Article body

But it has perhaps been wearisome to you, who, before I said a word, knew all about his venality. However, he calls it friendship and amity; and only just now he spoke of “the man who taunts me with the friendship of Alexander.” I taunt you with the friendship of Alexander! Where did you get it? How did you earn it? I am not out of my mind, and I would never call you the friend either of Philip or Alexander, unless we are to call a harvester or other hired laborer the friend of the man who pays him for his job. But it is not so. How could it be? Far from it! I call you Philip’s hireling of yesterday, and Alexander’s hireling of today, and so does every man in this Assembly.

Demosthenes[2]

In recent years, Dennis Thompson, Larry Lessig and others have articulated the concept of “institutional corruption,” characterized, on Lessig’s account, by improper organizational dependencies. In this paper, I propose to reveal a different face of institutional corruption, and its relevance for democracy and the rule of law.

The analysis proceeds through several steps. First, I argue that our existing concept of corruption in the individual context is actually two concepts. There is a moralized conception of corruption as transactional disloyalty, which I say is “moralized” because we ordinarily attribute moral blame to the agent who engages in actions that are corrupt in this sense; it is pro tanto blameworthy—it requires some novel argument to make any particular case not blameworthy. And there is a non-moralized conception of corruption as pollution, tainting by some external force; it is pro tanto not blameworthy. The disloyalty concept of corruption is about an agent’s actions, the tainting concept about the influences on an agent’s character. And the disloyalty and tainting conceptions of individual corruption trigger distinct social responses: someone who is corrupt in the disloyalty sense is punished, someone who is corrupt in the tainting sense is treated.

Second, I suggest that the existing scholarly literature contains a compelling non-moralized conception of institutional corruption, in the form of Lessig’s “dependence conception,” but does not contain a plausible moralized conception. I develop one, centering on the idea that an institutionally corrupt state in the moral sense is one whose institutions de jure or de facto permit public officials to treat the power of the state as their own.

Third, I explain the significance of moralized conceptions in the institutional context. While dependence corruption is a sufficient reason to carry out political reforms (the institutional equivalent to individual treatment), the presence of institutional corruption in the moralized sense suggests deeper failings in a democratic system: such corruption is inconsistent with popular sovereignty and the rule of law. A state that contains it to a substantial degree may lose its legitimate entitlement to power.

Individual corruption: Pollution or disloyalty?

The literature on the concept of corruption has been distorted (“corrupted,” even) by an equivocation between two notions that, in ordinary language, come under the same head. On the one hand is what I will call a tainting, or heteronomy, conception of corruption. On this conception, the corrupt individual has had his or her will or character twisted by some external influence; s/he has been polluted.[3] On the other is the notion of corrupt conduct, exemplified by bribe-taking, which is a form of disloyalty. Scholars have recognized these two ideas, but ordinarily suppose that they are both different faces of one thing;[4] here, I shall suggest that they’re really two things, which can, but need not, be present in the same person or transaction.

The tainting sense of corruption is the one at play in Lord Acton’s “power tends to corrupt, and absolute power corrupts absolutely.” Acton isn’t saying, here, that the one with power sells that power, in the bribe-taking sense of corruption, to the highest bidder. Indeed, the notion seems superfluous: if you have absolute power, you don’t need to trade favors for money. You have soldiers to get you money. Unsurprisingly, the tyrant through history, from the Thirty in Athens through Hitler, has preferred to expropriate rather than bargain. Rather, Lord Acton is saying that the soul of the one with power is polluted, warped, by that power.

This tainting sense of the notion of corruption dominates book 9 of Plato’s Republic: the best government devolves into oligarchy in virtue of breeding errors (!) which allow inferior characters to enter into power. Notice that Plato’s story moves from the individual corruption of the souls of those who take office to the institutional corruption of the state under their rule—a causal sequence that is typical of Republic’s famous state-soul analogy, but is not unique to Plato. Similar accounts of the pollution of political communities and the souls of those who lead them can be found in book 8 of Montesquieu’s Spirit of the Laws and in Machiavelli’s Discourses.

Money often features in these accounts (as in Plato’s), but when it does, it’s the corruption of the soul by the desire for money, which leads to all kinds of money-motivated misconduct, not just the bribe-taking associated with the second sense of corruption. And things other than money can also drive this kind of corruption. For example, we might describe a heroin addict as corrupted by the drug: her character and motivational structure have been warped by the desire for the drug; as a result, her will is no longer her own: she has become profoundly heteronomous.

