The most famous corruption indicator is Transparency International’s Corruption Perceptions Index. Its only problem is that the perceptions of their self-appointed experts have nothing to do with reality!
As I explained in previous posts on this blog, it suffers from numerous flaws. Part of it has to do with its questionable methodology: using changing mixes of different surveys to gauge a fluid, opaque-by-definition social phenomenon. Another is its reliance on its appeal to authority, the theory being that “experts” in business and think-tanks know more about corruption relative to anyone else. Countries with more regulations are systematically prejudged, as are those facing hostile media environments such as Russia or Venezuela. Above all, the CPI doesn’t pass the face validity test – in other words, many of its results are frankly ludicrous. Is it truly plausible that Russia (2.1) is as corrupt as failed states like Zimbabwe (2.4) or D.R. Congo (2.0), or that Italy (3.9) is more corrupt than Saudi Arabia (4.7) which is a feudalistic monarchy!?
This suggests that we urgently need another, more objective index. Thus I present the Corruption Realities Index (CRI)!
Unlike my previous attempt at this, the Karlin Corruption Index – which was rightly critiqued for being no less subjective than the CPI (though I do believe it was more accurate) – this time round I am drawing on real world data. In particular, there are three corruption indices that aren’t as well known as the CPI, but far more useful.
One of them is Transparency International’s less well-known Global Corruption Barometer. Every year, they poll respondents on the following question: “In the past 12 months have you or anyone living in your household paid a bribe?” The answers hint at the prevalence of corruption in everyday life, as experienced by a sample of normal people, and as such they almost certainly offer a better picture than the perceptions of experts who are prone to the narrative fallacy and are unduly influenced by the ideological biases of the international business media (e.g. op-eds in The Economist or the WSJ). A good example is the reputational fallout Russia experienced in the wake of the prosecution of Mikhail Khodorkovsky, which saw its CPI retreat despite the lack of any noticeable increase in corruption on the ground.
Another key resource is the Global Integrity Report, which evaluates countries on their “actually existing” Legal Frameworks and Actual Implementation on issues such as “the transparency of the public procurement process, media freedom, asset disclosure requirements, and conflicts of interest regulations.” This involves line by line examination of the laws in question, and the “de facto realities of practical implementation.” Crucially, the assessments are blindly reviewed by a panel of peer reviewers and outside experts, which is an important antidote to bias.
Finally, there is the International Budget Partnership, which – believe it or not – assesses budget transparency and accountability. It compiles an Open Budget Index on the basis of factors such as budget documents availability, and the effectiveness of oversight by legislatures and supreme audit institutions. People who think of Eastern Europe as a black hole of government corruption will be surprised to learn that it is the best performing region after the developed world, while the Middle East, China, and sub-Saharan Africa are distinguished by their opacity.
Data from all three sources – to the extent that it is available – is amalgamated and fed through a formula to produce the Corruption Realities Index.
GCB is the percentage of people saying they or their households paid a bribe in the past 12 months from 2010 data. GILF is the Global Integrity Legal Framework score. GIAI is the Global Integrity Actual Implementation score. Most Global Integrity data is from 2007-2010. OBI is the Open Budget Index score from 2010 data. The CRI is the Corruption Realities Index. More details are given at the bottom of the table.
