Reforming Reform: Revising the Anticorruption Playbook

Two generations of activism and research in support of corruption control have produced indifferent results at best. What kind of thinking – and rethinking – might lead to better outcomes? This essay offers critical commentary about the contemporary anticorruption movement, made from the perspective of a longtime friend and active participant. (When I write that “we” have struggled with this or failed to realize that, I include myself.) The arguments in this essay fall into two categories: common reform themes that need rethinking or, in some cases, replacement; and a discussion of what history might tell us about the drivers of both corruption and sustained opposition to it. My ideas in no way supplant reforms underway today, but rather are intended to build stronger foundations for them. The result is not a list of “best practices,” but ideas about better practices, tailored to address a range of challenges.

It would be wrong to suggest that the reform movement has failed. Corruption, for many years a nonissue in academe, business, and international policy, now has a prominent place on the global agenda. Annual governance rankings still make headlines. Aid programs and investment decisions treat corruption as a prominent concern, while official exploitation and misconduct are mobilizing grievances in political upheavals and “color revolutions” in many societies. We may take such heightened awareness of corruption for granted nowadays, but it was not always so. This itself is a significant accomplishment.

Positive results, however, are another matter. Success has been possible within specific agencies and locales, and the Hong Kong and Singapore reform sagas are familiar history. But clear-cut, sustained reductions in corruption in diverse societies on the state level have been few. Country-level indices point to a few cases – such as Japan and Belgium – where corruption is perceived to be in decline, but the measures used are problematic on the grounds of validity (most measure perceptions of corruption, not corruption itself) and reliability (country scores usually have large standard errors, making many comparisons suspect). More persuasive process-tracing evidence from the European Union's massive anticorrp research project has identified seven countries – Chile, Costa Rica, Estonia, Georgia, South Korea, Taiwan, and Uruguay – as having made progress on corruption; Rwanda and Botswana appear to be close behind.¹ Those cases offer valuable lessons, but as comparatively small (save for Korea) and ethnically homogeneous societies, they are unrepresentative of the full range of countries.

How can we do better?

Commitment is not the problem. Around the world people and groups are working hard, often at personal risk, to fight abuses of power and wealth. Strategies and tactics have evolved: less emphasis is now placed on privatization, scaling back the state, and relying on markets in place of public institutions; and there is greater appreciation of the global nature of many corruption problems. International cooperation has grown: the Organisation for Economic Cooperation and Development (oecd) Anti-Bribery Treaty, Group of States Against Corruption (greco), and United Nations Convention Against Corruption (uncac), along with the venerable U.S. Foreign Corrupt Practices Act and the uk Bribery Act of 2010, bring cross-border support to the struggle.² New metrics of corruption, and of the effects of reforms, continue to appear.³

Still, the same ideas and strategies tend to dominate even in quite different settings. Top-down national efforts emphasizing crime-prevention concepts and transparency, often built around anticorruption agencies (acas); improved administrative processes; independent judiciaries and news media; and poorly defined support from civil society, remain core themes.⁴ These are scarcely bad ideas in themselves, but making them work in an applied context is another matter. We acknowledge that historical and cultural variations among societies matter, yet devise reform strategies that are remarkably similar from one case to the next. We still have no clear sense of which ideas to apply in what sequence and which to avoid entirely when dealing with contrasting situations or problems. Equally frustrating, it has proven difficult to win lasting political and public support and credibility for reform initiatives, even though large majorities would benefit from effective corruption control.

A number of reform assumptions and distinctions go largely unquestioned; some seem to be repeated mostly because we hear others repeating them. Too little thought, for example, has been devoted to asking what the opposite of corruption might be and how we might build positive support for that. “No corruption” is neither possible nor a credible goal for engaging sustained citizen interests; and “less corruption,” or less disruptive varieties of it, while more realistic, will inspire few. Technocratic visions of “good governance” often reduce government to a referee in social and economic arenas, thereby undervaluing justice and sidestepping the political contention often required to pursue it. In politically or culturally complex and/or divided societies, reform agendas based on harmony and moral consensus are unlikely to succeed, and in any event give little concrete guidance on how to move forward.

