
The question is not simply whether anti-corruption institutions are technically capable. Nor is it adequately answered by asking whether a country has sufficient laws on the books. A state may possess an apparently comprehensive anti-corruption architecture while the political conditions necessary for effective enforcement remain absent. This distinction is central to the argument developed here. The existence of an institution is a formal fact. Its effective independence is a political and organisational condition. Its ability to generate accountability outcomes is a further empirical question.
This article therefore distinguishes three dimensions of anti-corruption governance. The first is formal institutional presence: the existence of laws, agencies, procedures and legally defined powers. The second is effective institutional autonomy: the practical capacity of institutions to investigate, prosecute, audit or sanction without being subordinated to the political interests of the actors they are expected to scrutinise. The third is accountability outcomes: whether violations produce credible investigations, judicial proceedings and consequences, including when the alleged perpetrators occupy senior political or economic positions. The central argument is that these three dimensions can diverge substantially.
The core thesis of the article is as follows. Corruption persists, in a large class of political systems, because those who control the state have incentives to keep anti-corruption institutions formally present but effectively subordinate to the political arrangements that patronage sustains. Persistence is therefore a political equilibrium and not an administrative failure. Where political protection shields strategically important actors from accountability, the expansion of formal institutions coexists with selective enforcement, fragmented oversight and symbolic compliance. The corollary is demanding: reform that adds laws and agencies without raising the political cost of protecting powerful actors leaves the equilibrium intact.
The argument does not claim that all corruption is produced by patronage or that every anti-corruption institution is politically controlled. Nor does it suggest that administrative capacity is irrelevant. Rather, it identifies a specific set of mechanisms through which formal institutional reform can fail: political actors who benefit from existing arrangements may influence the institutions that threaten those arrangements. This becomes particularly important when political competition is limited, judicial independence is weak, media scrutiny is constrained or public resources play a major role in maintaining political coalitions.
The theoretical foundation combines four literatures that are often treated separately: principal-agent approaches3, collective action approaches4, research on patronage and clientelism5 and political settlements analysis6. Together they suggest that the effectiveness of anti-corruption institutions depends not only on institutional design but also on the political equilibrium within which those institutions operate.
The article makes three contributions. First, it integrates these strands of literature into a single framework centred on the distinction between institutional presence and effective accountability. Second, it identifies four mechanisms through which anti-corruption institutions can be neutralised without being formally abolished: appointment dependence, selective enforcement, protective fragmentation and symbolic compliance. Third, it derives propositions and observable implications that make the framework testable rather than purely interpretive.
The argument proceeds as follows. Section 2 conceptualises corruption as a political-institutional problem and Section 3 sets out the theoretical foundations. Section 4 distinguishes formal institutions from effective accountability, and Section 5 presents the four mechanisms and the framework summarised in Figure 1. Section 6 derives propositions and observable implications, Section 7 describes the research design, and Section 8 discusses illustrative evidence. Section 9 examines alternative explanations, Section 10 discusses implications and Section 11 concludes.
2. Conceptualising Corruption as a Political-Institutional Problem
Corruption is commonly defined as the abuse of entrusted power for private gain7. This definition remains useful because it distinguishes legitimate public authority from its misuse. However, it can become analytically restrictive when corruption is examined only as an exchange between an individual official and an individual beneficiary. In many political systems, the benefits associated with corrupt practices are not confined to the official who receives them. They may flow to political parties, factions, family networks, firms connected to political elites or groups whose support is important for maintaining political power8.
A political analysis must therefore consider not only the transaction but also the relationships within which the transaction takes place. Scott9 demonstrated that corruption and machine politics can be connected to broader systems of political exchange. Johnston10 similarly emphasised that corruption varies according to relationships between wealth, power and institutions. Warren11 moved the discussion towards democracy and exclusion, arguing that corruption can distort collective decision-making by transferring influence away from citizens and towards those able to use political or economic power.
This broader conception does not eliminate the distinction between petty and grand corruption. Bribery by a low-level official and the diversion of public resources through politically connected firms differ greatly in scale and institutional consequence. The present argument is concerned instead with a common political question: how are benefits from the misuse of public authority embedded in relationships of dependence and power?
The working definition adopted here is therefore the use of entrusted public power in violation of governing rules for the benefit of the office-holder or a group to which the office-holder is politically, economically or socially connected. This formulation includes individual enrichment but also accommodates forms of political patronage and institutional manipulation. It does not imply that every patronage relationship is legally corrupt, nor that every political appointment is improper. The relevant issue is whether public authority is exercised contrary to rules and impartial standards for particularistic benefit.
This distinction matters because anti-corruption policy often treats the individual act as the fundamental unit of analysis. A bribe is detected, an official is disciplined and a procedure is changed. Yet if the official’s behaviour is part of a larger system of exchange, the removal of one individual does not necessarily change the incentives of others. The system can reproduce itself by replacing one participant with another. Understanding corruption as a political-institutional problem therefore requires attention to the structures that reproduce particularistic exchange even when individual actors are replaced.
The political-institutional approach also draws attention to the dual character of anti-corruption institutions themselves. These institutions are simultaneously instruments of rule enforcement and objects of political contestation. Their leaders are appointed through political processes; their budgets are authorised by political authorities; their investigations may affect the interests of those who control the state. Treating them solely as technical devices therefore risks missing the political conditions under which they can or cannot constrain powerful actors.
