Last week Matthew wrote that too much time and energy has been wasted trying to define corruption. While I agree, I don’t think sufficient attention has been paid to why so many spend so much time arguing about what “corruption” means. Matthew pointed to the reason in one of the first posts on this blog but stopped just short of the explanation. Let me take the last, short step in the hopes it will end the interminable, unproductive wrangling over definitions.
In the earlier post Matthew wrote that corruption “implies a deviation from some ideal state” and hence “involves an implicit or explicit selection of a baseline standard of ‘correct’ behavior.” He went on to explain that in the corruption literature the three most common baselines are the law, public opinion, and public interest. “Corruption” can then be conduct that deviates from what the law provides, that diverges from what the public thinks is wrongful, or that is at odds with the public interest. The definition of “corruption” depends upon the baseline; each baseline, as Matthew explained, leads to a different approach to defining “corruption.”
With one slight emendation, Matthew is correct. But I think making that correction is important for taking the analysis the next step and, I hope, ending, or at least reducing, fruitless debate over definitions.
Matthew wrote in his first post on definition that “’corruption’ implies a deviation from an ideal state.” Not quite. As its etymology shows, “corruption” does not imply a deviation from an ideal. Rather, deviation from an ideal is the core meaning of term. The original Latin, and the Old French word from which the term “corruption” passed into English, both convey the idea of spoliation or impurity. That sense is most clearly captured in one of the secondary definitions of “corruption” in modern English. When a computer file is damaged in transmission, the file is said to be “corrupt” or “corrupted.”
This secondary meaning of the term shows why an agreed upon definition of corruption remains beyond our reach. If an Adobe Acrobat file sent across the internet won’t open, it means the file is not in its ideal state. Computer users don’t debate whether the file is corrupt or not. They have a clear standard against for deciding the question: Adobe Reader can’t open it. Hence the file is corrupt. Period.
No “Adobe test” exists for determining whether a polity is corrupt because there is no agreement on the characteristics of an ideal polity. The continuing furor over Plato’s vision of the ideal polity is one reminder; a contemporary one is the interminable debate in the United States about the financing of political campaigns. One vision of the ideal is the current system: anyone can speak as much or as loudly about issues of the day as they please and spend as much they can afford advocating the election of candidates they choose: truth will out. Another is one where no citizen should have more influence than the next: equality first. Many proponents of the latter view claim that because current American policy deviates from their vision of the ideal state, the American polity is corrupt.
The point is that at bottom “corruption” is a normative term. This normativeness is sometimes masked because in many instances there is no debate about what the ideal is. No one disagrees, or at least no longer disagrees, that in the ideal polity public servants should make decisions free from the influence of secret payments by interested parties. Hence there is no argument that bribery is “corrupt.” But in many areas — campaign finance in the U.S., gift giving in other nations – no consensus on the ideal has been reached; no Adobe test has yet been developed. And until one is, or more accurately emerges as a result of an agreement on the underlying values at stake, argument about definitions will continue.
I think it would be to everyone’s benefit if arguments about definitions were re-framed as arguments about what the ideal state should be. It would first of all help to cool the rhetoric if the politically loaded charge of “corruption” were wielded less often. It would also improve the quality of the debate. Rather than defining some policy or conduct as “corrupt,” those seeking change would do better by showing why the ideal state they prefer is superior to the prevailing one. Who knows — casting the issue this way might even enhance their chances of bringing the change they seek.
GAB | The Global Anticorruption Blog 
Hmm. There’s much in here that I agree with, but I think I disagree with your bottom line, for two reasons.
First, as I emphasized in my last post (in slightly different words), although corruption always implies deviation from some normative baseline, not every deviation from that normative baseline is necessarily helpfully categorized as corruption. I do think there is an analytically useful category of “corruption” (even if its boundaries are inevitably fuzzy), and like you I don’t think we want to throw around the term corruption to broadly. So, in my ideal state, all public officials would be really smart and highly competent at their jobs. A world in which many public officials are not so bright, and fairly incompetent, is a pretty big deviation from my ideal state, but I tend to think it’s not so useful to think about this as a kind of “corruption.”
Second, I tend to disagree with the idea that we should try to define corruption by fully specifying our ideal state, precisely because we’re extremely unlikely ever to achieve consensus on what the ideal state looks like, but we might be able to achieve a fairly broad degree of agreement (if not full consensus) on what sorts of conduct ARE corrupt.
Third, another standard difficulty with a “public interest” approach to defining corruption (an approach you seem to endorse in your post) is that it limits our ability to ask objective social science questions about whether corruption is harmful (because on the public interest account corruption is harmful by definition–though I suppose we can still ask certain Theory of the Second Best-type questions about whether it might compensate for some other political imperfection.)