History offers one way to conceptualize the opposite of corruption: a reduction in corruption (to within reasonable limits) has often been a byproduct of prolonged political contention over the sources, uses, and limits of power. Ensuring that citizens have a voice in such processes and can defend themselves against official abuses is a process I have called deep democratization.⁵ Integrating citizens and their needs and wishes into governing lends new meaning to the notion of integrity, evoking honesty and transparency but also wholeness.

Reform strategies often place excessive faith in transparency, at times treating the term as synonymous with corruption control. Transparency is a laudable principle so long as it does not place vulnerable people at risk: political and legal scholar Richard Briffault has cited cases in which backers of controversial political causes have been threatened once their contributions have been disclosed, and fear of reprisal from employers has been cited as a justification for not disclosing individuals’ small contributions.⁶ But as a means of corruption control, transparency has problems. To begin, it addresses relatively few of the behaviors and social issues that many citizens view as corruption. For instance, a large majority of Americans see political contributions as corrupting democratic life despite (or perhaps precisely because of) the fact that most such money changes hands legally in publicly disclosed transactions.⁷ Transparency can also strengthen already-influential interests well-placed to capitalize upon access and openness, and can even facilitate corruption: according to some critics, the Legislative Reorganization Act of 1970, by making the workings of Congress and its committees more transparent and reducing the power of committee chairs, not only gave contributors more leverage over individual members of Congress, but also made it easier to track the activities of donation recipients through phases of the legislative process. In other words, contributors found it easier to ensure that they received value for their money.⁸ Transparency, therefore, is not an inherent good, and will accomplish little if citizens have little reason to “look in,” do not feel safe in doing so, or lack the political resources and opportunities to act on what they see.

Distinguishing between “grand” and so-called “petty” corruption has intuitive appeal: massive fraud in the course of building a dam and small protection payments regularly demanded by police are not the same thing. But precisely what the grand-versus-petty distinction helps us understand has never been clear. Indeed, this opposition may obscure more than it reveals, particularly if it is invoked to minimize the importance of the “petty”: “petty” corruption is a serious problem, keeping poor people poor and the powerless vulnerable. Far from being distinct problems, “grand” and “petty” corruption can enable each other.⁹ The spark that set off Tunisia's national upheaval was a public suicide resulting from police abuse of a young fruit seller in an open market: seemingly petty corruption that nonetheless enjoyed protection from the top of the system.¹⁰ In other instances, front-line functionaries share bribes with the superiors to whom they owe their jobs, creating a sizable upward flow of money into the hands of a select few. We do need to differentiate among the kinds of corruption problems (more on that below) but the value of the grand-versus-petty distinction – that is, what analytical work it does for us – remains unclear.

Political will – which we might define as politicians’ firm intention and desire to effect change, although reformers rarely say precisely what they mean when they invoke the term – is often considered the foundation for corruption control. Few would dispute that reforms will more likely succeed if backed by leaders. But will of any sort is a matter of intentions and dispositions and as such is fundamentally unknowable a priori. High-profile proclamations of “zero tolerance” may come to naught, and splashy reform campaigns can be mostly for show – or worse, aimed at jailing critics. In practice, we cannot assess political will until we have outcomes to analyze in their full political context, as public policy scholar Derick Brinkerhoff's framework for assessing political will suggests.¹¹ Good intentions can run headlong into historical constraints, social divisions, a lack of resources, or entrenched opposition. A leader or regime might overcome some obstacles by coercion, but that sort of “will” can do immense damage to state integrity and will scarcely foster anticorruption strength in the rest of society. If anything, many extensively corrupt societies suffer from an excess of political will – as powerful figures silence their critics and derail or prevent investigations – and from the weakness of countervailing institutions and interests.

