CPI 2018: Comments, questions, validity?

In: e-Civis Mundi # 82, April (II) 2019

CPI 2018: comments, questions, validity?

Civis Mundi Digitaal #82

door Michel van Hulten
25 Years after Transparency International (TI) was founded, TI issued on 29 January 2019 her next Corruption Perceptions Index, CPI 2018, continuing a series which began in 1995. See: https://www.transparency.org/cpi2018 .
Everybody should wonder how it is possible that can survive already for 25 years a research result which is also used as a tool for policy-makers, that offers only ‘perceptions’ of a global social-economic phenomenon, corruption. No facts.
Nevertheless numerous private and public policymakers (and researchers) believe in the reality of the scores and ranks offered to them by social-economic researchers who see countries climb or descent the ladders of more and less corruption, again without facts. Maybe that was acceptable in the earliest years of this endeavour when corruption indeed was hidden utmost. But now, with so many court proceedings, so many verdicts, penalties to be paid in the hundreds of millions, are there still not enough facts available allowing to qualify countries and/or population groups, and to quantify the presence of corruption on the basis of observed confirmed facts, in stead of believing that e.g. in Europe the presence of corruption is higher, the further South you travel? Because this is what many think? Swedes less corrupt than Germans, and they less corrupt than Italians from North-Italy while Southern Italians can without doubt be qualified as utterly corrupt? All this based on ‘perceptions’. I call this in Dutch ‘hersenspinsels’, the dictionary has many words for this concept: fiction, chimère, fantasma, etcetera, see also delusion, hallucination, figment.
Each successive edition of the CPI clearly states that no research has been done for that edition based on facts about corruption. Each one of the annual reports is called Corruption Perceptions Index, except the first one in 1995 which dared to name itself Corruption Index. That name caused at the time that quite a number of governments raised strong objections about the score and the resulting rank order reached by their countries. Interestingly, they saw their scores and the rank number received, comparing them with other countries, as the reflection of facts! Transparency International had to give in and changed the name, adding the word ‘Perceptions’. These governments did not want to be named as ‘corrupt’. Moreover the name ‘Perception’ also comes in plural in the titles of the annual reports in order to emphasize the wide diversity of thoughts on the subject.
 Less pertinence
The opening phrase for the CPI 2018 edition: “The index, which ranks 180 countries and territories by their perceived levels of public sector corruption according to experts and businesspeople, [ … ]”, claims less pertinence than the opening phrase for the CPI 2017 edition which in stead of using the 2018-qualification ‘perceived levels’ of corruption, called itself the leading global indicator of public sector corruption, a rather firm statement, do we go back to 1995?
I quote “Since its inception in 1995, the Corruption Perceptions Index, Transparency International’s flagship research product, has become the leading global indicator of public sector corruption.”
And indeed, nobody can deny that the CPI has become for many decisionmakers and students the measure to be used if you wish to be taken serious in your actions against corruption, or in your studies into the phenomenon ’corruption’ in our economic, political and social life.
The poor are corrupt?
However, I find most remarkable that the impression anyone may get from this annual list of around 180 countries ranked according to a perceived level of corruption, that – in general – this list can be seen running parallel to more or less similar lists of countries ranked according to an order from rich to poor.
Is that a harmless co-incidence? The poorest countries of the world are the most corrupt? The richer a country is, the less it is corrupt? Or is the suggested causality that the more corrupt you are, the poorer you will be? And how does this relates to TI’s definition of corruption as ‘the abuse of public power for private gain’? Is ‘power’ the issue and is ‘rich’ only a substitute? If so, how can the poor be the most corrupt?
The TI-ranking order shows that none of the OECD member countries is in the lower half of the CPI-list. And 38 of the 54 African countries figure in that lower part.
See also http://worldhappiness.report/ed/2018/, a landmark survey of the state of global happiness.[1] And recently we can see the Corruption Challenges Index 2019, first issued in 2017 by Risk Advisory see link: https://www.riskadvisory.com/campaigns/corruption-challenges-index-2019/. Later in this paper I will name more scores and rankings in other sources.
Of course, it is possible that this comparability reflects the reality in the world.
Is proof for ‘corruption’ already established through the work done by TI? I don’t think so.
Therefore the question should be asked: is this the reality, or is the similarity in ranking the outcome of the way research is conducted that makes this result inevitable? These questions lead us also to the question: is the result as TI publishes yearly, scientifically acceptable? And, if there are flaws in the research practiced by making this Index, can we repair?[2]
TI did already an unsuccessful effort to improve the quality of the yearly CPI reports between the annual reports of 2011 and 2012. In preparation, the highly qualified International Index Advisory Committee of TI was consulted, and a professional study was set up by Andres Gelman and Piero Stanig. They wrote Proposals to improve the index of corruption, August 27, 2010, and sent Comments regarding the views of the Advisory Committee, January 2011. As far as I could notice, all this did not materialize except for the proposal to move the comma in the scores with one digit. And I notice now that I cannot find any more on the internet neither on the TI-website these Proposals and Comments. Maybe I am not sufficient computer-educated to find them. If they are still there, anyhow they are difficult to find!
In my reporting about the CPI 1995-2005 which I published in 2007, I presented also my suggestions for improvements.[3]

Missing sources?