3. Theoretical Foundations
3.1 The principal-agent model
The principal-agent framework provides one of the foundational economic approaches to corruption. Public authority is delegated from principals to agents. Citizens delegate authority to political leaders, political leaders delegate authority to bureaucrats, and agencies delegate functions to individual officials. Because agents possess information and discretion that principals do not fully control, they may exploit their position for private benefit.
Klitgaard’s12 formulation captures the intuition that corruption increases with monopoly and discretion and decreases with accountability. Rose-Ackerman13 similarly emphasises the economic incentives created when officials control scarce benefits and can exchange public decisions for private payments. The principal-agent model generates clear policy prescriptions: reduce unnecessary discretion, increase transparency, improve monitoring, separate functions, raise the expected cost of corruption and strengthen sanctions. These remain valuable components of anti-corruption policy.
The difficulty arises when the principal is not actually committed to controlling corruption. In a simple administrative hierarchy, it may be reasonable to assume that the government wants bureaucrats to follow the law. In systemic corruption, however, senior political actors may themselves benefit from the same practices they are expected to control. Persson, Rothstein and Teorell14 argue that conventional anti-corruption reforms can therefore fail because they assume the existence of a non-corrupt principal who can discipline corrupt agents. Their analysis of systemic corruption as a collective action problem under conditions of mutually reinforcing expectations is particularly relevant here.
The implication for the present argument is important. The principal-agent model is not rejected. It is placed within a political context. Monitoring and sanctions remain necessary, but the question becomes: who controls the monitors? When those who appoint, fund and protect the monitors themselves benefit from the practices under scrutiny, the principal-agent logic is inverted. The agents of enforcement become potential agents of protection.
3.2 Collective action and systemic corruption
The collective action perspective shifts attention from individual incentives to expectations about the behaviour of others. If citizens believe that officials are generally corrupt, refusing to participate in corrupt exchanges may carry significant private costs. If firms believe that contracts are routinely allocated through political connections, refusing to cultivate those connections may reduce their competitiveness. If officials believe that political loyalty is necessary for promotion, strict adherence to formal rules may appear risky.
This produces a coordination problem. Many actors may prefer a system based on impartial rules, yet each individual may find it rational to participate in the existing system because unilateral compliance does not change the equilibrium. The result is a self-reinforcing equilibrium in which expectations of corruption sustain the very practices that generate those expectations. Mungiu-Pippidi15 expresses a similar idea through the contrast between particularism, in which resources are allocated according to personal and group ties, and ethical universalism, in which they are allocated by impersonal rules, and argues that the transition from one to the other is a collective rather than a purely institutional process.
This perspective explains why laws can coexist with persistent corruption. A new law changes the formal rules, but if actors expect the old rules to continue operating informally, the incentives created by the new law may remain weak. Institutions become performative rather than transformative. The collective action literature therefore complements the political argument developed here. It explains why institutional reform can fail even without explicit political interference in every individual case. Officials may internalise expectations about what is politically safe. Prosecutors may understand which cases are likely to receive support. Auditors may anticipate whether their findings will be acted upon. Civil servants may distinguish between formal rules and politically relevant informal rules. In this sense, institutional neutralisation does not always require direct orders. It can operate through anticipated consequences.
3.3 Patronage, clientelism and political exchange
The literature on patronage and clientelism provides the next component. Patronage involves the distribution of positions, resources or opportunities through political relationships. Clientelism similarly involves reciprocal exchanges between political patrons and clients, often combining material benefits with political support16.
Such relationships are not necessarily reducible to corruption. A government can legitimately appoint political allies to certain political positions, and parties can mobilise supporters through lawful political programmes. The analytical issue arises when access to public resources depends systematically on particularistic loyalty rather than impersonal rules.
Patronage can perform several political functions. It can distribute resources among members of a ruling coalition, thereby reinforcing coalition cohesion. It can create incentives for supporters to remain loyal by making continued access contingent on political alignment. It can reduce the likelihood of defection by raising the costs of exit. It can connect political leaders to business networks capable of providing campaign resources. And it can create shared exposure to wrongdoing, generating a form of mutual vulnerability that further stabilises the coalition.
This last mechanism requires caution. Participation in irregular practices does not necessarily create a deliberate conspiracy of mutual protection. But where several actors benefit from the same informal arrangements, each may possess information about the others. The potential consequences of disclosure can make political defection costly. The result may be a form of mutual dependence that reinforces coalition cohesion even in the absence of explicit coordination.
The key insight is therefore not that corruption is always consciously designed as a political instrument. It is that corrupt practices can become embedded in political relationships and subsequently acquire political functions. Once those functions become important to political survival, the actors who control the state acquire incentives to constrain the institutions that threaten the underlying arrangements.
3.4 Political settlements
Political settlements analysis provides a broader explanation of why institutions that appear similar on paper can function differently across countries. Khan17 defines a political settlement in terms of the relationship between the distribution of power and the distribution of benefits supported by institutions. Where formal institutions are inconsistent with the distribution of power, informal arrangements can modify how those institutions actually operate. The related concept of limited access orders18 points in the same direction: stable political orders are built on arrangements that give powerful groups privileged access to state resources in exchange for restraint, so that formal rules cannot be understood apart from those arrangements.