Political will or its absence can inform post hoc assessments of anticorruption efforts, particularly if we break the general concept down into more specific sources of support and opposition. But while it might be deemed a necessary condition for corruption control, it is usually insufficient on its own. Calls for political will oversimplify the complexities of building social support for reform, and outcomes usually reflect a great many interconnected and context-specific influences. Indeed, if reforms fail, they invite us to blame the victims: we gave them the right tools and ideas, but the leaders just wouldn't see them through. Finally, if a genuine anticorruption champion should appear, what happens when she or he leaves the stage, or when once-strong support begins to fade? Ronald MacLean-Abaroa's corruption-control efforts during his time as mayor of La Paz, Bolivia, produced quite positive results, but once he left office, many corrupt practices returned.¹²

The notion that civil society has a central role to play in corruption control is widely accepted, but often in ways that reflect a narrow view of civil society itself. Many civil-society tactics center on formal organizations advocating reform as a public good: that is, as a cause that all should actively support because all will benefit. In developing countries, many such groups are donor-funded, operate mostly in and around national capitals, and are guided by donors’ agendas; often they are as concerned with protecting their own resources as they are with introducing significant change.¹³ But the strength of civil society considered less narrowly is also found in groups and informal social activities that have little to do with public purposes yet still build social capital in the form of networks, skills, and trust that can be mobilized in many ways for many reasons. In our possibly overoptimistic scenarios about civil society in the United States, for example, residents wanting to clean up a park do not necessarily organize a formal group.¹⁴ Instead, they may draw upon their own and shared networks – friends, members of clubs and churches, and outdoor enthusiasts – to get the job done. Such civil-society mobilization is difficult to sustain from above, but it can happen organically; someone chooses to take the lead, and side benefits such as beer and a barbecue at the end of the day will make the cleanup all the more attractive.

The point is that collective action cannot rely solely on formal purpose-oriented organizations, but must engage a wider range of social ties and incentives.¹⁵ The democratic transition in 1970s Spain, for example, was aided by a civil society that proved just strong enough to sustain trust and discourage massive disorder. Later data showed, however, that Spain at the time of Franco's death had few autonomous social organizations.¹⁶ What it did have were long-standing, deep traditions of informal socializing in neighborhoods and local communities – arguably a durable substitute for formal organizations.¹⁷ Similar informal networks supported democratic reforms in neighboring Portugal. Particularly in postconflict and postauthoritarian settings, reformers might do well to encourage the formation of multipurpose women's, students', and farmers’ organizations; social clubs and music societies; labor unions; and neighborhood mutual-aid schemes that offer citizens things they want and need. Few such groups will have dedicated anticorruption agendas or acquire strength immediately, but over time all might contribute to networks and trust, and diffuse organizational skills in ways that are useful to challenge official exploitation. That approach will take time and patience, but the resulting social ties will likely be strong, versatile, and grounded in relationships unlikely to emerge from donor-driven advocacy of public goods in civil-society and nongovernmental organizations.

A final point in need of rethinking is reliance upon one-dimensional, country-level corruption indices and rankings. Such indices do keep corruption on the agenda and direct our attention to regimes that would rather we looked the other way. But some are of dubious validity and reliability, exaggerate the precision of results, and have problems tracking change.¹⁸ For example, by attributing corruption to the societies in which it is revealed, these indices may overlook cross-border dealings. Further, important reforms requiring major political capital frequently fail to “move the needle” on such indices, while the trials, scandals, and evidence that emerge when a country gets serious about corruption can ironically make perceptions worse. Analysts have made strides in assessing the overall scope of corruption in societies and regions, but measurement on the scale at which reforms are effected remains difficult.¹⁹ Do improvements in specific procurement or customs functions, for example, actually restrain corruption in those areas of government? As a result, it has been hard to assess tightly targeted controls to show that they are producing benefits, and thus to show citizens that reform is real and can improve their lives.