| Country | GCB /% | GILF /100 | GIAI /100 | OBI /100 | CRI | |
|---|---|---|---|---|---|---|
| 1 | Denmark | 0 | 10.0 | |||
| 2 | UK | 1 | 87 | 9.1 | ||
| 3 | Norway | 1 | 83 | 9.0 | ||
| 4 | Korea, South | 2 | 96 | 82 | 71 | 8.7 |
| 5 | Finland | 2 | 8.6 | |||
| 6 | Netherlands | 2 | 8.6 | |||
| 7 | Switzerland | 2 | 8.6 | |||
| 8 | Germany*† | 2 | 83 | 83 | 67 | 8.4 |
| 9 | Portugal*† | 3 | 86 | 86 | 58 | 8.3 |
| 10 | Iceland | 3 | 8.3 | |||
| 11 | Australia*† | 2 | 78 | 78 | 8.2 | |
| 12 | Israel*† | 4 | 83 | 83 | 8.2 | |
| 13 | USA | 5 | 90 | 78 | 82 | 8.1 |
| 14 | Slovenia | 4 | 68 | 8.1 | ||
| 15 | New Zealand | 4 | 90 | 8.0 | ||
| 16 | Ireland | 4 | 8.0 | |||
| 17 | Brazil | 4 | 85 | 66 | 71 | 7.9 |
| 18 | Spain | 5 | 84 | 74 | 63 | 7.8 |
| 19 | Canada | 4 | 90 | 61 | 7.8 | |
| 20 | Hong Kong | 5 | 7.8 | |||
| 21 | Georgia | 3 | 86 | 58 | 55 | 7.8 |
| 22 | Bulgaria | 8 | 97 | 73 | 56 | 7.7 |
| 23 | Croatia | 5 | 57 | 7.7 | ||
| 24 | France | 7 | 87 | 68 | 87 | 7.7 |
| 25 | Japan | 9 | 91 | 76 | 7.7 | |
| 26 | Argentina | 12 | 97 | 77 | 56 | 7.5 |
| 27 | Taiwan | 7 | 7.4 | |||
| 28 | Latvia | 15 | 93 | 76 | 7.3 | |
| 29 | Italy | 13 | 84 | 71 | 58 | 7.1 |
| 30 | Poland | 15 | 86 | 71 | 64 | 7.0 |
| 31 | Singapore | 9 | 7.0 | |||
| 32 | Austria | 9 | 7.0 | |||
| 33 | Czech Republic | 14 | 84 | 64 | 62 | 6.9 |
| 34 | Peru | 22 | 93 | 70 | 65 | 6.7 |
| 35 | Chile | 21 | 87 | 66 | 72 | 6.6 |
| 36 | Indonesia | 18 | 92 | 56 | 51 | 6.6 |
| 37 | FYR Macedonia | 21 | 90 | 65 | 49 | 6.5 |
| 38 | Vanuatu | 16 | 84 | 55 | 6.5 | |
| 39 | Fiji | 12 | 64 | 64 | 6.5 | |
| 40 | Kosovo | 16 | 78 | 60 | 6.5 | |
| 41 | Romania | 28 | 95 | 64 | 59 | 6.3 |
| 42 | Armenia* | 22 | 72 | 72 | 6.3 | |
| 43 | China | 9 | 76 | 43 | 13 | 6.2 |
| 44 | Hungary | 24 | 83 | 62 | 6.2 | |
| 45 | Serbia | 17 | 80 | 44 | 54 | 6.2 |
| 46 | Russia | 26 | 90 | 54 | 60 | 6.1 |
| 47 | Colombia | 24 | 89 | 49 | 61 | 6.1 |
| 48 | Philippines | 16 | 84 | 31 | 55 | 6.0 |
| 49 | Malaysia | 9 | 57 | 37 | 39 | 6.0 |
| 50 | Luxembourg | 16 | 6.0 | |||
| 51 | Papua New Guinea* | 26 | 69 | 69 | 57 | 6.0 |
| 52 | Bosnia & Herzegovina | 23 | 90 | 39 | 44 | 5.8 |
| 53 | Thailand | 23 | 79 | 50 | 42 | 5.8 |
| 54 | Venezuela | 20 | 84 | 39 | 34 | 5.8 |
| 55 | Lithuania | 34 | 85 | 63 | 5.8 | |
| 56 | Mexico | 31 | 83 | 59 | 52 | 5.8 |
| 57 | Greece | 18 | 5.8 | |||
| 58 | Moldova | 37 | 89 | 59 | 5.7 | |
| 59 | Solomon Islands | 20 | 63 | 52 | 5.6 | |
| 60 | Belarus | 27 | 81 | 48 | 5.6 | |
| 61 | Turkey | 33 | 75 | 57 | 57 | 5.5 |
| 62 | Bolivia | 30 | 78 | 56 | 13 | 5.3 |
| 63 | Ghana | 37 | 78 | 51 | 54 | 5.3 |
| 64 | Ukraine | 34 | 77 | 39 | 62 | 5.2 |
