The Persistence of Phony Statistics in Anticorruption Discourse

Early last month, UN Secretary General António Guterres delivered some brief opening remarks to the Security Council at a meeting on the relationship between corruption and conflict. In these remarks, Secretary General Guterres cited a couple of statistics about the economic costs of corruption: an estimate, attributed to the World Economic Forum (WEF), that the global cost of corruption is $2.6 trillion (or 5% of global GDP), as well as another estimate, attributed to the World Bank, that individuals and businesses cumulatively pay over $1 trillion in bribes each year. And last week, in her opening remarks at the International Anti-Corruption Conference, former Transparency International chair Huguette Labelle repeated these same figures.

Those statistics, as I’ve explained in prior posts (see here and here) are bogus. I realize that Secretary General Guterres’ invocation of those numbers shouldn’t bother me so much, since these figures had no substantive importance in his speech, and the speech itself was just the usual collection of platitudes and bromides about how corruption is bad, how the international community needs to do more to fight it, that the UN is a key player in the global effort against corruption, blah blah blah. Ditto for Ms. Labelle–her speech used these numbers kind of like a rhetorical garnish, to underscore the point that corruption is widespread and harmful, a point with which I very much agree. But just on principle, I feel like it’s important to set the right tone for evidence-based policymaking by eschewing impressive-sounding numbers that do not stand up to even mild scrutiny. Just to recap:

  • The “$1 trillion in annual bribes” number, often attributed to the World Bank as an institution, actually comes from the appendix to a chapter by Daniel Kaufmann (a World Bank researcher) in the WEF’s 2005-2006 Global Competitiveness Report. That appendix relies on data from surveys of firms or households that ask how much the respondents typically pay each year in bribes (expressed as a percentage of household income in the case of household surveys, and as a percentage of sales in the firm surveys). Kaufmann then tries to convert these responses to dollar amounts, relying on highly dubious assumptions. He then extrapolates the data to other countries not included in the surveys by using even more dubious assumptions (principally the notion that the amount paid in bribes per capita each year is about the same in countries with similar scores on country-level corruption perception indexes). Even if we put aside a lot of the obvious problems with the methodology, the uncertainty in the estimated bribe quantity is huge, ranging from $600 billion to over $1.5 trillion, depending on which (strong and dubious) assumptions one chooses to make. The $1 trillion point estimate comes from the fact that it’s roughly the midpoint of the range, and has no further justification. Really the punchline of Kaufmann’s appendix, properly interpreted, is that though we know that there’s a lot of bribery, but we have no real idea how much. And it might also be worth noting that the data on which this estimate is based are more than 14 years old.
  • The “$2.6 trillion annual economic cost of bribes” estimate (often also expressed as 5% of global GDP) is even worse. That number, attributed to the WEF, actually comes from basically nowhere. The WEF did not come up with this estimate itself. Rather, it cited to the same Kaufmann appendix that was used to come up with the $1 trillion in bribes figure. But the Kaufmann analysis didn’t purport to estimate the economic cost of all that bribery, so citing that study for any such estimate is clearly a simple and obvious error. So how did the WEF extract its figure for annual cost of bribery from the Kaufmann appendix? So far as I can tell, the Kaufmann study tried to “validate” its $1 trillion in annual bribes figure by seeing whether it was comparable t estimates of the magnitude of other sorts of illicit activity. In undertaking this “validation” exercise, Kaufmann cited two papers (from 1998 and 1999) that tried to estimate the amount (not the cost) of money laundering worldwide. One of those studies estimated the amount of money laundering at between $600 billion and $2.8 trillion, equivalent to 2%-5% of global GDP at the time (that is, in 1999). A 2007 World Bank document appeared to take up the high end of that range (5% of global GDP), misrepresent that estimate as the annual cost of bribery (rather than the annual amount of money laundering), and applied the percentage to 2007 global GDP, producing the $2.6 trillion figure, which was later cited by the WEF (without proper attribution).

This is clearly nonsense, as I explained in my posts on these figures from almost three years ago. I guess it’s mostly-harmless nonsense, given that no consequential policy or advocacy decisions depend on whether the annual cost of bribery is $500 billion or $1 trillion or $2.6 trillion or unknown. These numbers aren’t about rigorous cost-benefit analysis or even soft priority-setting; they’re just decoration, and their purpose would be served equally well by making up a random number ending in “-illion”. Still, I wish people would stop using made-up numbers as a rhetorical flourish, not only on general principle but because I think we should be collectively sending a message, and setting a tone, that facts and evidence and careful scrutiny of data are crucial for addressing corruption and other serious social problems.

14 thoughts on “The Persistence of Phony Statistics in Anticorruption Discourse

      • Often referenced as a World Bank figure (Brun, Jean-Pierre, Larissa Gray, Clive Scott, and Kevin M. Stephenson. 2011. Asset Recovery Handbook: A Guide for Practitioners. Washington, DC: The World Bank, United Nations Office on Drugs and Crime (UNODC), the International Bank for Reconstruction and Development), it seemingly came out of nowhere, yet is one of the most frequently used.

