Measurements and markets: deconstructing the corruption perception index

The Authors

William De Maria, UQ Business School, The University of Queensland, Brisbane, Australia

Abstract

Purpose – The paper aims to examine how the measurement of African “corruption” has been manipulated to serve western economic interests.

Design/methodology/approach – In depth secondary source analysis within a post-colonial framework.

Findings – The most popular measure of corruption, Transparency International's corruption perception index (CPI), is a flawed instrument. Capable only of calculating proxies of corruption, the measure is oblivious to cultural variance and is business-centric in style. The CPI is embraced in good faith by African governments and donor organisations oblivious to its deeper purpose of serving western economic and geo-political interests under the guise of weeding out something falsely portrayed as a universal negative.

Practical implications – The paper will assist efforts to ground the anti-corruption effort in the realities of Africa.

Originality/value – The paper is part of a minority scholarship that seeks to provide space for the consideration of alternatives to the dominant conceptions of corruption and its measurement.

Article Type:

Research paper

Keyword(s):

Corruption; Measurement; Africa.

Journal:

International Journal of Public Sector Management

Volume:

Number:

Year:

2008

pp:

777-797

Copyright ©

Emerald Group Publishing Limited

ISSN:

0951-3558

Introduction

“Make poverty history.” This was Nelson Mandela's clarion call before 22,000 people in Trafalgar Square, London, on 3 February 2005. A month later the Commission on Africa (2005), set up by the British Prime Minister Tony Blair, released its long awaited report proclaiming that “African poverty and stagnation is the greatest tragedy of our time.” “Saving” Africa is now squarely on the international agenda. The price tag is big and a huge public administration edifice required. The Commission on Africa estimate, an additional US$25 billion per year, was accepted at the July 2005 G-8 meeting at Gleneagles and the September 2005 UN World Summit[1].

This unprecedented output is not simply about the West slowly rising to its greatest aid challenge since the Marshall Plan that followed World War II, for more is at stake then African well being. There is abundant anecdotal evidence that this new donor energy is following a golden rule; aid must serve first and foremost the West's economic and geo-political priorities[2]. A new consensus, “corruption” is inimical to first world economic development, is sending a nervous West back into Africa with patterns of engagement strikingly reminiscent of the bad days of colonial rule[3].

This consensus has a mesmeric grip on international aid decisions and accounts for billions of heavily tagged anti-“corruption” dollars now flowing annually from the West to Africa in a massive (yet grievously uncoordinated) public administration mission to rid the continent of “corruption”. The West is conducting this mission on its own terms and in its own style (Osabu-Kle, 2000, p. 516). The result is a rapidly growing anti-corruption movement, or “industry”, spreading under close western supervision across Africa (particularly sub-Saharan Africa) that is detached from the street level realities of Nairobi, Abuja, Luanda and the like[4].

Accompanying donor dollars to Africa is a very narrow and very western definition of “corruption”, which in turns pivots the equally narrow methods employed to measure it. Powerful definitions are the definitions of the powerful and in the history worn asymmetric encounters between Africa and the West the latter's international business-centric view of “corruption” prevails. Despite a valiant counter-insurgency from African scholars (Wiredu, 1996; Abdi, 1999; Osabu-Kle, 2000; Adekson, 2004; Mushanga, 2004; Prempeh, 2004) the indigenous prerogative to make the final call on what constitutes “corruption”, and by extension how to measure and control it, has been buried under a semantic and methodological tsunami that is flooding in from the West[5].

The paper will attempt to illustrate this point by focussing on the most prominent anti-corruption NGO in the world – Transparency International (TI) and particularly its “corruption” measurement instrument, the corruption perception index (CPI). Since 1995 this annual reckoning of (now) 163 countries is promulgated by the Berlin-based NGO in conjunction with the Internet Centre for Corruption Research at the University of Passau in Germany. The CPI is now the best known “measure” of “corruption” in the world. This is particularly so around that time each year when TI publishes its new rankings. Then we see a flurry of activity in media, government, NGO and academic circles. Position is everything. A low spot on this “corruption” league table has been credited with the fall of the Bhutto Government in Pakistan in 1996, the fall of the Banzer Government in Bolivia a year later (Galtung, 2006, p. 107) and grounding the anti-“corruption” drive in many countries, for example Turkey (Kurtulis, 2002). It has also been recognised for putting “corruption” on the front page of newspapers throughout the developing world and acting as an incentive to governments to get or stay in the top ten of the index (Galtung, 2006, p. 108; Johnston, 2006, p. 1).

For too long CPI results have been unquestionably accepted as a valid and reliable indicator of “corruption” by African governments, donors, media and academia (Wilhelm, 2002; You and Khagram, 2005, p. 141; Serra, 2006; Knack, 2007). The time is ripe for a critical reappraisal. A central aim of this paper then is to excavate the hidden assumptions that drive the CPI Index and structure the debate about Africa “corruption” and its antidotes. The paper offers a counterpoint to the western ownership of the definition of African “corruption”. It leans towards Bayart's (1993, p. 235) view that “corruption” in Africa is all to do with “the politics of the belly”; in other words the daily struggles against poverty, disease and exploitation[6]. This needs based approach to “corruption” (as opposed to a greed-based approach) turns our gaze towards local histories, local culture and local economic conditions, as the prime incubators of “corruption”. It also turns our attention towards a wider role for the African Union (AU) in the sponsorship of culturally-sensitive approaches to the issue of “corruption”.

