Marginal Revolution: Perceptions of Corruption

November
S	M	T	W	Th	F	St
1	2	3	4	5	6	7
8	9	10	11	12	13	14
15	16	17	18	19	20	21
22	23	24	25	26	27	28
29	30

« What went wrong with subprime loans in 2006? | Main | How much is America taxing its rich? »

Perceptions of Corruption

Transparency International produces a much cited index of corruption, the Corruption Perceptions Index (CPI). But here is something, shall we say... interesting.

"Transparency International commissions the CPI from Johann Graf Lambsdorff." Lambsdorff, who likes to be called the "father" of the CPI, has another kid on the side, a firm called Anti-Corruption Training and Consulting. And what does this firm do? Well I will let them speak for themselves:

Following an invitation of the Chinese Ministry of Supervsion Prof. Graf Lambsdorff and Mathias Nell went to China from July 22 to July 29 2007. The trip encompassed anti-corruption consultations in Beijing, Nanjing and Chengdu as well as the release ceremony at Tsinghua University of the Chinese version of Prof. Graf Lambsdorff’s new book “The Institutional Economics of Corruption and Reform: Theory, Evidence and Policy”.

China, let us recall, scores a 3.5 out of 10 on TI's Corruption Index where the most corrupt country in the world, Somalia, has a score of 1.4. Pretty corrupt, eh? Here is a picture, from the ACTC website illustrating some of ACTC's consulting:

Hat tip to CPI-Watch.

Posted by Alex Tabarrok on October 23, 2007 at 07:32 AM in Current Affairs, Data Source, Law | Permalink

Comments

Ratings agencies and CDOs all over again. Different players, same results.

What really suprises me is that numerous econ papers actually use all these CPI scores for calculating stuff.

Posted by: sa at Oct 23, 2007 7:57:47 AM

Wow!

It's going to be *really* interesting to see how Transparency International spins these perceptions ...

Posted by: Abi at Oct 23, 2007 8:29:40 AM

You might also be interested in this research from two Wharton professors that were able to replicate the TI rankings by running a series of web searches and recording how many documents contained the word "corruption" within the same paragraph as a country’s name. The corresponding ranking very closely mirrored Transparency International’s corruption rankings.

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=990021
http://knowledge.wharton.upenn.edu/article.cfm?articleid=1786

Posted by: Pablo H. at Oct 23, 2007 8:42:41 AM

Thanks for pointing this out. I have always had thought that TI after meeting Peter Eigen and seeing him try to shake coins from the pockets of everyone in the room. TI is more proof that not-for-profits and NGOs behave like traditional firms.

Posted by: Peter at Oct 23, 2007 9:15:33 AM

Crazy world where a numerical ranking would be worth spending $ to influence.
Don't doubt it is, though.

Posted by: different jeff at Oct 23, 2007 9:21:17 AM

The guy with white socks is obviously uncorruptable.

Posted by: biklett at Oct 23, 2007 12:11:24 PM

What is your point Alex? Do you think a typical photo of a typical meeting with Chinese officials means anything? Would it be better if they didn't try to meet with officials? Unless you can convince me that boycott-mode is the only solution, I'd be as happy as the next man to declare that China is generally opague and corrupt, but this is a pathetic means of trying to convey this point. Seriously. This blog certainly has seen better days.

Posted by: CW at Oct 23, 2007 12:25:03 PM

I'm asking the following question as someone who is genuinely interested in the answer but doesn't have the time to look up the answer: How well does this survey correlate with something like actual corruption?

I guess my worry is that societies with anti-government bents might perceive their governments as being more corrupt than they are and that societies with pro-government bents might perceive their governments as being less corrupt than they are. So, suppose you have two societies with governments that have identical levels of corruption, but one society is libertarian and the other is statist/socialist. We might expect the libertarian society to perceive their government as more corrupt than the socialist society perceives its government.

Posted by: J. at Oct 23, 2007 12:35:11 PM

CW needs to read the post more closely. Hint: the post isn't about China.

Posted by: Alex Tabarrok at Oct 23, 2007 1:03:22 PM

I realize that it wouldn't cost China much (relatively) to buy
CPI's services (and possibly a small uptick in their rating--
watch out Romania), but what tangible benefit would they receive?

Would this number, relayed through the media and academic research
parlay into more business? (Hell, I guess it might.)

Is this roughly equivalent with corporations who purchase TV ad
time where they show feel good images and give their name, but
don't even mention what they do? And does that sort of ad
campaign pay dividends and drive up the stock? I've always been
a curious on that point.