| 65 | India | 54 | 86 | 55 | 67 | 5.0 |
| 66 | Pakistan | 49 | 91 | 47 | 38 | 4.9 |
| 67 | El Salvador | 31 | 37 | 4.8 | ||
| 68 | Azerbaijan | 47 | 88 | 40 | 43 | 4.8 |
| 69 | Lebanon | 34 | 65 | 39 | 32 | 4.7 |
| 70 | Mongolia | 48 | 70 | 43 | 60 | 4.6 |
| 71 | Kenya | 45 | 62 | 45 | 49 | 4.5 |
| 72 | Palestine | 51 | 73 | 41 | 4.3 | |
| 73 | Sierra Leone | 71 | 79 | 58 | 4.2 | |
| 74 | Zambia | 42 | 36 | 4.1 | ||
| 75 | Senegal* | 56 | 65 | 65 | 3 | 4.0 |
| 76 | Uganda | 86 | 99 | 45 | 55 | 4.0 |
| 77 | Nigeria | 63 | 73 | 44 | 18 | 3.8 |
| 78 | Vietnam | 44 | 56 | 28 | 14 | 3.7 |
| 79 | Cameroon | 54 | 69 | 39 | 2 | 3.6 |
| 80 | Iraq | 56 | 75 | 32 | 0 | 3.4 |
| 81 | Liberia | 89 | 64 | 43 | 40 | 3.1 |
| 82 | Afghanistan | 61 | 21 | 2.8 | ||
| 83 | Cambodia | 84 | 58 | 30 | 15 | 2.6 |
* The Global Integrity scores for Legal Framework and Actual Implementation were given as one averaged figure, so bear in mind that
† Global Integrity scores were collected for 2006 or earlier, so may no longer be up to date.
The CRI for countries in italics was generated on the basis of only one piece of data, the percentage of people saying they or someone in their household paid a bribe in the last year. As such, their CRI has a high margin of error.
Formulas
- For countries with all four data points. (10-sqrt(GCB))*5 + (GILF + GIAI)/5 + sqrt(OBI) = CRI, with the GCB/GILF/GIAI/OBI weighted 50-20-20-10.
- For countries with no OBI. (10-sqrt(GCB))*5 + (GILF + GIAI)/4 = CRI, with the GCB/GILF/GIAI weighted 50-25-25.
- For countries with no Global Integrity data. (10-sqrt(GCB))*7.5 + sqrt(OBI)*2.5 = CRI, with the GCB/OBI weighted 75-25.
- For countries with only GCB. (10-sqrt(GCB))*10, with the GCB being the only weight by definition.
Needless to say, the accuracy of any CRI score increases with the amount of data it is based on.
I did not bother including any country that doesn’t have polling data on the prevalence of bribery over the past year, as it is an indispensable indicator. One can only hope that the Global Corruption Barometer will expand its coverage in the coming years, as the CRI depends so much on its data.
One major group of countries that isn’t assessed here, because of a lack of data, are the Gulf monarchies. If we make the (rather generous?) assumption than only 5% of their households pay a bribe in any given year, then based on the UAE’s and Kuwait’s Global Integrity scores and Saudi Arabia’s OBI, these countries will have the following CRI: Kuwait (6.2), UAE (6.7), Saudi Arabia (4.0).
This is but the beginning. I hope to search out more sources of data like the GCB, and keep expanding the Corruption Realities Index in the years ahead.