Importantly, there need not be anything morally blameworthy about this kind of corruption. One’s soul might become corrupted through no fault of one’s own. For example, we often say that a child is “spoiled” (that is, corrupted) by overly permissive or protective parents. This kind of corruption might, on the folk psychological account of the process, ruin the child’s character later in life, making him weak, needy, dependent, demanding—but this isn’t really the child’s fault: it’s something someone else did to him.[5] And the appropriate response to the child or adult who has been thus corrupted (as with the heroin addict) isn’t punishment but therapy, treatment, an organized effort to remove the corrupting influence from his soul.

The disloyalty sense of corruption should be, but often is not, kept distinct from the tainting sense.[6] The “corrupt cop” is disloyal, as are Hesiod’s βασιληας δωροφάγος[7] who issue corrupt judgments. Neither need be polluted; we’re not primarily concerned with the state of their souls but with the wickedness of their behavior. The quintessential form of disloyal corruption is government bribe-taking: an official betrays his public duties for money. (In the next subsection I will generalize this conception to other forms of what I will call “transactional disloyalty,” but for now, let us assume we just mean bribe-taking.) And that behavior need not represent anything about his character as a whole: a cop might succumb to the temptation to take just one bribe, and otherwise be an exemplary citizen in her public and private life.

This kind of corruption is always a pro tanto wrongful act. By “pro tanto wrongful,” I mean that ordinarily we will say that the bribe-taking sort of corruption is morally blamable. There might be reasons to surrender that judgment in an individual case—the blamability of bribe-taking corruption is defeasible—but the burden of persuasion is on the person who wants to redeem the act. And, in consequence, the remedy for bribe-taking corruption isn’t treatment or therapy, it’s punishment. The corrupt cop hasn’t had the autonomy of his will overthrown by some kind of external influence. Bribe taking is an intentional act: we ordinarily attribute moral responsibility to those who do it.

Of course, the two flavors of corruption aren’t wholly independent. It’s very plausible to think that someone’s character might be corrupted by a habit of bribe-taking (an implication of the general Aristotelian proposition that one’s character is primarily composed of one’s habitual behaviors); it’s also plausible to think that someone with a twisted soul is more likely to take bribes.[8] And we might easily think that a community or organization as a whole could be warped by a widespread practice of bribe-taking.[9] But these are contingent, rather than necessary, facts about any particular person or social arrangement. It is possible to take just one bribe; it is also possible to be twisted in ways that don’t implicate the notions of bribe-taking or disloyalty. Yet scholars routinely merge the two ideas together; Underkuffler, to give the most recent example, gives a conception of corruption as a whole as a form of tainted character that always comes attached to moral blame and captures the intuitive core of corrupt acts like bribe-taking.[10] But this is a mistake: twistedness may be (contingently) causally related to blameworthy acts including, but not limited to, bribe-taking, but they are conceptually distinct and there is no necessary relationship between the two.[11]

In the remainder of the first half of this paper, I set aside the tainting kind of corruption to draw out the individual version of the disloyalty kind. The object of the exercise is to provide a base for discussing, in the second half of the paper, how disloyalty-corruption might be transposed to the institutional context.

Moralized individual corruption as transactional disloyalty

Begin with a paradigm case of individual corruption in the bribe-taking sense.[12] Alice is an income tax collector employed by her government. She learns that Barry has lied about his income, and calls him in for an audit. Barry happens to be the director of admissions at Yale University, and he offers to see to it that Alice’s son is admitted to the freshman class if she buries the evidence of his tax evasion. Alice agrees.

Obviously, Alice’s behavior is corrupt. (So is Barry’s, but I will focus on Alice.) She has sold the powers of her official position for personal gain. But what’s wrong with it?

I would like to suggest that Alice has committed a kind of theft. She’s stolen the property of the government, in the sense that she’s taken her official powers, which are entrusted to her for public purposes and in pursuit of the public good, and converted them to her own use. She treats, that is, her official powers as her own property.