This perspective is particularly relevant to anti-corruption institutions. A law can formally establish an independent agency, but the agency still operates within a political settlement. Leaders must be appointed. Budgets must be approved. Investigations require cooperation from other institutions. Courts must enforce decisions. Political actors can influence the wider environment in which the institution operates.
The important question is therefore not whether formal rules exist but whether the distribution of political power supports their enforcement. This insight helps explain the apparent paradox of institutional proliferation. International governance reforms often diffuse similar institutional forms across countries. Yet similar agencies can perform very differently because their political environments differ. Institutional design is therefore a necessary but insufficient condition for effective accountability.
4. From Formal Institutions to Effective Accountability
4.1 The institutional presence problem
A recurrent weakness in anti-corruption analysis is the tendency to treat institutional existence as evidence of institutional strength19. A country is said to have an anti-corruption agency because legislation established one. It is said to have judicial independence because the constitution provides for it. It is said to have transparency because a freedom-of-information law exists. These statements are formally correct but analytically incomplete. The standard of impartiality in the exercise of public authority is itself a defining feature of the quality of government20, which is why institutions that exist formally but do not apply rules impartially cannot be taken as evidence of effective governance.
An institution has formal presence when its legal existence, mandate and procedures are established. It has effective autonomy when it can exercise those powers without unacceptable political interference. It produces accountability outcomes when its actions generate credible consequences for violations. The three form a sequence in which formal presence does not guarantee effective autonomy, and effective autonomy does not automatically guarantee accountability outcomes.
An anti-corruption agency may possess investigative powers but lack sufficient resources. It may possess resources but depend on politically controlled appointments. It may investigate cases but depend on prosecutors to proceed. Prosecutors may proceed but courts may delay decisions. Courts may decide cases but sanctions may not be implemented. Accountability is therefore a chain rather than a single institutional function.
4.2 Dimensions of effective institutional autonomy
Formal independence is normally expressed through legal language. An agency may be described as autonomous, independent or impartial. Yet autonomy has multiple dimensions that can vary independently: appointment autonomy (who selects institutional leaders and under what criteria); tenure security (whether leaders can be removed easily); budgetary autonomy (whether resources are predictable and protected from political retaliation); operational autonomy (whether sensitive cases can be initiated without political authorisation); and institutional protection (whether investigators, prosecutors, judges, auditors and whistleblowers are protected when their work threatens powerful interests).
A law may provide tenure security while budgetary dependence remains high. An agency may have budgetary resources but politically controlled leadership. Courts may have formal independence but financial dependence on the executive. The concept of effective autonomy therefore refers not to a single legal attribute but to the practical capacity to act against politically powerful actors.
4.3 Accountability as an institutional chain
Accountability also depends on interdependence. Anti-corruption institutions rarely operate alone. An audit institution may identify irregularities. A prosecutor may investigate them. A court may determine criminal responsibility. Parliament may demand explanations. The executive may need to implement administrative consequences. Civil society and the media may publicise the case. Weakness at any stage can reduce the final accountability outcome.
This creates an important analytical distinction between institutional fragmentation and institutional checks and balances. Multiple institutions can strengthen accountability when each provides an independent counterweight to the others. But fragmentation can also weaken accountability when responsibilities are unclear and each institution can shift responsibility elsewhere. The difference depends on coordination, information flows and the ability of institutions to compel one another to act.
4.4 The accountability gap
The distance between formal institutional presence and accountability outcomes can be described as the accountability gap. A large accountability gap exists when extensive laws and institutions are formally present; institutions possess significant legal mandates; enforcement remains weak or uneven; politically connected actors are less exposed to consequences than peripheral actors; and institutional failures cannot easily be attributed to a single identifiable decision-maker.
The accountability gap is not itself proof of political capture. Administrative capacity, legal complexity and resource constraints can also produce it. The value of the concept is diagnostic: it directs research towards explaining why formal capacity does not translate into outcomes. The central hypothesis of this article is that patronage-based political orders can sustain large accountability gaps because political actors have incentives to preserve institutions formally while constraining their practical independence.
5. The Political Mechanisms of Corruption Persistence
5.1 The political functions of corruption
Corruption can perform political functions even when those functions are not the stated purpose of corrupt acts. The first function is coalition cohesion: access to public resources can help distribute benefits among actors with different interests within a ruling coalition. The second is rewarding loyalty: political leaders can use control over appointments and public resources to reward individuals whose political support is valuable. The third is resource mobilisation: where legitimate political finance is weakly regulated or insufficient, informal access to state resources or politically connected business networks may become important. The fourth is discipline through dependence: individuals who depend on political protection may face high costs if they defect, and shared knowledge about irregular practices can increase these costs.
These functions create a theoretical problem for anti-corruption reform. If corruption is integrated into the political organisation of power, eliminating it may change the distribution of resources and therefore alter the coalition itself. Anti-corruption reform consequently becomes politically distributive.
5.2 Mechanism 1: Appointment dependence
The first mechanism is appointment dependence. Anti-corruption institutions require leaders. Those leaders must be appointed through procedures established elsewhere in the political system. If appointment power is concentrated in political executives or ruling majorities, formal institutional independence may coexist with practical dependence.