EDIT 05/26/2011: Note that there are going to be substantial internal variations for corruption within a country, as different regions will have varyingly corrupt bureaucracies and police forces. To take the example of Russia, for instance, a recent FOM poll indicates that the frequency of requests for bribes ranged from, say, 7% in Omsk oblast, to 31% in St.-Petersburg. This would translate into CRI scores of 7.3 and 5.9, respectively.
Is the UK really that clean?
That’s what the available figures say… though I should point out that the UK lacks data from Global Integrity. I doubt they would change anything cardinally though.
Its ranking would probably go down slightly if you add things like tax compliance by big companies and banks, or the “legal corruption” that some may argue is engendered by its rigid class hierarchy and old boys networks, but such things are very hard to measure and are therefore excluded.
Is there a correlation between corruption and specific social phenomena? I’m curious if societies with high rates of cousin marriage are more susceptible to corruption than countries in which it’s rare.
Above all, the CPI doesn’t pass the face validity test – in other words, many of its results are frankly ludicrous. Is it truly plausible that Russia (2.1) is as corrupt as failed states like Zimbabwe (2.4) or D.R. Congo (2.0), or that Italy (3.9) is more corrupt than Saudi Arabia (4.7) which is a feudalistic monarchy!?
This would carry more weight if you had any experience with which to conduct a face validity test. Where have you done business in your career? The Gulf states are not considered particularly corrupt for two reasons: 1) you only have to pay the top guy once, as opposed to every man and his dog, and 2) in the Gulf you merely bribe somebody to bend the rules, whereas in truly corrupt places you have to bribe people just to let you do what you’re supposed to be allowed to do anyway.
And if you’re going to accuse Barclays of dodgy tax compliance, you’d do well to avoid citing editorials in The Guardian, whose staff appears not to include anyone who knows much about economics. There are good reasons why the Barclays didn’t pay much UK tax, and it isn’t a form of corruption.
No, I don’t have business experience outside the US (which is quite clean to date). However, contrary to what you imply, personal experiences, and specifically business corruption, aren’t the be all and end all of corruption.
As I told you the last time you raised this point, I have a friend who runs a UK higher education program (engineering) with a big international student intake, with whom I’ve discussed the corruption issue on several occasions it being a matter of mutual interest.
There is strong evidence to suggest that many of the Saudis and Omanis got on the program only because they had the right connections. Many of those students don’t do any work, don’t even turn up to lessons, coast along, get their Third-class degrees (the university takes pains not to fail any of them – even though in many causes it should – because they are a very lucrative revenue stream, which in itself may be a mild form of corruption), and effortlessly land lucrative bureaucratic jobs upon their return home. These students from the oil-rich Gulf states are also the only ones from whom he ever received bribe offers (which he refused needless to say).
He estimates the percentages of students who got in по блату to be much less or non-existent for other countries that regularly participate in the program, which include (among others) China, several East European countries, Italy, and Botswana. They generally provide good students, with the exception of Botswanans who are weak, but unlike many Saudis and Omanis they are hardworking and honest.
Finally, you haven’t addressed one of the main points of this blog post that undermines the picture of corruption painted by the CPI – that where detailed, objective data on anti-corruption laws and their implementation is available, the Gulf states do worse than Russia, let alone Italy. The Saudis reveal hardly any of their budget (a wise decision, perhaps, given that so much of it flows directly into the coffers of members of the House of Saud), scoring 1/100 in the OBI 2010, compared to 58/100 for Italy and 60/100 for Russia. According to Global Integrity, on Legal Framework and Actual Implementation, respective, Italy scores 84-71/100 and Russia scores 90-56/100 (the big gap isn’t a surprise since everyone’s gripe there is that enforcement of laws is weak, as opposed to the laws themselves); the UAE scores 63-68/100 and Kuwait scores an average of 55/100.
Re-Barclays. Well, good for the Brits if banks aren’t as nefarious as the Guardian makes them out to be. The UK is near the top in most corruption ratings anyway, including this one.