To see that this is an accurate expression of our intuitions about what’s wrong with Alice’s bribe-taking, compare her to Carol. Carol is not an employee of the tax collection agency, she’s a tax farmer, like those in use in Ancien régime France and many other countries on the Continent up to the 18th century.[13] She has purchased the right to collect taxes from the government, and is entitled by the terms of that purchase to keep what she can collect. When Carol meets Barry and agrees to forget about Barry’s tax debt in exchange for preferential university admissions, she is, intuitively, not doing anything wrong. The power to collect taxes from Barry isn’t someone else’s, it’s her personal right, which she has purchased, fair and square, from the government. That power being her personal property, she’s free to trade it however she pleases in the marketplace, and if her son’s admission to Yale is worth more to her than the amount of money that she would otherwise receive from Barry, there are no moral or legal reasons for her not to take the deal.[14]

With this intuition tentatively in place, it is now necessary to further interrogate the idea of using someone else’s property as one’s own. What is it about Alice’s bribe-taking that feels so much like theft?

We can easily isolate several things that are irrelevant. First, it’s not relevant that Alice receives a personal benefit from her bribe-taking; we can easily substitute a case where what she receives is wholly altruistic (“forget about my tax fraud and I’ll go volunteer in this hospital for a year”)[15]. Second, the particular form of currency that Alice receives isn’t relevant: she could get money, she could get college admission for her child, she could have a building named after her—it’s still corruption.

Rather, what seems to be key here is the kind of relationship that Alice has with the government, and with the official powers that the government has delegated (not granted) to her. In legal (and economic) jargon, Alice is an agent, and the government is the principal. The defining feature of a principal-agent relationship is that the principal gives the agent authority—the power to carry out actions in the principal’s name—which is to be used for the principal.

The key idea is “for the principal”: the authority delegated to the agent is not the agent’s to use however she or he wants. “For the principal” encompasses a wide but not limitless scope of variation.[16] Sometimes “for the principal” means “according to the principal’s commands”—in the most basic kind of employer-employee relationship, the duty of the employee is just to do what he’s told—and those commands can be more or less abstract, depending on the extent to which the principal is to use her own judgment to carry out the employer’s stated goals (“maximize the value of my stock portfolio”), or just to complete discrete tasks (“pick these berries and carry them to the barn”). Other times, “for the principal” means “in order to serve ends that we attribute to the principal” or “in the principal’s best interests,” regardless of what the principal says she or he wants (or when the principal is an entity that can’t give commands except through agents, like a corporation)—for example, we can conceive of the guardian of a child or an incompetent person as an agent of that person, delegated the authority to conduct the principal’s affairs according to the agent’s best judgment of the principal’s interests.[17]

Whatever “for the principal” means in a given case, it clearly doesn’t mean “for the satisfaction of the personal preferences, interests, or whims of the agent.” In this sense, an agent who uses her authority for her own ends, even if they’re altruistic, does engage in a kind of stealing: she takes something rightfully owned (in the sense that the owner of something is entitled to receive the benefits of that thing) by another and uses it to serve ends not encompassed by the true owner. For example, in ordinary English, it’s admissible to say that an employee who starts an internet side-business from the office, using his employer’s equipment and during time for which he is being paid a salary, has stolen the time and the use of the equipment from his employer.

So far, it looks like the principal-agent relationship is all burden for the agent and no benefit; it even looks like I’ve said that the agent is a thief if she manages to get any personal benefit out of the relationship. But this isn’t right: obviously, nobody would ever willingly serve as an agent if the relationship were so one-sided. And this is why, typically, agency relationships are established by contract, for compensation. The principal and the agent agree on the ways in which the agent’s preferences are to be served by the relationship, and these are the incentive to get the agent to assume an obligation of loyalty to the principal. Compensation comes in many forms (including, inter alia, salaries, tips from third parties, commissions, perks like the personal use of a company car, etc.), and is sometimes implicit, drawing on shared social customs, general legal ideas, practices in a given industry, etc.

We now have a schematic description of the principal-agent relationship: a principal delegates authority to an agent, where the authority is to be used for the principal, and, in exchange for taking on the obligation to act for the principal, the agent receives compensation, which is specified by the terms of the agreement, or by shared norms governing the parties.

That schematic allows us to see where Alice went wrong in our initial example: she didn’t use her authority for her principal, the government. She used it for herself (or her son), without the (actual or attributed) agreement of the principal per her compensation scheme.