What might work better ? First, we might stop thinking of corruption as a national characteristic attributable to all parts and levels of a society and recall that a one-number score may distort more than it reveals. Corruption often arises in small niches: a procurement process; a relationship between a politician and a contributor, or between officials and vendors; or in a tax assessor's use of discretion. Indicators of government performance – how long it takes to get a license or permit, the variability of inspections or tax assessments, prices paid for comparable commodities like fuel or concrete – benchmarked over time and across jurisdictions, can signal the effects of corruption and the incentives sustaining it.²⁰ They will not measure corrupt dealings directly or generate headline numbers about whole regimes, but may well give reformers critical insights into points of vulnerability and the effects of new controls.

No master plan will suffice to check corruption as a singular problem, for that is not the kind of problem it is. Progress must take many, and evolving, forms, as has been the case historically, and the impact of our efforts must be assessed in numerous ways. “Best practices” will be elusive; indeed, what is effective in Country A can be impossible in Country B, irrelevant in Country C, and downright harmful in Country D (as with privatizations in Russia in the 1990s).²¹ We can, however, work toward better practices adapted to the diverse contexts in which corruption is embedded, and reflecting a fuller understanding of the ways both it and sustained anticorruption opposition take root.

To a surprising extent, we treat corruption as though it were essentially the same thing everywhere, but this is not so in two senses. First, “corruption” has long been implicitly equated with “bribery.” A United Nations document on corruption-prevention processes, for example, while allowing that corrupt activities arise in many situations, nevertheless simply equates them with bribery and extortion.²² This kind of normative thinking is understandable in part because bribery is likely the most common corrupt practice, and in part because direct quid pro quo transactions between parties on relatively equal footing are easily modeled, and so receive more attention.²³ But that is to underestimate the sheer diversity of the phenomenon: nepotism, official theft, and extended rings of collusion in privatization or customs functions (to name a few variations) involve diverse interactions, timelines, risks, and gains, and can have differing origins and consequences. Some corrupt dealings revolve around the activities of middlemen who make temporary, but lucrative, connections between citizens and officials. Others involve repressive uses of authority or outright violence, and are scarcely equal trades. Examples include more lurid forms of extortion as well as the plata o plomo (“silver or lead”) choices forced upon state and local officials by Mexican drug cartels.²⁴ Some corruption is carried out openly and with impunity. Some corruption undermines order; elsewhere it is doled out as patronage – and functions as a means of control and of maintaining social order (albeit a dysfunctional one).²⁵ In still other cases, elite collusion unifies ruling coalitions facing rising competition, sustaining a de facto political predictability that can coexist with sustained economic growth. Recent research argues that not all corruption is illegal: some of it works through, not in defiance of, laws and institutions.²⁶

The second sense in which corruption is not uniform has to do with the deeper origins of corruption. What are the most important contrasts to understand? Useful distinctions – for example, “need” corruption versus “greed” corruption²⁷ – have emerged in the literature. Anticorruption expert Adam Graycar's TASP (Type, Activities, Sectors, Places) framework can map occurrences and vulnerabilities.²⁸ But most other typologies categorize derivative details, not fundamental contrasts. My own work points to four broad syndromes of corruption defined by the openness in political and economic arenas and by the strength of state, political, and social institutions.²⁹Official Moguls cases (such as Egypt, Nigeria, and in a larger and more decentralized way, China) are dominated by a few elites in a setting of very weak institutions, monopolizing power and resources for themselves and for clients and ruling more by personal power than official authority. Oligarch-and-Clans situations (for example, in Mexico and the Philippines) likewise involve very weak institutions, but in these cases several contending powerful figures and their followers amass both wealth and power in a setting of pervasive insecurity. There is little doubt about who is in charge in Official Mogul cases, while in Oligarch-and-Clan situations, it may not be clear that anyone is in control: corruption is often linked to violence as oligarchs struggle to protect their gains and enterprises from predation by other oligarchs. Elite Cartel societies (such as the Republic of Korea, Botswana, and Poland) are often new or reforming democracies with emerging markets; their institutions are only moderately strong, but they are dominated – and stabilized – by networks of colluding elites sharing the proceeds of corruption while seeing off would-be competitors. Influence Market cases (such as the United States, Japan, and Australia) tend to be affluent market democracies with open, generally well-institutionalized politics and economies. In these states, however, political influence (usually over specific decisions and benefits) is traded as a commodity, often legally. Influence Markets might seem relatively benign, but they affect policy – and often limit political and economic competition – within large and important economies, and affect many other societies via global markets. These syndromes can point to contrasting challenges of building support for reforms. Thus, more work and creative thinking is needed regarding contrasts in the underlying causes, inner workings, evolution, and consequences of corruption problems.