So what’s my preferred approach? I don’t really know — as I said, I try not to think too much about super-precise definitions. But my default is to start with the standard “abuse of entrusted power for private gain” idea, then identify a series of practices that would likely be considered corrupt under any of the major approaches to defining corruption: bribery, embezzlement, etc.,; we can focus on those as the “core” of the phenomenon, and recognize that there are some fuzzy borderline cases that may be in or out depending on the context.
One person that has thought long and hard about this issue is Mark Philp. I won’t repeat his arguments here. They should be read in their entirety (not violated in a short blog) but I like his point that we should be able to distinguish between a crime, of whatever sorts, and corruption, or whatever sorts. Corruption should not be confused with ordinary criminal acts or misdemeanours. Corruption is political, something that distorts the political process or system, and the public interest dimension is therefore inherent to the definition (although Philp also recognises that there is a principal-agent element to a corrupt transaction). He has proposed a definition that is unfortunately not elegant (short) enough to replace the standard definition we operate under: “Corruption in politics occur where a public official (A), acting in ways that violate the rules and norms of office, and that involves personal, partisan or sectional gain, harms the interests of the public (B) (or some sub-section there-of) who is the designated beneficiary of that office, to benefit themselves and/or a third party (C) who rewards or otherwise incentivises A to gain access to goods or services they would not otherwise obtain.”
I won’t go into why the standard definition has some serious drawbacks when people try to operationalise it, but at the core is the importance of being able to distinguish between ordinary criminal acts and corruption.
Yes, I’m familiar with the attempts of Philp (and others) to try to come up with a more refined definition of corruption, as well as the arguments for a “public interest”-based definition. I remain unconvinced that this definitional enterprise is worth the amount of intellectual effort that seems to have been poured into it. Just to illustrate, I’m not clear on why the Philp definition that you quote is so much more helpful, when it comes to operationalizing the concept, than the standard, rough-and-ready “abuse of entrusted power for private gain”. There’s some extra verbiage, but not in my view a lot of additional precision or clarity.
I also confess confusion as to why it is so important to distinguish “ordinary” criminal acts and corruption. Much (though not all) of what we ordinarily think of as corruption is a crime in most systems; much (most) criminal conduct would not be considered corruption. I also agree that we generally use the term “corruption” only to describe a certain subset of acts, generally those that involve an agent using authority vested in her to benefit herself rather than the her principal.
I remain of the view that it’s not all that helpful to continue to run around in definitional circles. All the basic approaches to defining the concept of corruption, and the problems with each, have been fairly well known for 40 years. We all basically know what we’re talking about — primarily bribery and embezzlement, and as well as some other bad conduct relating to conflict of interest (e.g. nepotism) and shirking (e.g. absenteeism by public employees), though there the line-drawing problems get harder. Why not just accept that when most of us use the term “corruption,” we’re talking about that cluster of stuff, and charge ahead with our research?
Matthew shows an understandable degree of impatience with the definitional wrangling over corruption, which is unlikely ever to be resolved given – as Rick points out – the inherently normative notion of what any ‘ideal state’ should look like. Equally, it is probably true both that, grosso modo, we can agree on what corruption entails and that is broadly captured in the widely-used ‘abuse of entrusted power for private gain’ or some close variant thereof.
The key problem, though, is twofold. First, such a definition is a very poor guide to the kind of research that Matthew enjoins us to get on with; second it has given rise to far too much in the way of sloppy conceptualisation in existing research. Basically, the problem with such a definition is that it is simply too broad to be usefully operationalised: it encompasses activities of such vastly different scope, dimension, range, location and impact that it serves little explanatory purpose. The term ‘corruption’, without adjectives, can be hugely misleading and results in the kind of confusion that we are all familiar with over cause and effect – and endless articles claiming that ‘more’ or ‘less’ corruption will have this or that outcome.
In practice, what we need to do is be more systematic in our research about what exactly we are looking at: corruption, after all, takes place in concrete settings at particular times involving actual people. We need, in Sartori’s terms, to move down the ladder of abstraction from the abstract to the concrete. In practice, not enough research does that – especially that which seeks to measure ‘how much’ corruption exists and what its effects are. As Matthew has pointed out before, a lot of that kind of research depends on specifying intervening variables, that are often little short of being plain silly.
We need to focus more on disaggregating the broad notion of ‘corruption’, identifying more carefully its core forms and manifestations. That way, we can understand better how it operates in specific contextual and cultural settings in a given environment, and hopefully we can not only move on from the definitional debate to a more practical one, but also understand better why so many of the proposed intervention strategies over the last two decades have failed to deliver.
Thanks for the thoughtful response. I think I mostly agree… but not entirely. I can’t claim to have thought long or deeply about this, but my general inclination is that there is no “correct” level of generality/specificity for studying social-science concepts–everything depends on the nature of the research question (and, more generally, what we’re trying to figure out).