And that is what I mean by saying that Alice used the authority of the government as her own. If Alice, like Carol, owned the right to receive taxation, she would be entitled to use it to serve her own preferences, whatever those preferences might be (absent some other principle barring it). The extent to which Carol is allowed to serve her own preferences with the power to collect taxes isn’t specified by some preexisting compensation agreement, but is importantly open-ended, and that’s what it means to have some power for one’s own: one has an open-ended entitlement to use it in whatever way one wants. So, the distinction between Alice and Carol is that Alice has a power which is only to be used either to serve the government, or to obtain compensation that the government’s agreed to; Carol has a general-use power.

The conversion of a principal’s power—the use of one’s principal’s property or authority as one’s own—appears to be a necessary condition for classifying some act as corrupt in the disloyalty sense. We can test this against the most classic cases of corruption. A football player who takes a kickback from a gambler to throw the game counts as corrupt because she’s converting the authority of her license to participate on the team, which was delegated to her so that she may pursue the team’s interest in winning. Similarly, a legislator who sells his vote is corrupt, because he uses the authority delegated to him by the demos, but not for the demos. On the other hand, a politician who changes his legislative vote in order to win electoral votes is not corrupt, because an implicit condition of the agency agreement between the legislator and the demos is that the legislator is allowed to use his vote to win the support of constituents, and hence get compensation in the form of the status and salary of the office—it’s a use of authority contemplated by the relationship.

However, such conversion can’t be a sufficient condition. There are many cases of agents converting their authority that we would not describe as corruption. For example, the delivery person who takes the company vehicle to go drag racing clearly is converting his authorization to use his employer’s property, but isn’t engaging in corruption. The same goes for the police officer who abuses her public powers by harassing her personal enemies. Both of those wrongdoers are offending against morality as well as (in every state of which I’m aware) law, but they’re not engaging in corruption. In our ordinary use of the term, to mere malfeasance of office, bribe-taking as the quintessential case of corruption adds a relational element: the corrupt agent exchanges his relationship with his principal for a relationship with someone else. And the way in which the corrupt agent uses the property of his principal as his own is transactional: she makes use of the principal’s authority to conduct a personal transaction with some third party, a transaction for himself or for someone else rather than for the principal; in doing so, his loyalty to his principal is undermined and at least partially transferred to that other person.

This final idea gives us individual corruption as disloyalty: a property of the actions of agents, such that an agent’s action is disloyalty-corrupt when 1) she treats the authority delegated to her by her principal as her own, by using that authority other than for the principal or in pursuit of the agent’s compensation as ordinarily understood by the principal and agent to be built into the relationship, and 2) does so as part of a transaction with some third party in which she betrays her loyalty to the principal.[18] This kind of corruption is pro tanto morally wrong, because we value loyalty and suppose that it is a virtue, but any specific act of corruption may be more or less wrong, depending on the moral value of the agency relationship that has been betrayed: a public official who violates the trust of a legitimate democracy or a doctor who takes bribes to neglect her patients is extremely blameworthy; a mafia hit man who becomes a snitch for money may not be blameworthy at all.

Of course, this isn’t terribly novel material. Scholars have been debating what exactly corruption is for decades. Fortunately for me, Laura Underkuffler has written a book comprehensively reviewing the major streams of argument and criticizing them.[19] I will take a moment to discuss how this conception avoids her main lines of criticism.

Against conceptions of corruption that build the moral idea of corruption out of the violation of legal entitlements, Underkuffler points out that we can have violations of legal entitlements that are not wrongful, if the law itself is morally objectionable.[20] Even though my account of corruption draws on legal concepts, and depends on the violation of some entitlement (of the principal, to having the authority she delegates used for her), it is not subject to this objection, because nothing in my account requires that the entitlements be legal entitlements. We can conceive of a legal kind of corruption, in which an agent uses the principal’s authority as her own, and we say that the authority in question is “the principal’s” because the law designates it as such. Such law-corruption is pro tanto wrongful to the extent legal entitlements track forms of loyalty that we have reason to value, such as promise-keeping, loyalty to employers or governments, etc. Alternatively, we can conceive of a moral kind of corruption, in which an agent uses the principal’s authority as her own, and we say that the authority is rightfully the principal’s because the principal has a moral entitlement to it. Such moral corruption may or may not be against the law.[21] On the conception articulated in this paper, disloyalty-corruption is a form of transaction and relationship that can be discovered in the moral domain or the legal.[22]

Moreover, while Underkuffler is right to argue that the concept of (disloyalty-) corruption inherently implies moral condemnation, she misses that this condemnation is merely pro tanto: the corruption of an act always gives some reason not to do that act, but the reason can be outweighed by other moral reasons in appropriate situations.[23] To see the defeasibility of the moral condemnation of disloyalty-corruption, consider an extreme example. Suppose a Nazi official takes a bribe from a Jewish family to free them from a death camp. The act is corrupt. It has to be. Bribe-taking by a public official is the central case of corruption; if anything is corrupt, that is. But being corrupt doesn’t make it evil. Because Nazi Germany can’t generate legitimate claims to its officials’ loyalty, and because of the overriding moral importance of not participating in genocide, the bribed Nazi’s corruption is positively laudable. The pro tanto wrongness of corruption has been totally overridden.