Much can be learned from a fresh historical look at corruption and reform. Most theories of change supporting reform efforts turn history upside down or get it backward.³⁰ The former involves overemphasizing reform from above while taking it as a given that political support from below will develop naturally. Missing from such scenarios are the political contention and bottom-up demand needed to use diverse grievances to mobilize broad support for controls and check the powerful. Getting history backward means that we mistake outcomes – legislation, institutions, a middle class, an active civil society, anticorruption agencies, checks by the press and the courts – for the causes of better government.³¹ But what initially propelled well-governed countries toward good government is not necessarily what sustains it now, and there is no guarantee that things will not get worse. For struggling societies, simply emulating the laws and institutions of successful countries without ensuring solid social demand for reform, grounded in lasting social values and interests, will be like pushing on one end of a string.

We see variations on that theme today. Many countries have anticorruption laws on the books (even if penalties need updating), an ACA of some sort, and numerous externally funded governance projects. Few in society, however, have a compelling stake in their success; support from courts and prosecutors is weak or absent, and enforcement is ineffective. Strong, effective laws and institutions found elsewhere, by contrast, were rarely if ever implemented from the start, but rather emerged out of long and contentious processes of deep democratization, driven by citizens’ demands for better treatment and ways to protect their interests. Frederick Douglass put it best in 1857: “Power concedes nothing without a demand. It never did and it never will.”³² Those demands, in turn, cannot be taken for granted, and usually require broad social support: it is essential that leadership have real social roots and knowledge of what concerns citizens most.³³

The argument follows – correctly – that to succeed, deep democratization requires some degree of liberty and security, or at least of political space to express oneself as well as a diversity of active voices in society. But that does not mean reform must await the arrival of institutionalized democracy (which, after all, creates corruption risks of its own). Economist Jonathan Isham and governance experts Daniel Kaufmann and Lant Pritchett have shown that even in undemocratic societies, basic civil liberties – such as freedom to criticize the regime in public occasionally – are linked to better use of aid resources.³⁴ The real lesson here is that basic political changes must often be underway before corruption controls can gather force.

How might deep democratization work in practice, particularly where political, administrative, and law-enforcement institutions are weak or manipulated from above? The four syndromes offer some clues, not in the form of specific reform menus, but rather in terms of the social and political foundations that must be built to support existing and future measures.³⁵ Particularly for the Official Moguls and Oligarchs-and-Clans syndromes, this process must usually be long-term and indirect, focused not on “fixes” but rather on building lasting resistance to corruption and opportunities for its expression. Even periods of rapid change often build upon underlying longer-term developments that have brought new interests into being and enabled them to make demands. The argument here is not that struggling societies should let corruption go unchallenged until some developmental checklist has been completed. Rather, caution and pragmatism are in order: confronting entrenched corrupt regimes prematurely can end in tragedy, as with the Tiananmen Square demonstrators of 1989, on whose lists of grievances corruption ranked highly.³⁶

Therefore, where Official Moguls monopolize power, a key reform task is to increase political pluralism over time by enabling more people to voice their interests. Corruption may be just one of many entrenched governance issues a country faces as it begins moving toward more pluralism; in a way, the specific grievances people raise matter less than their ability to raise them at all. Repressive regimes will not welcome new voices, so increasing pluralism will be a long and difficult process, just as it was, historically, in many of today's well-governed societies.³⁷ Reformers might seek to restrain abuses by police and exploitation by officials interacting directly with citizens; set a higher standard of professionalism and independence for judiciaries and the press; and pursue meaningful (if realistically limited) opportunities for association and expression. Nonpolitical groups, social and recreational networks, ethnic or migrant communities, and some religious groups may offer safe ways to build strength and mutual trust.