It is just about always true, with virtually any concept (at least in the social sciences) that it actually comprises a bunch of potentially distinct sub-concepts–each of which can be further subdivided, refined, etc. The same is true in the other direction–virtually any concept can be subsumed under a larger concept. We can always play the lumping and splitting game, in both directions, but we can’t settle on the “correct” level of generality unless we have some sense of WHY we’re trying to isolate the concept.
To illustrate with “corruption,” we could ask “How much corruption?”, taking that concept in more or less the way I (and many others before me) suggested defining it (“abuse of entrusted power for private gain”, basically bribery, embezzlement, and a few other closely related things). Of course, that broad concept includes a lot of things that are quite dissimilar–for example, administrative corruption and political corruption. But of course, “administrative corruption” (and “political corruption”) are also broad categories that also contain within them activities of “vastly different scope, dimension, range, and location”. So do we need to add more adjectives? How many? Does each separate act of “corruption” get its own separate definition, in order to avoid lumping together dissimilar activities? We need some criteria to determine how similar but non-identical events to be in order to group them into a single category. And we can go in the other direction (up the ladder of generality) as well. Why should we talk about “corruption” as opposed to some broader category (such as, to go back to the subject of the original post, “system failures” or “principal-agent problems”)? Again, we need some criteria to decide when a set of activities are sufficiently distinct from other activities that have some common features (but also some differences) that they should be treated as separate, rather than grouped together.
Perhaps this is stating the obvious, but this is not a problem limited to corruption. Think about a whole bunch of other concepts that we use (in both academic research and policy discussions) all the time: democracy, war, revolution, economic growth, natural resources, city, legislator, bureaucrat, crime, etc. Can we study or measure these things? Do we need to split them into smaller, more specific categories? (For example, we could split “war” into “civil war”, “territorial war”, “humanitarian intervention”, etc.) Do we need to lump some of these categories together, on the grounds that they are different manifestations of a similar underlying phenomenon? (For example, we could group “war” and “revolution” together as part of the larger category “political violence”.)
It all depends. Sometimes it makes sense to study “corruption” as a concept; sometimes it doesn’t. Sometimes we need to refine the concept, through the addition of more adjectives, in order to understand what we’re really interested in. Sometimes we don’t. But we can’t have these discussions in the abstract, untethered to some sense of why we’re asking the questions, and what we hope to do with the answers.
All that said, I don’t think we’re all that far apart on this question. I agree that there has been a lot of sloppy operationalization, and I also think we’ve reached the point where moving down the ladder of abstraction, and doing more research on how different forms of corruption manifest in particular contexts, is the way to go at this point. Though I reach this conclusion not so much because I think there’s anything inherently/definitionally wrong with researching the causes and consequences of “corruption” more broadly, but rather because I think there’s already been a bunch of research in that vein, we’ve learned a lot from it, but we’ve hit the point of rapidly diminishing returns, so it’s time to start asking some new questions, with new approaches.
I see that while we all agree that debates about the definition of corruption are distracting and of little value, we are all discussing definitional issues. But that is understandable because so many use the term so carelessly. I suspect we all hope that our discussion will help curb some of that carelessness. In that spirit thanks to Matthew, Jesper, and Paul for their comments.
Matthew, I take your point that calling “corruption” a deviation from an ideal state sweeps to broadly. Today, we would not call a state “corrupt” just because many public officials are dumb or incompetent or both (although respondents to a survey might – who knows). But as I read Pocock’s history of political thought, some political thinkers under the sway of classical republicanism likely would or would have in the past. As he explains in The Machiavellian Moment, for those writing in this tradition “corruption” is the antithesis of the qualities that make for a functioning, durable state. Surely stupidity and incompetence would qualify as “corrupt” by this test.
This earlier usage underlines the point that the term “corruption” today has its roots in the notion of a deviation from the ideal. But in writing that “corruption” is a deviation from the ideal polity, I did not mean to say that we should spend our time trying to define what the ideal looked like. We can’t. That was my point. Since we can’t define the ideal, we will never be able to define a deviation from it and thus precisely define “corruption.” I think Paul’s corruption with adjectives approach is helpful, and certainly will help focus research on the subject (as Collier and Levitsky’s “democracy with adjectives” article has helped sharpen research about democracy), but even “corruption with adjectives” is unlikely to produce the precision we would like. I doubt terming the debate about campaign finance as a question about “political corruption” as opposed to simply “corruption” will bring advocates of limits on contributions and expenditures and supporters of no limits on either any closer to agreement on a definition — or anything else for that matter.
I did not mean to endorse the public interest approach to defining corruption. It falls short for the same reason that approaching corruption as a deviation from the ideal state falls short. The “public interest” is no more capable of exact definition than the ideal state. But I do think it the best (only?) starting point for research on corruption as it leads directly to the use of modern social science tools, tools that I don’t think necessarily bias research outcomes. An analyst could examine, per Melissa Thomas’ Govern Like Us, whether scrapping patronage would lead to civil war or anarchy without any priors about whether patronage is “corrupt” affecting the result.