Yet, Underkuffler might object, that example actually seems to suggest that I’m wrong to say that disloyalty corruption is only pro tanto blamable. After all, intuitively, surely the official who only releases Jews when bribed does something wrong. To answer this point, I submit that even if one shares that intuition, we can explain it from corruption-independent moral principles, such that it does not require the conclusion that corruption is more than pro tanto blamable.

Consider an alternate case: the whimsical Nazi. The whimsical Nazi frees Jews for wholly arbitrary reasons, say, by rolling dice. The whimsical Nazi isn’t corrupt on my definition (there’s no transactional element), and doesn’t look like a bribe-taker. Yet we should still have a strong intuition that the whimsical Nazi is doing wrong, and we can explain it from all the major moral perspectives and see that it is the same intuition that applies to the bribed Nazi. For a utilitarian, it’s not wrong for the whimsical Nazi to free the lucky Jews or the bribed Nazi to free Jews who pay; the real wrong is the failure to free the unlucky or non-bribing Jews. For a Kantian, the bribed and whimsical Nazis free the Jews for the wrong reasons: they’re supposed to free them from a recognition that it’s morally wrong to murder them, not from pecuniary or random motives. Similarly, for a Humean or virtue ethicist, the bribed and whimsical Nazis are supposed to free the Jews out of a sympathy for the wrong done them, not out of avarice or caprice. None of these general moral critiques of the bribed and whimsical Nazis have anything to do with corruption, and do nothing to undermine the proposition that the bribed Nazis corruption, standing alone, is virtuous.

We now have the analytic tools in hand to put the tainting/disloyalty typology to work in understanding institutional corruption.

Institutional corruption: Extending the typology

In recent years, there has been growing interest in a notion of “institutional corruption,” a kind of corruption that is a property not of individuals but of political, economic, or social organizations. A legislature like the U.S. Congress, for example, might be said to be institutionally corrupt in virtue of the relationship between campaign finance and the ability to get elected, and this might be true independent of whether we can say that any individual member of or candidate for Congress is corrupt.

The best discussions of institutional corruption, however, appear to be solely of the tainting sort. As yet, I have been unable to find a plausible version of institutional corruption equivalent to the disloyalty conception which I’ve described for the individual case. In the remainder of this paper, I will first discuss the leading account of institutional corruption as tainting, Lessig’s “improper dependence” conception, and then give an account of institutional corruption as disloyalty. I will then criticize Dennis Thompson’s version of institutional corruption, which appears to draw from ideas of disloyalty, but does not do so convincingly. Finally, I’ll describe the moral significance of disloyalty-corruption in the institutional domain.

Non-moralized institutional corruption: The dependence conception

Lessig has, over the last several years, been advancing a “dependence conception” of institutional corruption. For Lessig, an institution is corrupt when it or its members have become dependent on the wrong thing. Congress has become dependent on funders, rather than on the people, therefore, Congress is corrupt.[24] Elsewhere, Lessig defined institutional corruption as follows:

a systemic and strategic influence which is legal, or even currently ethical, that undermines the institution’s effectiveness by diverting it from its purpose or weakening its ability to achieve its purpose, including, to the extent relevant to its purpose, weakening either the public’s trust in that institution or the institution’s inherent trustworthiness.[25]

I take these two definitions to be cover essentially similar territory: the “systematic and strategic influence” of the second is the improper dependence of the first, and the purpose of the second is the proper dependence, the one that an institution ought to have, of the first.