In Oligarchs-and-Clans scenarios, pluralism abounds. Indeed, notwithstanding oligarchs’ personal clout, it can be unclear whether anyone is really in charge. Potential opponents of corruption are numerous but find it risky in a climate of danger and insecurity to challenge oligarchs, who may be linked to organized crime, drug cartels, or private armies. In this case, the primary reform goals – which are far from easy to achieve – are to reduce citizens’ pervasive sense of insecurity and create safe and valued political and economic spaces where in – dividuals can pursue and defend their own interests. This includes strengthening electoral, financial, law-enforcement, and judicial bodies so they are neither colonized by oligarchs nor supplanted by mafias. Corruption may be a common thread linking a large number of specific grievances around which people can be mobilized once they have a safe space in which to act.

Elite Cartel cases, in which dominant colluding elites face pressure from competing political and economic forces, are characterized by multiple active interests and a moderately well-institutionalized political space. In these cases, in addition to mobilizing social interests and maintaining safe political space, a major goal is to increase political and economic competition and openness, which often entails direct challenges to collusive corruption. Familiar anticorruption and institution-building measures can be effective if they enjoy social support. In addition, many Elite Cartel societies have experienced sustained economic growth, increasing the range of active interests in society and reinforcing citizens’ incentives to be their own advocates. Entrenched elites will not give ground gladly, but unlike Official Moguls and Oligarchs, for whom defeat may mean ruin or death, they may have the option of making way for more competition. Moreover, the costs of outright repression, both economically and in terms of public image, may well counsel political accommodation.³⁸

Finally we come to Influence Markets, and to a paradox: these regimes have institutionalized anticorruption laws and ideas and flagrant abuses are uncommon, yet much activity commonly seen as corrupting takes place within the limits and protection of the law. Monetary contributions to political campaigns are just one example. The challenge in such environments is to demand enforcement of corruption controls, increase competitiveness in economies and politics, and roll back legislation creating unfair advantages and suspect rents. Such efforts typically must traverse political, economic, and legal landscapes long ago reshaped to suit the wealthy, who have the added advantage of defending a status quo many people see as legitimate. Transparency, ironically, may do as much harm as good, persuading many that wealth alone is what really decides elections and shapes public policy. From one election cycle to the next, donor transparency produces what amount to “target lists” of likely contributors (for politicians) and sympathetic recipients (for both candidates and contributors with influence agendas).³⁹ Some forms of confidential contributions might be helpful, but “dark money” has obvious risks too.⁴⁰ Influence Market corruption – seemingly less damaging than other syndromes – may actually be among the toughest varieties to control and, because of wealthy countries’ economic clout, should be a global as well as a domestic concern.

These long-term, politicized, indirect reform scenarios will be unsatisfying to anyone looking for quick fixes via direct attacks on corrupt practices. The ideas here seek to replicate and gradually accelerate the political contention that enabled today's relatively well-governed peoples to check abuses of power. There is no guarantee of success: useful stalemates, in which contending parties gradually arrive at workable settlements that can become institutionalized, are more likely than civic breakthroughs; and as I argue above, solutions implemented in different polities have different results. Followers may lose heart; collective action problems may also set in. Therefore, reforms must be closely linked to citizens’ well-being. The indicators and benchmarks of government functions outlined above may help persuade citizens that they are benefiting from reform as they see improved and more fairly distributed services. They can also enable leaders and managers to claim some credit for progress, thus placing advocates of better government in a less adversarial relationship with ruling elites.