What I miss in the introductory reporting by TI on her website is an answer to the question why there are these 13 sources? And why only these 13 sources? Which other possible sources did TI consider for inclusion in her analysis but did reject (and, if so, for what reason?). The least we may expect is that over the years TI looked into all possibilities to include also others? And explains the choices made. I have some suggestions.
In 2012 TI used as one of the sources for the CPI of that year, her own Bribe Payers Index (BPI) which evaluates the supply side of corruption, i.e. ‘the likelihood of firms from the world’s industrialized countries to bribe abroad’. I quote more from the TI-website: ‘Some of the world’s richest countries turn a blind eye to corruption. When their companies use bribes to win business abroad and are allowed to get away with it, governments are effectively complicit in exporting corruption.’ (Pro memorie: The Netherlands scored first on the BPI – which is worst – in 2011).
‘Our Bribe Payers Index ranks the world’s wealthiest countries by the propensity of their firms to bribe abroad and looks at which industrial sectors are the worst offenders. The index is based on the views of thousands of senior business executives from developed and developing countries.’
This year (Jan 2019) the BPI 2018 was available, but is not used as a source. Why?
Same question
I also wonder and ask, why for instance has not been used the Mo Ibrahim Index of African Governance (http://iiag.online/)? Would not this have been better than to use two sources (#1, African Development Bank and #10, World Bank), which both use data from the same fieldwork: the CPIA – Country Policy and Institutional Assessment? One of these would have been enough?
The more so as these two sources, used for the CPI 2018, collected their data from the same source but that the one, the AfDB, used the 2016 edition of the CPIA, and the other, the WB, used the 2017 edition of the CPIA? I also think that a warning note should have been included that for the CPI 2018 the same AfDB assessment 2016 had been used as was already used also for the CPI 2017? Similar remark with regard to the source #3, Bertelsmann Stiftung Transformation Index, which had been used for the CPI 2017 and is now again used for the 2018 CPI.
Another source which is produced by the TI-Secretariat annually is not used either. Why? Title: Exporting Corruption, progress report 2018 (or possibly, if that one came too late available, the 2017 edition) Assessing enforcement of the OECD Anti-Bribery Convention, by Gillian Dell and Andrew McDevitt; Contributors: Huma Haider, Julius Hinks, Casey Kelso and Johannes Wendt; Editor: Stephanie Debere. This report was produced by the Transparency International Secretariat as part of a project funded by the Siemens Integrity Initiative. ISBN of full report: 978-3-96076-097-9. https://www.transparency.org/files/content/publication/Download_a_short_version_of_the_report.pdf
Why did IHS stop deliveries of data?
I found it noteworthy that TI remarks on her website under the full disclosure of the ‘Sources’ for the CPI 2018 (and last year also for CPI 2017) that ‘The data for CPI 2018 from IHS Global Insight was accessed through the World Bank World Governance Indicators portal, as IHS Global Insight stopped providing data to Transparency International in 2015.’
The least we should know is why IHS Global Insight stopped providing data to TI, although I am happy that a bypass of this obstacle has been created and that those data anyhow can be used as they are available at the WB, as is indicated correctly.
Relation corruption – democracy?
What I do not want to discuss in this paper is the claim laid by TI on the opening page of the CPI 2018, that the Index “reveals that the continued failure of most countries to significantly control corruption is contributing to a crisis in democracy around the world.” This part of the opening statement of the published CPI 2018 is also a good subject for closer scrutiny, but it has nothing to do with the questions to be posed about the validity of the CPI itself.
In forthcoming articles in following editions of e-Civis Mundi I will subsequently ask questions related to four subjects which are treated in the 2018 CPI report and focus my analysis of the CPI 2018 on:
1. Who does the work in collecting and processing the data used in classifying the countries ?
2. How precise is ‘2018’? When have data been collected and processed?
3. What definition for ‘Corruption’ has been used and which questions have been asked?
4. How ‘global’ is the coverage of the report?
I hope that all readers will agree that my arguments are strong enough so that they can agree with already so many others and with me, that the qualifications given by TI and the resulting ranking order as published under the title Corruption Perceptions Index cannot pass any quality check, and therefore are unreliable and harmful to the fight against corruption.

[1] Helliwell, J., Layard, R., & Sachs, J. (2018). World Happiness Report 2018, New York: Sustainable Development Solutions Network. http://worldhappiness.report/ed/2018/

[2] See my earlier reports on more or less the same topic: Onverantwoord klasseren van ‘corruptheid’ van landen, by Michel van Hulten in Civis Mundi, e-magazine, https://www.civismundi.nl/index.php?p=artikel&aid=1998. This article in Dutch was expanded in an article in English: The ‘new’ CPI by TI, scientifically not defendable (december 4, 2013), http://www.corruptie.org/the-new-cpi-by-ti-scientifically-not-defendable-2/

[3] Earlier, in 2007, I had already compared and analysed all CPI-reports from 1995-2005 and published my critical report, go to: https://www.academia.edu/38000426/070131_CPI_analysis_and_comments_final_.docx



Gelieve dit veld leeg te laten.