Lessig’s conception is clearly equivalent to what in the individual domain I have called tainting-corruption. The idea of dependence is closely intertwined with that of tainting: pollutants corrupt the motivational structure of an individual as well as of an institution by generating dependence. Thus, the heroin addict example is tainting-corrupt in virtue of her dependence on heroin; it is the dependence that warps her character, and that subverts her will—it is that which makes her heteronomous. Similarly, Tsarina Alexandra was corrupted when she became dependent on Rasputin as a political, religious, and personal adviser, rather than on her own judgment or the advice of her non-evil advisers. The vain have become corrupted because they are addicted to the praise and admiration of others. And so forth.

Appropriately for a tainting conception, Lessig’s conception of institutional corruption depends on a preexisting story about what an uncorrupted institution would look like, in the form of the requirement that the institution’s effectiveness (i.e., its actual purpose, the content of its true collective will unaffected by the external, heteronomy-producing influence, its proper dependence) actually is undermined. Moreover, Lessig’s conception seems to do a very good job at explaining our intuitions for concrete cases: the role of campaign finance in a legislature, or the role of gift-giving in pharmaceutical marketing, two of the main cases of institutional corruption on Lessig’s account, do seem to have this sense of pollution: these practices are an insidious influence that taint the essence of the institutions in question.[26]

To see that the dependency corruption is not a moralized conception, and does not depend on disloyalty, consider that an institution can be dependency corrupt where individual agents are not doing anything even pro tanto wrong. Lessig gives as an example of institutional corruption the improper dependency of a judicial system on judicial ideologies,[27] but this dependency can exist without any individual judge doing anything morally suspect. Accordingly, we might think, in accordance with the suggestion made at the beginning of this paper in the individual context, that the appropriate social response to dependence corruption is something like treatment: the corrupting influence of campaign finance, or judicial ideology, or pharmaceutical research support, ought to be purged from the system; being purged, the institution can go on, with no moral critique addressed to either people or organizations.

Moralized institutional corruption

However, there is conceptual space for a form of institutional corruption in the moralized, disloyalty, sense. Here, I describe one.

Elsewhere, I have shown that the rule of law is a property of a state’s legal systems, and one of its requirements is regularity: a state that comports with regularity is one in which officials use their power only under the constraint of reasonably specific legal rules.[28] In a regular state, officials maintain a role separation between their institutional roles and their personal identities: they set down their official powers when they leave the job; when they’re on the job, they exercise their powers as given by law, for the ends encompassed by the law, when they clock out, they’re free to use their personal resources to pursue their personal ends.[29] Perhaps the exemplar of the official under the rule of law is the Anglo-American judge, whose institutional role is most strictly separate from her personal identity, a separation enforced not only by legal and ethical demands such as recusal requirements, but even by ritual: a judge puts on her role, and her impartiality, with a robe (and, traditionally in some countries, a wig), a symbolic reaffirmation and reminder of the difference between the judge’s ordinary life and her life within the law.

By contrast, when regularity fails, this role separation does not obtain. Officials no longer are constrained to treat their powers as a trust to be used in accordance with the law. Instead, the institutions of the state permit officials to use their powers as part of their personal endowments (either by not making this use illegal, or not enforcing or being able to enforce such laws as exist), to be used however they want. Judges are allowed to rule for their friends, police officers are allowed to harass their enemies. That is, officials are allowed to treat their powers, delegated to them by the state and for the state’s purposes, as their own personal property, for their own purposes.

It now becomes clear that individual disloyalty corruption and failures of regularity have the same basic feature of conversion of authority. And this gives us a candidate definition of institutional disloyalty corruption in a state: institutional corruption in a state is one way in which regularity can fail, characterized by officials being permitted by the de jure or de facto norms of the state to use their power as their own personal property in transactions with others. Institutional corruption is just a failure of the rule of law in transactional form.

The contrast between institutional and individual disloyalty corruption, then, is one of norms. Individual corruption violates some kind of meaningful legal, social, or economic (contractual) norm: an individually corrupt agent is one who not only converts his authority and switches his loyalties from his principal to someone else, but who violates the expectations embedded into his agency relationship by doing so. By contrast, in a case of institutional corruption, the agency relationship—typically with the state, although I will consider below cases of non-state institutional corruption—itself is defective, such that agents are permitted by the norms that govern it to convert their powers and switch their loyalties.

On this definition, there are two quintessential cases of institutional disloyalty-corruption. The first is a state where the laws make no attempt to control the transactional conversion of public power, such as a dictatorship in which officials are permitted to use their powers to extract money and favors from the populace. The second is a state that formally defines the duties of its officials, but where, in reality, the laws are so ineffective at controlling them that officials routinely sell their powers for a price; this sale has become the new norm within the state.