Appointment dependence does not mean that every politically appointed official is corrupt or that every appointment results in interference. The mechanism operates through incentives. Suppose an institutional leader knows that the executive controls appointment, that renewal or removal depends on political authorities, that sensitive investigations can affect the political coalition, and that institutional protection is uncertain. Even without receiving explicit instructions, the leader may anticipate the political consequences of certain actions. This is a form of anticipatory compliance. Political influence can therefore operate without direct intervention.
Appointment dependence can also affect the broader institution. Senior officials appoint subordinate managers. Subordinates observe the career consequences associated with particular types of enforcement. Over time, an organisational culture may develop in which certain cases are perceived as politically dangerous. The key prediction is therefore not simply that appointed officials will be corrupt. It is that the probability of politically costly enforcement may decline when institutional leaders depend on actors who could themselves be affected by enforcement.
5.3 Mechanism 2: Selective enforcement
The second mechanism is selective enforcement. Selective enforcement does not require complete inaction. On the contrary, visible enforcement can coexist with political protection. An anti-corruption agency may prosecute numerous officials while rarely reaching the political centre. This can produce the appearance of institutional effectiveness while leaving the most politically consequential networks intact.
Selective enforcement can occur along several dimensions: rank of accused officials; political affiliation; current versus former political allies; economic importance; proximity to senior decision-makers; and timing of political transitions. The critical distinction is between capacity to enforce and willingness or ability to enforce against protected actors.
This mechanism also helps explain why isolated high-profile prosecutions do not automatically disprove a theory of political protection. The relevant question is whether enforcement is systematic and impartial or whether it changes according to political relationships. Selective enforcement may also become more likely after political transitions. Former allies may lose protection, while new coalitions have incentives to expose misconduct associated with previous leadership. Such cases can generate genuine accountability while simultaneously reflecting changes in political protection. This does not mean that every prosecution of a former official is politically motivated. It means that political protection is an empirical variable that should be examined alongside legal evidence.
5.4 Mechanism 3: Protective fragmentation
The third mechanism is protective fragmentation. Modern states typically distribute accountability functions across multiple institutions. This can strengthen checks and balances. But where coordination is weak, fragmentation can create opportunities for institutional responsibility to be displaced.
Consider a simplified chain of audit, investigation, prosecution, trial and sanction. If the audit body identifies irregularities but prosecutors do not act, the prosecutor may attribute the problem to incomplete evidence. If prosecutors investigate but courts delay, prosecutors can point to judicial procedures. If courts issue decisions but enforcement is weak, responsibility can be shifted again. Each institution can therefore possess a partial justification for failure.
Protective fragmentation becomes especially powerful when mandates overlap, information is not systematically shared, procedural thresholds are high, institutional reporting is opaque, judicial proceedings are slow, and no institution is clearly responsible for monitoring the entire accountability chain. The mechanism does not require intentional coordination among institutions. It can emerge structurally from the distribution of responsibilities. The political effect is that no single institution appears responsible for the persistence of impunity.
5.5 Mechanism 4: Symbolic compliance
The fourth mechanism is symbolic compliance. States have incentives to adopt international norms. Ratifying conventions, establishing agencies and publishing strategies can generate legitimacy, facilitate donor relations and demonstrate commitment to governance reform. These actions can be genuinely meaningful. The argument is not that formal compliance is necessarily insincere.
The problem arises when formal adoption is not accompanied by the political and organisational conditions necessary for implementation. Symbolic compliance can therefore produce a stable equilibrium in which an international norm leads to domestic law, institutional creation and reporting, but only to limited implementation. The state gains reputational benefits from the earlier stages while avoiding the political costs associated with deep enforcement.
This is particularly relevant where international actors measure reform through observable outputs such as legislation, strategies and institutional establishment. These outputs are easier to verify than politically sensitive outcomes such as whether prosecutors can investigate a sitting minister or whether courts can sanction powerful economic actors. Symbolic compliance therefore creates a measurement problem. Reform can appear successful according to institutional indicators while remaining weak according to accountability outcomes. Pritchett, Woolcock and Andrews21 describe a related pattern in which states adopt the outward forms of effective institutions without acquiring their functions.
5.6 Conceptual framework and the cycle of persistence
The four mechanisms should not be treated as independent. Appointment dependence can make selective enforcement more likely. Selective enforcement can reduce the political pressure for institutional reform. Fragmentation can make selective enforcement difficult to detect. Symbolic compliance can provide legitimacy for institutions whose effective autonomy remains limited.
The conceptual relationship proposed in this article is summarised in Figure 1. The framework does not assume that patronage automatically produces corruption or that anti-corruption institutions are inherently ineffective. Rather, it proposes that where political survival depends substantially on particularistic networks, political actors may have incentives to protect strategically important relationships from effective accountability. This protection can operate through appointment dependence, selective enforcement, protective fragmentation and symbolic compliance. The resulting gap between formal institutional presence and effective autonomy can weaken accountability outcomes and contribute to the persistence of corruption. Where weak accountability subsequently preserves the resources and relationships underlying patronage, the process can become self-reinforcing.