I respect the effort to come up with the CRI. But the fact that it ranks Kosovo and Georgia higher than Russia tells me it is a failure. Transparency International is a propaganda outfit tasked to create a “citable source” for media pieces. So the GCB is worthless. Seriously, Georgia 3 and Russia 26? Don’t make me laugh.
It’s based on concrete data. Like it or not, but only 3% of Georgians said they or their household paid a bribe in the past year. Last year’s results were similarly low, so it isn’t a statistical fluke.
Russia’s 26% is roughly in line with other surveys, such as FOM and Levada.
Sorry but I don’t buy the 26% of Russians pay bribes vs 3% Georgians nonsense. None of my relatives paid any bribes in the last 10 years and they are middle of the road people living in Saratov. This BS survey claims that 1 in 4 Russians have to pay bribes, a number beyond any hope of plausibility. The Levada numbers say 15% paid bribes in the last four years and that the average bribe was 8887.4 roubles in 2010, which is nearly $300, and a lot of money for most people in Russia. There is no distinction being made between bribes imposed by corrupt officials on the populace and bribes given to officials for the purposes of queue jumping by people with too much initiative. If hospital stays required bribes then everyone would be paying them and not 1 out of 3.4. For examples of real corruption take a look at certain African countries where 100% of the people wanting hospital treatment have to pay some sort of bribe. Paying a nurse some cash to get better treatment is not in the same class of problem.
Erm, paying nurses to get better treatment is still a bribe. The criteria for all countries in the GCB are in any case the same, so I don’t exactly see your point.
$300 may be a lot for an average bribe, but don’t forget its the mean bribe size. Not the median. The median may be something like $20, i.e. the cost of escaping a fine for some minor traffic violation. But very big business bribes, i.e. in the $10,000’s, may inflate the mean bribe size to serious levels.
Georgia’s very low bribe incidence may be partly ascribed to the big deregulation seen there in recent years. Note that Saakashvili is a libertarian ideologue. With fewer rules to have to go around, the incidence of corruption will logically decline. Nonetheless, it’s still a huge achievement on their part.
The 3% figure for Georgia indicates to me some selective filtering of the facts. There is nothing structurally and culturally different between Georgia and the other Caucasian republics and the rest of the former USSR. It ain’t the baltics. Sakashvili’s regime can no more remove this than Putin’s regime. I am quite sure that any money and gifts being handed over to Georgian officials and civil servants on a voluntary basis is not being treated as bribery in the case of Georgia. If it is considered normal and not coercive then it is not being counted. Well, most of the bribery in Russia is in the voluntary category.
One example is the speed traps created in rural Russia where cops “extort” money once they catch you for “speeding”. This is coercive but such examples exist in the USA as well. It appears such corruption types get counted by Transparency International selectively.
I’ve read – in a Georgian media source, the Georgia Media Centre, if I recall correctly – that data sources for Transparency International in Georgia are all in Tbilisi and are all offices of the Georgian government or businesses connected to it. Hence, corruption is wildly underreported.
However, contrary to what you imply, personal experiences, and specifically business corruption, aren’t the be all and end all of corruption.
No, but personal experiences are the be all and end all when you say that calculated figures don’t pass the face validity test.
I assumed you were talking about personal experiences with business corruption. And the face validity test was far from my only criticism of the CPI.
Well, have to admit that relative alignment of countries in your list complies with my intuitive perception of the issue quite well. Respectable effort!
What is a bribe?
In Germany and other countries, it’s a widespread custom to give Christmas presents to business contacts. In Germany, there’s also a regulation how expensive presents can be in order not to be considered corruption. These regulations are regularly ignored. Recipents of such gifts include politicians, but also journalists. There’s no implicit demand for certain decisions included, just a feel good and be nice to each other. That of course does affect public works and information in a very subtil way as they create a group of “dependants”. The distribution is not so much via families paying bakshish, but via companies giving families presents, so it might not turn up in all these indices.
