Where Does the $2.6 Trillion Corruption Cost Estimate Come From?

Last year, I published a post expressing my puzzlement regarding the source of the widely-cited statistic that over $1 trillion dollars are paid annually. I was pleasantly surprised by the speed with which several GAB readers pointed me to the original source that described the methods and data used to calculate that number—and while I wasn’t entirely satisfied, it was at least nice to know where the number came from.

I’m hoping someone out there can help me with a very similar question: Within the last few months, I’ve been at several conferences and meetings where someone has quoted the figure that worldwide corruption (not just bribery) imposes annual costs to the global economy of approximately $2.6 trillion, roughly 5% of global GDP. I’ve looked and looked, and I cannot for the life of me figure out where this number comes from.

Most sources I’ve found trace back to an estimate purportedly done by the World Economic Forum (WEF), with no citations to any other sources (other than sources which, in turn, cite the WEF). But most of the WEF documents I’ve found just quote the $2.6 trillion figure without any source or explanation (see, e.g., here and here). As far as I can tell, the earliest WEF document to contain this figure (and the only one to contain anything resembling a source reference) is an undated advocacy document (published jointly with the International Chamber of Commerce, Transparency International, and UN Global Compact). That document states that “[e]stimates show that the cost of corruption equals more than 5% of global GDP (US $2.6 trillion), with over US $1 trillion paid in bribes each year,” and cites the World Bank (with no further detail) as the source of these figures.

As discussed in my earlier post, there is indeed a World Bank research document that derives the “$1 trillion in annual bribes” figure (subsequently discussed in a chapter in one of the WEF’s annual reports). But that published chapter (the only source for this figure that I can access) doesn’t contain the $2.6 trillion total cost estimate. It is, I suppose, possible that either the unpublished version of the World Bank document contains the $2.6 trillion figure, or that the WEF document was citing to a different World Bank source. That said, it’s notable that no other source I’ve found, including those published by the World Bank, has cited to the World Bank rather than the WEF as the source of the $2.6 trillion number, which casts some doubt on the idea that there’s some unpublished World Bank document that derives this estimate.

I’m at a loss. Does anyone know how the WEF (or World Bank, or whoever was the original source of this estimate) came up with the $2.6 trillion cost figure? It’s a bit troubling that it’s so difficult to find the original source and methodology for such a widely quoted statistic.

13 thoughts on “Where Does the $2.6 Trillion Corruption Cost Estimate Come From?

    • No, I haven’t asked anyone at the UN directly, partly because I’m not sure whom to ask. (If you have any ideas, or could do some digging yourself, I’d be most grateful!) Every public document I can find online (including the UN fact sheet you linked to) seems to trace back to the WEF “Clean Business Is Good Business” document I mentioned. But as I said in the post, that document just cites “the World Bank,” and there the trail goes cold. Like you, I assumed there’d be a good (or at least identifiable) source for the number, but I can’t find it. Hence the post, expressing my befuddlement and seeking clarification.

  1. This is certainly puzzling. I did a bit of searching from a somewhat different starting point, specifically looking for some World Bank document pegging the cost of corruption figure at five percent of global GDP. I found one report from June 2007 that might be WEF’s source: http://siteresources.worldbank.org/NEWS/Resources/Star-rep-full.pdf (see specifically footnote 7 and the accompanying text) (Google dates the undated WEF document you linked above as coming from June 25, 2008). Interestingly, this figure is cited within a section of the report entitled “Estimates of the Size of the Problem and Potential Benefits from Tackling High-level Corruption”

    But the works cited in that World Bank report do not directly refer to corruption costs, making it unclear whether the five percent proposition in the report is actually being supported by the material it cites. Specifically, the World Bank document (and this earlier UNODC page citing similar figures (https://www.unodc.org/unodc/en/money-laundering/globalization.html)) are referring to the global costs of money laundering. The World Bank report also includes a citation to this 1998 speech by then-IMF Managing Director Michel Camdessus, which focused on the macroeconomic effects of money laundering, with some discussion of corruption: http://www.imf.org/external/np/speeches/1998/021098.htm.