Figure 1. Conceptual Framework of Corruption Persistence
5.7 Political protection and the costs of impunity
The framework also suggests that enforcement against powerful actors becomes more likely when the costs of protection increase. These costs can originate from electoral competition, judicial independence, independent media, civil society organisations, international pressure and elite fragmentation. These forces do not guarantee accountability. They alter the political incentives surrounding enforcement. The framework therefore predicts that anti-corruption institutions are most effective when they are embedded within a wider environment in which political protection is costly.
6. Propositions and Observable Implications
Two constructs in the propositions require operational definitions. Patronage dependence refers to the extent to which the cohesion and survival of a ruling coalition rely on the discretionary distribution of public jobs, contracts and resources. It can be approximated through indicators such as the share of senior public appointments made at political discretion, the use of public enterprises and procurement for political purposes, and the weakness of political-finance regulation. Political protection refers to the probability that an actor with significant political position or connections will face credible investigation and sanction for conduct that would lead to such consequences for an ordinary official. It can be assessed through the political rank and affiliation of those investigated, the timing of cases in relation to political events, and the documented fate of cases involving connected actors. Both constructs are matters of degree and are best assessed through multiple indicators rather than a single score.
6.1 Proposition 1
Proposition 1. The greater the dependence of a ruling coalition on patronage for political cohesion and survival, the greater the expected divergence between the formal independence and effective operational autonomy of anti-corruption institutions.
The proposition does not predict that formal independence will necessarily be absent. Indeed, political systems dependent on patronage may adopt formal independence precisely because international legitimacy requires it. The predicted relationship concerns the gap between formal and effective independence. Observable indicators include concentration of appointment power, frequency of leadership turnover, security of institutional tenure, budgetary dependence, ability to initiate sensitive investigations, documented interference, and treatment of cases involving senior officials.
6.2 Proposition 2
Proposition 2a. The probability that anti-corruption enforcement reaches politically central actors should decline as the strength of their political protection increases.
Proposition 2b. That probability should rise when political protection weakens or when the political costs of non-enforcement increase.
Observable implications include: cases involving politically peripheral actors should be more common than cases involving the core of a ruling coalition when protection is strong; enforcement against senior actors should increase following coalition fragmentation, political transition or significant external pressure; former allies should be more vulnerable after losing political protection than while occupying central positions; and high-profile enforcement episodes should coincide with changes in the political environment more frequently than would be expected if enforcement were entirely independent of political relationships.
6.3 Proposition 3
Proposition 3. Anti-corruption reforms are more likely to produce sustained accountability when they alter the incentives and constraints facing political decision-makers than when they merely expand the formal powers of enforcement institutions.
Increasing an agency’s legal mandate may improve its capacity. But if the agency remains dependent on political appointments, the reform may not substantially alter enforcement. By contrast, reforms that change appointment procedures, strengthen judicial tenure, protect investigative bodies, increase transparency, improve political finance regulation or strengthen independent media can alter the political costs of interference.
6.4 Observable implications for comparative research
The framework generates a broader set of empirical expectations concerning appointment patterns, case distribution by political rank and affiliation, timing of enforcement relative to political transitions, audit-to-sanction conversion rates, transparency of cases involving senior officials, and clustering of legislative reforms around international review processes without corresponding increases in accountability outcomes. These indicators make the theory falsifiable. If strong institutional autonomy consistently produces impartial enforcement even when political protection is strong, or if enforcement against powerful actors is unrelated to changes in political protection, the framework would require modification.
7. Research Design and Analytical Approach
7.1 Nature of the inquiry
This article is a conceptual and theory-building contribution rather than a causal identification exercise. Its primary purpose is to integrate existing theoretical traditions into a coherent framework that explains the persistence of corruption under conditions of formal institutional expansion, and to derive mechanisms and propositions that can subsequently be tested. The analysis does not rely on a single-country case study design.
The choice of a conceptual rather than econometric design follows from the nature of the research question. The question is not whether a particular variable has a statistically significant average effect on corruption levels across a large sample of countries. It is how and why formal anti-corruption institutions can fail to constrain powerful actors even when those institutions appear, on paper, to possess the necessary legal powers. Answering that question requires attention to mechanisms, sequences and political incentives that are not easily captured by cross-national regressions alone.
7.2 Analytical strategy
The analytical strategy is process-oriented and mechanism-focused. For each of the four mechanisms, the article specifies expected observable implications. Consistency between these implications and documented institutional patterns in different contexts does not constitute causal proof. It does, however, allow assessment of whether the framework provides a coherent interpretation of available evidence.
Particular care is taken to distinguish between (a) relatively uncontested institutional facts (for example, the legal establishment of specialised agencies or the ratification of international conventions), (b) assessments offered by external observers (for example, international indices or civil-society reports on institutional autonomy), and (c) the theoretical interpretation advanced by this article. Assessments from external sources are attributed to those sources and are not treated as independent findings of the present analysis.
7.3 Limitations
Several limitations should be acknowledged. First, the absence of systematic, comparable data on the political rank and affiliation of individuals investigated and prosecuted for corruption across countries constrains the ability to test Proposition 2 rigorously at present. Second, the article cannot observe the private calculations of institutional leaders; anticipatory compliance is inferred from structural incentives rather than demonstrated through direct evidence of intent. Third, comparative research across countries with varying degrees of patronage dependence would be required to assess the generalisability of the framework. Fourth, the analysis is necessarily limited to publicly available information; internal processes of investigation and decision-making often remain opaque.