Hello AK, I’ve been reading your blog for a few weeks now and it is always very interesting reading. This new index seems interesting and it would be nice if more countries could be added. Have you thought about using The Latin American Public Opinion Project (LAPOP) and other similar regional barometers? From what I recall they sometimes ask respondents similar questions to what would be found in the Global Corruption Barometer (basically – have you paid a bribe? have you been asked to pay a bribe? etc). A good way to determine if LAPOP could be used to augment your GCB results would be to compare the results of countries covered in both (such as Brazil, Argentina, Mexico, Peru, Chile, Canada, the United States, Colombia and Venezuela) and see if they match.
Thanks for the suggestion, Hunter.
I will consider using it when compiling the 2011 Index. 🙂
Hi AK, it’s the first time I comment here but I’ve been reading your blog for a few months already and it’s quickly becoming one of my favorite blogs, good job!
I like your index more than the infamous CPI. That being said, I think Indonesia ranking at #36, just 6 places behind Poland is hardly believable. I’m Polish and I used to live in Indonesia for some time, and I gotta say that the difference in corruption prevalence was huge. Three years ago the corruption in Indonesia was at levels that are almost unimaginable to an EU citizen so unless some truly monumental change took place there in those three years since I’ve last been there I don’t think they should be ranked anywhere close to Chile and Czech Republic. The data may be skewed because corruption in Indonesia is mostly voluntary. That is, on one hand you would hardly ever come across someone openly demanding a bribe but on the other you can bribe pretty much anyone if you want to and achieve unbelievable things for laughably low bribes*, all it takes is some contacts and know-how.
*An example: I was offered an Indonesian ID card – a real one made in a government office from original materials, only with fake personal details – for mere $30! I don’t know how much it would cost one to pull a similar trick in Poland but I’m guessing the price would be in the thousands if it is at all possible.
…
Greetings AK from AKL !
I came to your blog via a Google alert I have for Transparency International and am well impressed with your CRI work, especially as it rates New Zealand well down the table rather than at the top.
However my warm fuzzies came to a screeching halt when I read your comment from 25th May 2011:
“No, I don’t have business experience outside the US (which is quite clean to date).”
Am I misreading or missing something?
Yes, the US does have the admirable Foreign Corrupt Practices Act, and an enviable record of investigation and enforcement. However efforts like the FCPA dwindle to fig leaf status when most of the corruption is domestic, not foreign. This is well documented, with no less than Time magazine noting that, on the strength of onshore jurisdictions like Delaware alone, ‘ arguably ‘ earn the US top (or, more accurately, bottom) listing as most corrupt in the world.
Maybe I need to read further to understand your comment more fully. However I wonder if in your efforts to attain completeness via correlation and compilation of various indicative indices, the CRI misses more concrete measures, such as the KPMG Fraud Barometer, which measures actual dollar values for prosecutions.
Admittedly, the KPMG efforts are rather selective and do not yet include the US, so I’ve shot my own argument in the foot, but there are other fraud watchdogs who keep an eye on such things.
Nothing more from me than a vague impression, and am glad to indulge my confirmation bias when it comes to the short comings of the TI CPI. I first began doubting that august institution when it issued a report on my adoptive home country of the Cook Islands – co-authored by the sister of a local lawyer who was up to his eyeballs in the country’s biggest corruption scandal, a $50 million Italian hotel project, but did not mention it once, attributing corruption to “cultural” values.
Ha.
I wrote to various TI bigwigs asking about their report author selection process, got no reply, then emailed all 91 country chapters asking for their own processes, but only got one response, from Papua New Guinea.
So, yes, scepticism well deserved. I’ll be reading lots more here to ensure I haven’t taken your comment out of context and look forward to learning more.
Jason Brown
Auckland (AKL), New Zealand
Australia is far less corrupt in almost every way than the USA, yet Australia and the USA have a very similar ranking. That can’t be right. Australians don’t even tip, because we think it’s corrupt. And there’s no way a crazy ideological dictatorship like Israel is the same as Australia. Also Palestine is run by Israel and part of Israel, so you need to average the Palestinian data into the Israeli data.