The punchline of this essay, in the end, is that when it comes to reform there is no punchline. No standard “toolkit” is likely to address corruption in all its forms. Until we trade whole-country perception ratings for evidence-driven assessments of trends in the quality of government we will understand neither the effects of corruption controls nor which practices might be “best” in a particular setting.⁴¹ At the same time, however, if we do pursue deep democratization, some cautionary tales should be kept in mind.

First, do no harm.⁴² Even seemingly promising reforms are likely to fail if they lack a solid base of political support grounded in lasting interests. Indeed, they may do more harm than good: not only to backers vulnerable to reprisals, but also to society as a whole if reform opportunities are lost; or if citizens, investors, and aid partners conclude that corruption cannot be challenged. Meanwhile, corrupt operators may decide that even a failed reform push means they had better take as much as they can – and take it quickly.⁴³ These forms of backlash must be considered when reforms are undertaken.

Trust and credibility are essential. Citizens have heard anticorruption appeals before, and have likely seen them fail. Indeed, they may well have seen corruption control used as a pretext for seizing power or for continued elite enrichment. In postconflict or deeply divided societies, citizens may distrust each other as much as corrupt officials, which further complicates collective action problems. But rather than targeting corruption in general, those in governance roles should listen to specific grievances – about, for example, poor utility services, health systems devoid of resources, police who work harder at collecting bribes than at protecting the public – to identify issues in which people have a shared stake. Demonstrable improvements in those areas can build the credibility of reform, reduce collective-action problems, and foster trust in the more honest officials. In that connection, “working with civil society” must involve the whole country as much as possible, reaching beyond the orbit of familiar ngos and into grassroots networks.

What you do, do well. As noted, credibility is a primary challenge for reformers, particularly in Official Moguls and Oligarchs-and-Clans situations. At the outset it is likely better to attain modest goals on a regular basis – and, of course, to call attention to accomplishments – than to proclaim massive campaigns that will once again come to naught. We cannot make “picking the low-hanging fruit” a permanent approach, but it is a first step toward building the credibility and relative strength needed to confront entrenched interests.

Establishing an anticorruption agency is not always a wise idea. Particularly in response to entrenched Moguls and Oligarchs, such organizations make the extensive resources and support needed by them hard to justify. Where moguls are in charge, the agency is vulnerable to capture from above, if it was not set up that way from the beginning. Where oligarchs slug it out, an aca may be ineffective in the face of their muscular networks, which can colonize the courts and law enforcement. The two most successful aca stories are those of Hong Kong's Independent Commission Against Corruption (icac) and Singapore's Corrupt Practices Investigation Bureau. But both of these societies are small and, in the case of Hong Kong, ethnically homogeneous, and well-positioned to capitalize upon low-corruption reputations. Most other societies are considerably larger, more diverse, and more economically differentiated, and thus face challenges on a greater scale. Moreover, since neither Hong Kong nor Singapore is a democracy, their acas face little opposition. By contrast, the excellent icac of New South Wales, Australia, has often had to fend off accusations of favoritism by one political party or another.

In the end, the deep democratization argument reminds us why corruption is worth worrying about in the first place: justice. Can people be governed – and ideally, govern themselves – in ways that are both effective and fair? Corruption is by no means the only reason why societies, even when outwardly successful, fall short of those ideals.⁴⁴ Still, the best way to link the grassroots and high-level parts of thean-ticorruption movement may be to harness political aspirations to broader social justice and set them as the guiding principles of reform thinking. Linking reform to aspirations to fair treatment by officials, secure property rights, responsive representation, and better public services could be the best way to mobilize lasting support for the long fight against corruption. Indeed, in some instances it might be advantageous not to discuss corruption at all, but rather to keep the focus on fairness, freedom, and human dignity. Demonstrating improvements in those areas, in turn, may be the most effective way to show that progress is being made and reform is for real.