These limitations do not invalidate the conceptual contribution. They define the boundary between what the present article can claim and what must be left to subsequent empirical research.
8. Illustrative Evidence and Cross-Context Applicability
The framework is intended to be general, and its value lies in directing attention to features that recur across contexts. This section does not test the propositions. It draws on documented cases and comparative studies to show what the distinction between formal presence and effective autonomy looks like in practice, and under what conditions the gap between them narrows.
Comparative studies illustrate the first two dimensions. De Sousa22 observes that, whatever their format and powers, specialised anti-corruption agencies face constraints on their mandate that explain the meagre results obtained by some of them. Read through the present framework, such constraints include the political conditions under which agencies operate: an agency can be formally present while its practical reach depends on actors and decisions it does not control.
Cross-national evidence points in the same direction. In a comparison of ten Asian countries with different patterns of corruption control, Quah23 concludes that only Singapore and Hong Kong succeeded in minimising corruption, owing to the commitment of their political leaders, favourable policy contexts and the impartial implementation of effective measures, while the other eight countries failed because of a lack of political will, unfavourable contexts and reliance on ineffective measures. This is what Proposition 3 would lead us to expect: the decisive difference lay in the political conditions surrounding anti-corruption measures, and not in the mere existence of laws or agencies, which most of the failing countries also had.
A single-country illustration may help to clarify Proposition 2b. In Mozambique, loans of more than two billion United States dollars to state-linked companies were concealed from parliament and the public until 2016, when their disclosure led the International Monetary Fund and other donors to suspend financial support24. An independent forensic audit by Kroll Associates UK Ltd25 was conducted under the mandate of the Attorney General’s Office. Criminal proceedings followed in the Maputo City Court against 19 defendants, including the son of a former president, and the former finance minister was extradited to the United States in 2023 after a legal contest in South Africa in which Mozambique had sought his return to stand trial at home26. The sequence shows that enforcement reached politically connected actors, but only after intense external scrutiny, the withdrawal of donor support and the involvement of foreign courts, which is consistent with the expectation that the costs of impunity matter. It also shows the limits of the argument: enforcement did occur, and how far it reached remains a matter of public and scholarly debate. The case is used here as an illustration and not as a test.
The cross-contextual applicability of the framework rests on the presence of the underlying political conditions rather than on geography or level of economic development. It can therefore be applied to dominant-party systems, competitive authoritarian regimes, newly democratising states and even certain configurations within established democracies in which particularistic networks retain significant influence over public resources. The relative weight of the four mechanisms will vary, but the underlying logic, that political incentives shape the effective autonomy of formal institutions, remains constant.
9. Alternative Explanations and Scope Conditions
9.1 Administrative capacity
The most straightforward alternative explanation is administrative capacity. Anti-corruption institutions may fail because they lack investigators, prosecutors, forensic specialists, information systems or training. Weak capacity can delay investigations and reduce successful prosecutions. This explanation should not be dismissed. The distinction is empirical. A capacity explanation predicts that enforcement weakness should be relatively broad and not systematically correlated with political status. A political protection explanation predicts that capacity may coexist with uneven enforcement. The two explanations can also coexist: political leaders may strategically under-resource institutions, making capacity itself a mechanism of political control.
9.2 Legal complexity
Corruption cases are often inherently difficult to prosecute. Financial crimes involve complex transactions, multiple jurisdictions, corporate structures and evidentiary challenges. Legal complexity can produce delays without political interference. The distinction again lies in variation across cases. If complexity alone explains delays, cases of similar complexity should experience similar institutional treatment. If politically sensitive cases receive systematically different treatment after controlling for complexity, political protection becomes more plausible.
9.3 Social norms and economic incentives
Cultural acceptance of corruption and low official salaries can contribute to the persistence of petty corruption. These factors matter, but they do not adequately explain high-level corruption involving actors with substantial legal incomes and political resources. The framework developed here is most relevant to explaining the persistence of politically consequential corruption, particularly where corruption is connected to state contracts, public enterprises, political finance or elite networks. Norms may also be shaped by institutional performance: repeated observation of impunity for powerful actors can revise citizens’ expectations about what behaviour is realistic27.
9.4 Institutional maturation
New institutions may simply require time to mature. Staff gain experience, procedures become clearer and professional norms develop. However, maturation is not automatic. If the political incentives that initially constrained an institution remain unchanged, time may improve administrative competence without producing greater independence. Maturation should therefore be analysed together with changes in political incentives.
9.5 Scope conditions
The theory should apply most strongly where a dominant political coalition controls state appointments, public resources are important to political mobilisation, electoral competition is weak or uneven, courts have limited independence, media access is constrained, civil society has limited capacity, and accountability institutions depend financially on political authorities. It should apply less strongly where political competition is genuine, courts have secure tenure, appointments are dispersed, media organisations are independent and political finance is transparent. These scope conditions indicate that corruption persistence is not an inevitable feature of any particular region or income category. The framework is political, not geographical.