Here, however, a problem with this understanding begins to appear. The notion of a conversion of authority presupposes some preexisting standard to determine for whom the authority in question is to be used (that is, who is the principal), and what constitutes using that authority for the principal. In the case of individual corruption, that standard can be supplied by preexisting legal, social, or economic/contractual norms. But institutional disloyalty-corruption is defined in part by the absence of effective norms. To understand institutional disloyalty-corruption, we must have the concept of an agent who is allowed to use her power in a way that is nonetheless unauthorized.

Allowedness can be a purely positive notion: to say that A is allowed to X might just mean that nothing is stopping A from X. For example, we can say that General McClellan allowed the Southern army to escape after the Battle of Antietam; by this we would merely mean that McClellan failed to stop the Southern army from escaping, even though he could have done so. By contrast, authorization is most naturally a normative notion. To say that A is authorized to X carries with it the implication that there is some evaluative standard (which need not be all-things-considered morality) according to which it is right that A X. Thus, it makes sense for ordinary users of English to say that the Southern army was allowed, but not authorized, to escape. The difference depends on the existence of an evaluative standard (norm).

Where there are existing de jure norms regulating the use of official power, but those norms are toothless and ignored, the norm against which we judge official power to be converted without authorization can be internal, drawn from the ambitions that the law declares, but fails to meet. We might also draw such internal norms from the non-legal ambitions of a political community. For example, if there are documents, historical declarations, or cultural or religious ideals that express the guiding purposes of a political community, as for example the U.S. Declaration of Independence and preamble to the Constitution, we might say that a set of operational social or legal norms that allow officials to use their power inconsistent with those ideals is institutionally corrupt. However, where no such internal norms are available, we must appeal to external norms, such as moral principles or general ideas about what government power is for and how it ought to work. In such cases, the idea of institutional corruption as disloyalty may have no independent normative content: it may simply refer to a norm of behavior that violates independent moral principles governing the way an entity ought to regulate its agents. When the state is involved, often this independent principle will be the rule of law, which is the moral norm governing the way that states ought to control their officials.[30] A state that gives officials too much discretion violates the principle of regularity, in doing so, it may also allow officials treat the power delegated to them as their own, in which case it will be institutionally corrupt in the external sense.[31]

Fully defined, then, institutional disloyalty-corruption is a species of norm-conflict. A property of systems of norms, it occurs when agents are allowed by existing de jure or de facto norms to treat the authority delegated to them by their principals as their own, by using that authority other than for the principal or in pursuit of the agent’s compensation (as in individual disloyalty-corruption), and as part of loyalty-shifting transactions (as in individual disloyalty-corruption), where those norms conflict with other moral, legal, social, or economic-contractual norms applying to the principal-agent relationship that indicate that the use to which the agent is putting her authority is not for her principal.

An alternative moralized conception?

Contrast the loyalty conception given with that of Thompson. Thompson argues that the difference between institutional and individual corruption is that in the former, “the benefit an official receives is political rather than formal, the service the official provides is systematic rather than episodic, and the connection between the benefit and the service manifests a tendency that disregards the democratic purpose,” concluding that “institutional corruption occurs when an institution or its agent receives a benefit that is directly useful to performing an institutional function, and systematically provides a service to the benefactor under conditions that tend to undermine legitimate procedures of the institution.”[32] The transactional element seems to suggest that Thompson is aiming at something like a disloyalty conception.

However, taken as a disloyalty conception, Thompson’s definition of institutional corruption seems too broad. Consider the practice of legislators making use of briefing papers prepared by lobbyists. Under some circumstances, this would meet Thompson’s definition: legislators receive a benefit (information) directly useful to performing their legislative functions, they systematically provide a service to the lobbyists in the form of an audience for their views as presented in those briefing papers, and, to the extent lobbyist-provided information drowns out other information, the practice appears to undermine legitimate procedures of a legislature by undermining the extent to which constituents have access to their representatives on equal terms, as well as undermining legislators’ access to objective and unbiased information.