    Perhaps at some point someone made a leap to tie the amount of money being laundered to the price of corruption, but it seems like a stretch. Still, this is one place where researching about the cost of corruption totaling five percent of global GDP leads.

    • This is great, nice job digging into this further. I think you may well be right that this is the possible “World Bank” source for the estimate in the June 2008 WEF document. If it is, though, then it would mean that the $2.6 trillion/5% figure for the economic costs of corruption is completely wrong, based on a misreading of the underlying sources. As you point out, the World Bank report gives this figure as an estimate not of the costs of corruption, but as an estimate of the volume of money laundering—not the same thing at all. Moreover, the World Bank report says, in the footnote you mention, that although this estimate “has been widely quoted in the literature …. a documented basis for it could not be found.” And neither of the other two sources you mention, which the World Bank source cites, actually give a source for the “5% of GDP” estimate.


  2. Hi Matt,

    A little tongue in cheek, but the WEF includes in its list of Industry Partners (http://www.weforum.org/industry-partners) the likes of Siemens and BAE, Walmart, HSBC & other luminaries from FCPA cases over recent years. Perhaps they have opened their books to the WEF and the figures are far more accurate than we think…

    On that disturbing thought, Merry Christmas to all.

  3. Pingback: Where Does the $2.6 Trillion Corruption Cost Estimate Come From?  | Anti Corruption Digest

  4. Dear Matthew,

    The ‘cost’ of corruption has been one of my episodic interests, not in the least because it appears difficult to define, quantify and verify. I had tracked the commonly-quoted $1 trillion to a 2004 announcement by the World Bank Institute (WBI). This claimed the figure to be based on 2001-02 worldwide surveys of enterprises and ‘an estimate’ on bribes paid by household users of public services ‘derived’ from the WBI governance and anti-corruption diagnostic surveys. The 5 per cent figure appears to come from a comparison with ‘estimates of the world economy’.

    This allowed researchers and commentators to continue to use an assumption that the cost of bribery was around 3.5-5% of global GDP. Thus the World Economic Forum figure you mentioned has claimed that ‘estimates’ suggest the cost of corruption amounts to more than 5% of global GDP (US$ 2.6 trillion) with more than US$ 1 trillion paid in bribes each year. Other organisations, such as the OECD and U4, reproduce the figures without verifying or testing the data sources.

    The related problems are what we mean by ‘corruption’ and what data sets we use; the WBI extrapolation did not, for example, appear to include embezzlement or theft (or misuse) of public assets When the figure of $1 trillion was used in 2010 as the NGO Global Financial Integrity as the totality of financial crimes it included corruption, tax evasion and other financial crimes ‘lost by developing economies’. It arrived at this figure on the basis that unrecorded capital on state balance of payments (source and use of funds) must be illicit and must have left the state. The main area of loss was trade mispricing/misinvoicing which accounted for some 80 percent of cumulative illicit flows, leaving an annual estimate of some $200 billion in bribes being remitted abroad.

    Other agencies which have tried to quantify the cost of bribes have a much lower figure: FATF has used (in 2011) a figure of $50 billion, although its analysis of main cases suggests $27 billion. The StAR initiative claims that $20 to $40 billion is stolen annually because of corruption (on the basis of ‘conservative estimates’ and representing ‘20 to 40 percent of annual international development assistance’). Of course trying to define, quantify and verify what is ‘international development assistance’ – and in particular what is spent on anti-corruption work – is another matter entirely.

    Alan Doig

  5. Just came across this interesting string in doing research for the third ed of my textbook on Global B2B marketing. This reminds me of the problems I have had in tracking down the widely cited number (interestingly also 5% ) used for the amount of counterfeit goods in world trade. I was unable to verify any reliable source for the estimates that were quoted in literally thousands of Google references. Please let me know if yoiu have found a better source for the data on corruption.

  6. Pingback: Beheading the Hydra: How the IMF Fights Corruption – IMF Blog

Leave a Reply

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

WordPress.com Logo

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

Twitter picture

You are commenting using your Twitter 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.