I also find it kinda ironic (in the typical misuse sense of the
word) that this company would promote itself with a photo that
on one hand might lend credibility (very official looking/oppulent
digs/etc.) but on the other hand looks as though the company is
doing business as usual with a huge beaurocracy-- which to me
connotes inefficiency and corruption. It'd look better if he
were standing in an industrial park pointing off in the distance
while the onlookers stand about, rapt.

Posted by: different jeff at Oct 23, 2007 1:42:16 PM

I agree that there is a conflict of interest here... that said, not obvious which way it goes. On the one hand TI has the incentive to portray its clients in a positive light; on the other hand, it may have an incentive to portrary corruption as being worse than it really is in order to generate demand for its services.

I don't understand Alex's point: "China, let us recall, scores a 3.5 out of 10 on TI's Corruption Index where the most corrupt country in the world, Somalia, has a score of 1.4. Pretty corrupt, eh? Here is a picture, from the ACTC website illustrating some of ACTC's consulting..."

I don't have any qualitative problem with those numbers -- what was Alex's point?

Posted by: nobody at Oct 23, 2007 2:31:03 PM

The CPI certainly has its flaws, and quantitative analysis based on the index is undoubtedly soft at its core, but provided the analysis is up front about the limitations and doesn't infer too much from the findings (i.e. makes it very clear both in discussing the data and in interpreting the findings that we're discussing perceptions of corruption rather than corruption proper, I don't see why we shouldn't expect to see journals publishing articles that use it. Not until there's a better (and equally accessible) measure, at least.

Posted by: Mike at Oct 23, 2007 8:24:37 PM

Posted by: SEO at Oct 23, 2007 11:49:18 PM

Wow - this is one of the least meaningful posts I've ever read in a blog. Are you simply checking whether we're awake or not? Sure hope so.

Posted by: Odd at Oct 24, 2007 5:18:00 AM

Checking whether Google is awake more likely.

Posted by: Christine at Oct 24, 2007 6:45:47 AM

Prof. Tabarrok wrote: "Is this roughly equivalent with corporations who purchase TV ad time where they show feel good images and give their name, but don't even mention what they do?"

That seems a much more valid conjecture (though perhaps not quite so interesting or provocative as) than what I originally took away from the post (which was more along the lines of, "Ah, he's suggesting TI is corrupt," as some other posters mentioned).

Maybe it should be read as more a critique of TI's signaling (in this particular photograph) than a critique of TI's integrity?

Posted by: Philip at Oct 24, 2007 12:46:16 PM

I just participated in the Economics of corruption in Passau, a workshop led by Prof. Lambsdorf, and this discussion is very timely.

It's obvious from the original post that there is a conflict of interest, but some people appear to be mixing up Lambsdorf (who is a professor and consultant to TI and ACTC) with TI itself. I wonder if TI knows about Lambsdorf's activities?

Second, there are a number of reasons that the perceptions index should NOT be used. The biggest one -- as Lambsdorf says -- is that it's invalid for panel regressions. (It's debatable whether it's useful for cross country regressions; I think not.) For those seeking alternatives, I suggest some objective, cross-country surveys of actual bribes taken, TI's bribe payer's survey or the WB's Doing Business surveys.

As far as China's incentives are concerned, be VERY CLEAR that the corruption survey does impact business investment decisions. Given that the Party does business as well as politics, note the strong incentives for them to improve the PERCEPTION of corruption. (Actual corruption is another matter.)

Posted by: David Zetland at Oct 24, 2007 1:18:09 PM

In some recent work at the World Bank, we tried to look at the quality of public financial management systems in countries and corruption -- the thesis being that better public financial management systems reduce the risk of public corruption. The best data we could find was perception data. There is of course no way of measuring actual corruption. For example, if one country's supreme audit institution (SAI) detects public sector corruption and publishes the information, while another country's SAI does not, that does not mean the former has more corruption than the latter.

For those interested in the work, see a recent IMF Blog posting (http://blog-pfm.imf.org/pfmblog/2007/09/public-financia.html), or visit the blog itself at http://blog-pfm.imf.org

Posted by: Bill Dorotinsky at Oct 25, 2007 11:55:36 AM

H i B e s t w i s h e s。Allow me to offer my heartiest wishes.x i c a o l o v e s-流水线娱乐博客常年提供高、中、低压锅炉钢管、流体钢管、结构钢管、化肥专用钢管、石油裂化钢管、地质钢管、液压支柱钢管及合金钢管-无缝管-无缝钢管等利用同声传译设备来论文发表资讯/刊物信息，协助客户制定论文发表方案