Doubtless there are things to be said against the briefing paper system. By distorting the information legislators receive, it may bring about worse public policies; by making it easier for special interests to be heard than ordinary citizens, it may make the system less democratic. Yet it seems strained to describe these vices as “corruption.” The intuitive judgment that briefing papers aren’t corrupt proceeds, I submit, from the fact that the state retains legislators’ undivided loyalties: the “benefit” received by a legislator is the ability to do her job more efficiently, she accepts this benefit in order to achieve that end, and there’s no reason to think that the provision of briefing papers in any way obligates legislators to provide special advantages to lobbyists. The special advantages lobbyists happen to receive from legislators as a result of their providing briefing papers are just the incidental consequences of their having made themselves useful to the government in a personally beneficial way, not the object of some kind of quid pro quo.

Moreover, Thompson’s definition requires bad consequences (“tends to undermine the legitimate purposes”). But our ordinary notions of disloyalty-corruption depend on the character of the act, not on its consequences. We would ordinarily say, for example, that an institutionalized system of bribery that citizens and officials use to evade an ossified and inefficient bureaucracy is a form of corruption, even though it brings about good consequences. It’s not clear that Thompson’s account has the resources to explain that judgment.

My definition avoids these problems. It accounts for the non-corruption of the briefing paper system, because it requires transactions and shifting loyalties. And it depends only on the character of an act, rather than its consequences[33].

Concluding remarks: The moral crisis of corruption

Let us return for a moment to the case of the U.S. Congress. Lessig has given us a compelling account of how it is corrupt in the non-moralized sense, and thus a suitable candidate for treatment—for the removal of the polluting influence of money. But we can also give a case for Congress being corrupt in the moralized, disloyalty, sense, and thus a suitable candidate for moral condemnation.

Regularity is not the only principle of the rule of law. Another is generality, which requires that the positive law apply to all citizens on equal terms, or, more precisely, treat all citizens as equals.[34] While the principle of generality applies to all states, it has particular importance in democracies, which come prepackaged with the notion that each citizen is to be equal under law. Thicker theories of democracy, such as Rousseau’s, have the notion of general law built right in, in the form of the claim that the general will cannot speak except through general laws—that is, that a mode of government that depends on the notion that each law is the will of all citizens must necessarily presuppose that each law is justifiable by reasons that include each citizen.[35]

The principle of generality can give the external normative standard against which we can see institutional disloyalty-corruption in a democracy. A legislator who is individually disloyalty-corrupt betrays the public by making law that is not general—that fails to treat citizens as equals, that cannot be defended in terms of reasons that are applicable to the population as a whole—in the context of corrupting transactions. And a state that permits such lawmaking is institutionally disloyalty-corrupt.

The difference between the applications of the dependency conception of corruption and disloyalty-corruption is that disloyalty-corruption requires the laws actually be bought: legislators must actually make deals leading to non-general laws; the mere influence of campaign finance on elections is insufficient. For example, suppose we lived in a (bizarre) state in which the wealthy cared about exactly one political issue: how physically attractive their representatives happened to be. In this world, they would donate lots of money to the beautiful politicians and none to the ugly; only the beautiful would win what Lessig has called the “funders election” and have a chance to go before the people. Assuming that the beautiful aren’t systematically different from ordinary people, such a world would be characterized by an improper dependency, just as ours, but that dependency would have no political effect: politicians would feel no need to do anything special in order to appeal to the funders, their political positions would not change as a result of the funders election, and their loyalty would continue to lie entirely with the people. In such a world, Congress would count as corrupt under the dependency conception of institutional corruption but not the disloyalty conception.

Importantly, such a state could still count as a Rousseauian democracy, since the laws could still be aimed at the public good. For the same reason, such a state could still comport with the rule of law. By contrast, in a state in which Congress actually makes non-general laws as a result of being bought, the laws are no longer general, and it is no longer a Rousseauian democracy. Bought legislators are no longer loyal to the people: those whose interests are disregarded (and there must be some in order for us to say that law is not general) are no longer treated as part of the polis to which the bought legislator is accountable.[36] Similarly, a state in which judges and police are bought fails by the standards of democracy in addition to the rule of law: being unconstrained by law, the bribed officials no longer serve the sovereign demos, which no longer controls the power exercised by its officials.[37]

It is for that reason that a disloyalty-corrupt state warrants moral condemnation as well as reform.[38] A state that is institutionally corrupt in the dependency sense is merely broken. It is likely to function sub-optimally; we probably won’t get politicians as good or legislation as wise as we might get otherwise. Such corruption calls for reform. A state that is institutionally corrupt in the disloyalty sense is illegitimate. Failing by the standards of both democracy and the rule of law, such a state presents the public with an urgent moral crisis.