10. Discussion
10.1 Theoretical contribution
The central theoretical contribution of the framework is the distinction between institutional presence and effective accountability, together with the specification of four mechanisms that can sustain the gap between them. Much anti-corruption reform focuses on creating institutions. Yet institutional creation is an observable output that does not necessarily change the political equilibrium. A more demanding framework asks: Who appoints institutional leaders? Who controls institutional resources? Who can remove institutional leaders? Who decides whether sensitive cases proceed? Who controls information? Who can protect investigators and whistleblowers? Who ensures that judicial decisions are implemented? Who bears the political cost when enforcement fails? These questions move the analysis from institutional design to political incentives.
10.2 Implications for comparative research
The framework can be tested comparatively by selecting countries with similar formal anti-corruption architectures but different political environments and comparing appointment procedures, budgetary autonomy, judicial tenure, media independence, civil-society access, political competition, prosecution rates by political rank, and implementation of audit recommendations. Process-tracing of major corruption cases from allegation to final sanction, institutional event analysis of leadership changes, and network analysis of relationships between political office-holders, businesses and public contracts offer complementary strategies.
10.3 Policy implications
Governments and donors should distinguish between outputs (laws adopted, agencies established, officials trained, strategies published) and outcomes (investigations initiated, cases prosecuted, cases reaching judgment, sanctions implemented, assets recovered, senior officials investigated, audit recommendations implemented). If appointment dependence is a mechanism of political neutralisation, appointment procedures become a central reform target. Transparent selection criteria, public nomination processes, professional qualification requirements, fixed terms, clearly defined removal procedures, multi-institutional oversight and protection against arbitrary dismissal can reduce the concentration of appointment power.
Reform should also address institutional interdependence: clear referral procedures from audit to investigation, access to financial information for prosecutors, adequate judicial resources for complex cases, public accessibility of judgments, and monitoring of asset recovery. Information is a critical counterweight to political protection. Anti-corruption institutions require constituencies that value and defend their independence. External actors can increase the costs of impunity, but episodic pressure may produce episodic enforcement. Sustained engagement focused on institutional independence, judicial capacity, financial investigation, asset recovery, information systems and civil-society capacity is more likely to strengthen domestic accountability than episodic conditionality tied solely to legislative adoption.
10.4 Reform as a change in political equilibrium
The most important implication is that anti-corruption reform should be understood as a change in political equilibrium. If corruption helps distribute benefits within a coalition, reducing corruption changes who receives those benefits. If appointments are used to reward loyalty, transparent appointments reduce discretionary political control. If selective enforcement protects allies, impartial enforcement increases uncertainty for powerful actors. If fragmentation diffuses responsibility, integrated accountability makes failures more visible. If symbolic compliance provides legitimacy without implementation, outcome-based evaluation increases the cost of superficial reform. In each case, reform changes incentives. This is why legal reform can be simultaneously necessary and insufficient. The cases discussed in Section 8 suggest that episodes of significant progress in reducing high-level impunity have often coincided with changes in the broader political environment, such as intensified competition, elite fragmentation, sustained civic mobilisation or external pressure that raises the costs of protection, rather than with institutional redesign alone. This does not render institutional design irrelevant; it situates design within a wider set of political conditions that determine whether design can be made effective.
11. Conclusion
Why does corruption persist when governments have laws, agencies and international commitments designed to combat it? This article has argued that the answer lies in the political functions that corruption performs and in the incentives created by the distribution of political power. Where patronage contributes to coalition cohesion, loyalty and political survival, corruption becomes embedded in the organisation of political authority, and the institutions created to control it operate within the same political environment that sustains the practices they are meant to constrain. Persistence is therefore a political equilibrium and not an administrative failure.
The article developed a framework based on four mechanisms (appointment dependence, selective enforcement, protective fragmentation and symbolic compliance) that explain how anti-corruption institutions can remain formally present while their effective autonomy and accountability outcomes remain limited. The distinction between formal institutional presence, effective institutional autonomy and actual accountability outcomes is central. Accountability must be analysed as a chain of institutions and incentives rather than as the existence of a single agency.
The three propositions provide a basis for empirical testing. They link the gap between formal and effective independence, the distribution of enforcement and the success of reform to the political incentives facing those who control the state.
The argument does not imply that administrative capacity, legal reform, social norms or economic incentives are unimportant. Nor does it suggest that all anti-corruption institutions are captured or that all prosecutions of political actors are selective. Rather, it identifies a political mechanism that should be considered alongside these explanations.
The central empirical question for future research is therefore not simply whether anti-corruption institutions exist. It is whether they possess the autonomy, protection, information and political support necessary to constrain actors who control or benefit from the existing distribution of power. Durable anti-corruption reform ultimately depends on closing the gap between formal institutional presence, effective institutional autonomy and actual accountability outcomes. Laws and institutions remain necessary foundations. But where the political incentives of those who control institutions remain aligned with impunity, institutional design alone is unlikely to be sufficient. The deeper challenge is to create political conditions under which impartial enforcement becomes both possible and sustainable.
Endnotes
1. United Nations. (2003). United Nations Convention against Corruption. United Nations Office on Drugs and Crime.
2. de Sousa, L. (2010). Anti-corruption agencies: Between empowerment and irrelevance. Crime, Law and Social Change, 53(1), 5-22. https://doi.org/10.1007/s10611-009-9211-3; Quah, J. S. T. (2011). Curbing corruption in Asian countries: An impossible dream? Emerald Group Publishing.