        The furthest back I have been able to trace it is to a 2001 Nyanga Declaration ( No reference made to any methodology. Has anyone gotten any further?

  1. I think batting around such numbers is harmful. If the same gigantic numbers are repated year after year, there is a risk it will lead to defeatism. Many will conclude no progress can be made in tackling the problem, view any effort to combat corruption as futile, and not lift a finger (or cast a vote) to advance it. Alternatively, the constant repitition of such numbers will lend support to those proposing drastic, illiberal and undemocratic, remedies.

    • Precisely! In recent posts it has been discussed the relationship between anticorruption rethoric and populism/extremism, and I think this is a point that must be addressed. The use of inaccurate numbers (or conclusions in general) to artificially inflate indignation and mobilize crowds against corruption can give room to simplistic and radical solutions (exactly the environment on which populist/radical movements thrive).

      From my perspective, the best contribution academia can give to counter such radical movements is to be as accurate as possible in its description and analysis of the problem.

  2. Thanks, Professor, for this helpful post. I suppose I understand why these numbers are regularly trotted out in speeches that attack global corruption: erroneous or not, the numbers help make the corruption problem—and anticorruption work—seem important and tangible. Given that policymakers and advocates seem to value the rhetorical impact of quoting global corruption numbers, I’m wondering if you know of any studies that have more recently and more rigorously attempted to measure the scale of global corruption. I glanced through your corruption bibliography and nothing immediately jumped out at me, but this is obviously a question of intense interest and, daunting or not, I imagine someone must have tackled this problem in the intervening years between 2006 and today. With the understanding that any such estimate can only be very tentative, to your knowledge are there better, more reliable numbers out there?

    • The short answer is no, I’m not aware of any rigorous attempts to estimate these aggregate quantities or costs on a global basis, and I’m skeptical that doing so would be feasible. There is, by contrast, a wealth of evidence from more narrowly focused studies that corruption has significant adverse consequences for the economy, social stability, health outcomes, etc. My instincts are that we can and should rely on that latter body of evidence to make the case that corruption is a real and serious problem, and abandon the attempt to come up with a global estimate.

  3. Thank you Professor Stephenson for highlighting the fact that those statistics do not have any real basis to rely on. While those incorrect statistics are indeed nothing more than “decoration” and do not have any real effect in terms of policymaking, the fact that they have been mentioned in UN Secretary General António Guterres’s remarks might unfortunately lead them into academic publications (not to mention op-eds) citing Guterres’s remarks as the source.

  4. Another zombie stat that often gets repeated: “25% of Africa’s GDP, amounting to US$148 billion, is lost to corruption every year.”

    It gets attributed to an African Union study (sometimes cited as 2002, other times to the late 1990s), but I’ve never been able to track down the original source or methodology.

    I too would be very interested in any more recent and credible research that’s tried to approximate the cost, size, scale, etc. of corruption. In my research, I’ve not found much.

    Perhaps an area for academics, the World Bank, IMF, OECD, etc. to dig into?

  5. Hey, Matt. I never cite these figures myself, and share many of the same concerns that you have. And, like you, I can be kept up at night over such things. In fairness, though, you should note that a 1 followed by many 0s gives an immediate clue that this is an order of magnitude type of guess, rather than a “statistic”. (Talking only about the original figure, not the more recent ones that give the illusion of more precision.) In fact, the tone of your earlier post was much more charitable toward Dani’s annex. At least the methodology is presented and the sense of uncertainty is there… (Does he ever even call it a “statistic”?)

    Conceptually, TI used to give a great example of why it is misleading to even think of the cost of corruption as the flow of bribes. I can’t find the citation, but it went something like this: Imagine a $100,000 bribe is paid to win a contract to build a hospital, and the cost ends up being $1,000,000 higher as a result. More bribes are paid to the inspectors so the contractor can cheat on building materials, increasing the profits further. Imagine the hospital collapses. What is the cost of corruption? The bribes? Surely not. The amount of the overpayment for the hospital? The entire cost of the hospital since it needs to be rebuilt? All of those plus the value of the lives lost?…


  6. Thank you, Professor Stephenson, for sharing your insight on the unimpressive methodology that underlies the statistics you mention. Reading some of your colleagues’ comments on this post, it seems like there already is some agreement on how unreliable such statistics are. One of the biggest problems I see, even given the distrust from some anticorruption experts, is just how prevalent these unsupported statistics are in media reportage on corruption. When news articles mention such numbers, the public generally has a certain trust that these numbers are somehow supported—even if exactly how is unknown. The thought is, I assume, that someone qualified out there must have done the thinking to come up with them. And these numbers, without questioning the methodology that supports them, become what people consider reality. This has a huge effect on public morale and regard of their own governments, not to mention that buffed up numbers make corruption seem like a problem too difficult and too nebulous to effectively resolve.

Leave a Reply to Matthew Stephenson Cancel reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.