3. Klitgaard, R. (1988). Controlling corruption. University of California Press.; Rose-Ackerman, S. (1999). Corruption and government: Causes, consequences, and reform. Cambridge University Press.
4. Persson, A., Rothstein, B., & Teorell, J. (2013). Why anticorruption reforms fail: Systemic corruption as a collective action problem. Governance, 26(3), 449-471. https://doi.org/10.1111/j.1468-0491.2012.01604.x
5. Kitschelt, H., & Wilkinson, S. I. (Eds.). (2007). Patrons, clients, and policies: Patterns of democratic accountability and political competition. Cambridge University Press.
6. Khan, M. H. (2010). Political settlements and the governance of growth-enhancing institutions [Unpublished research paper]. SOAS, University of London. https://eprints.soas.ac.uk/9968/
7. Transparency International. (2009). The anti-corruption plain language guide. Transparency International. https://www.transparency.org/whatwedo/publication/the_anti_corruption_plain_language_guide
8. Kitschelt, H., & Wilkinson, S. I. (Eds.). (2007). Patrons, clients, and policies: Patterns of democratic accountability and political competition. Cambridge University Press.; Khan, M. H. (2010). Political settlements and the governance of growth-enhancing institutions [Unpublished research paper]. SOAS, University of London. https://eprints.soas.ac.uk/9968/
9. Scott, J. C. (1969). Corruption, machine politics, and political change. American Political Science Review, 63(4), 1142-1158.
10. Johnston, M. (2005). Syndromes of corruption: Wealth, power, and democracy. Cambridge University Press.
11. Warren, M. E. (2004). What does corruption mean in a democracy? American Journal of Political Science, 48(2), 328-343.
12. Klitgaard, R. (1988). Controlling corruption. University of California Press.
13. Rose-Ackerman, S. (1999). Corruption and government: Causes, consequences, and reform. Cambridge University Press.
14. Persson, A., Rothstein, B., & Teorell, J. (2013). Why anticorruption reforms fail: Systemic corruption as a collective action problem. Governance, 26(3), 449-471. https://doi.org/10.1111/j.1468-0491.2012.01604.x
15. Mungiu-Pippidi, A. (2015). The quest for good governance: How societies develop control of corruption. Cambridge University Press.
16. Kitschelt, H., & Wilkinson, S. I. (Eds.). (2007). Patrons, clients, and policies: Patterns of democratic accountability and political competition. Cambridge University Press.
17. Khan, M. H. (2010). Political settlements and the governance of growth-enhancing institutions [Unpublished research paper]. SOAS, University of London. https://eprints.soas.ac.uk/9968/
18. North, D. C., Wallis, J. J., & Weingast, B. R. (2009). Violence and social orders: A conceptual framework for interpreting recorded human history. Cambridge University Press.
19. de Sousa, L. (2010). Anti-corruption agencies: Between empowerment and irrelevance. Crime, Law and Social Change, 53(1), 5-22. https://doi.org/10.1007/s10611-009-9211-3; Pritchett, L., Woolcock, M., & Andrews, M. (2013). Looking like a state: Techniques of persistent failure in state capability for implementation. Journal of Development Studies, 49(1), 1-18.
20. Rothstein, B., & Teorell, J. (2008). What is quality of government? A theory of impartial government institutions. Governance, 21(2), 165–190. https://doi.org/10.1111/j.1468-0491.2008.00391.x; Rothstein, B. (2011). The quality of government: Corruption, social trust, and inequality in international perspective. University of Chicago Press.
21. Pritchett, L., Woolcock, M., & Andrews, M. (2013). Looking like a state: Techniques of persistent failure in state capability for implementation. Journal of Development Studies, 49(1), 1-18.
22. de Sousa, L. (2010). Anti-corruption agencies: Between empowerment and irrelevance. Crime, Law and Social Change, 53(1), 5-22. https://doi.org/10.1007/s10611-009-9211-3
23. Quah, J. S. T. (2011). Curbing corruption in Asian countries: An impossible dream? Emerald Group Publishing.
24. Al Jazeera. (2022, November 30). Mozambique court hands out verdicts in $2bn corruption case. https://www.aljazeera.com/news/2022/11/30/mozambique-court-hands-out-verdicts-in-2bn-corruption-case
25. Kroll Associates UK Ltd. (2017). Independent audit related to ProIndicus S.A., EMATUM S.A. and Mozambique Asset Management S.A. [Project Montague, executive summary, redacted for publishing].
26. Al Jazeera. (2022, November 30). Mozambique court hands out verdicts in $2bn corruption case. https://www.aljazeera.com/news/2022/11/30/mozambique-court-hands-out-verdicts-in-2bn-corruption-case; Associated Press. (2023, July 12). Former Mozambique finance minister is extradited to the US to face trial over $2 billion scandal. KSAT. https://www.ksat.com/business/2023/07/12/former-mozambique-finance-minister-is-extradited-to-the-us-to-face-trial-over-2-billion-scandal/
27. Persson, A., Rothstein, B., & Teorell, J. (2013). Why anticorruption reforms fail: Systemic corruption as a collective action problem. Governance, 26(3), 449-471. https://doi.org/10.1111/j.1468-0491.2012.01604.x