The paper proceeds in the following way. The choice of post-colonialism as the theoretical backbone of the paper is explained first. The CPI is then deconstructed. This is the main part of the paper. The final section presents a general commentary on an alternative epistemology to the CPI. This revolves around two speculations; a reassertion of the African voice in the “corruption” debate and the active involvement of the AU in the management of African “corruption”.

Digging with post-colonial theory

Post-colonial theory seeks to expose and subvert the sovereignty of western constructs-epistemological, moral, economic and political (Prasad, 2003, p. 7)[7]. Its special contribution to development and anti-“corruption” literatures is that it keeps a close watch on power asymmetries between Africa and the West. When this is articulated into the paper the inquiry becomes one of understanding the forced diffusion of the single western approach to “corruption” that has been accumulating since colonial times. We can use post-colonial theory like a naïve tourist uses a travel guide to reconnoitre unfamiliar places. Post-colonial theory offers a priori guidance about what to expect when we traverse the “landscape” of “corruption” and its measurement. Two proto-typical western features stand out in the post-colonial guide to “corruption”. The first is the self-determined prerogative of the West to define and manage African “corruption” and the second is the value place on scientific measurement to provide accurate readings of this phenomenon.

The western prerogative

de Sardan (1999) and Sandholtz and Koetzle (1998) have pondered the enormous distance between judicial, political, media and donor condemnation of African “corruption” on one hand and its frequency, banalisation and outright cultural legitimation by ordinary African people on the other. Could it have something to do with the way “corruption” is defined? Philp has recently commented that this is not just a nice semantic point. It raises concerns about the very feasibility of the “corruption” measurement project (Philp, 2006, p. 50).

Post-colonial theory guides us towards constructions that are exogenously developed in the West and force fed into weak national administrative systems in Africa. This process becomes obvious with respect to the meaning attached to “corruption”. A specific definition of “corruption” created and then repeatedly endorsed in the western literature masquerades as a world applicable definition, and forces into oblivion other less powerful constructions of “corruption” (Duncan, 2006, p. 131). We will see in the analysis below that a standard off-the-shelf, business-centric, culture blind definition of “corruption” is active in the CPI polls. The force of culture, the vagaries of situation, the weight of history and the overwhelming evidence of business initiated “corruption” are swept aside or diminished in this definition[8].

Smith's study of the Igbo of Nigeria is a valuable but overlooked condemnation of this process of definitional imperialism through standardisation. He diverts our gaze from explanations of acts of deviance against a “neutral” state onto kinship realities where something called “corruption” is reproduced by ordinary people out of necessity. Smith (2001, p. 345) concluded from his study that ordinary Nigerians had a stake in reproducing “corruption” and that what may appear to be “corruption” to the onlooker is actually survivalist moral behaviour as people navigate Nigeria's clientelistic political economy.

Smith found that the Igbos accessed needed resources through reciprocity and obligation, practices long embedded in family, lineage and community. The western state has, over the last 130 years, moved to formalise reciprocity and obligation through mechanisms such as welfare states which provide basic resources to citizens as a human right. Judgements about what is and what is not “corrupt” are made as if the welfare state was part of Nigeria's political economy. At this level of cultural blindness “corruption” is falsely portrayed as the universal extraction of private benefit from public duty (Marquette, 2001, p. 6, fn 26; Rose-Ackerman, 2005, p. 290; Lambsdorff, 2006, p. 82) (emphasis added). The World Bank, an important CPI data source, sees “corruption” as “The abuse of public power for private gain.” Another CPI data source, the Economist Intelligence Unit, echoes with “Misuse of public office for personal (or party political) financial gain” (emphasis added).

This obsession with the “corruption” of the public agent can be traced back to 1989 with the publication of the influential works of Heidenheimer and Nye. Heidenheimer's (1989) three ideal types: public office-centred “corruption” market-centred “corruption” and public interest-centred “corruption” are variants of the crooked bureaucrat type. No space was made in his taxonomy for crooked business people (1989). Similarly, Nye's (1989, p. 966) definitions of “corruption” as “… behaviour which deviates from the formal duties of a public role … ” (emphasis added) set the definitional scene right up to the present.

“Corruption” then is falsely constructed as a trans-cultural “disease”, carried by (often minor) public officials that must be surgically removed worldwide for economic wellbeing (Hope, 2000, pp. 17, 19; Wolfensohn, 2000, p. 1; Commission on Africa, 2005, p. 32). Portraying “corruption” in this way does nothing to make us smart about how to prevent and contain this wrongdoing phenomenon in conditions of great social and economic variance. It simply licences a fantasy that “corruption” is (like all diseases) no respecter of culture. Riley is one author who commends this uniformity. He sees nothing “African” about African “corruption”. “It is”, he says, “part of the strong global growth in corruption” (Riley, 2000, p. 138). This opens the gate for standardised international intervention and collides with recent (but nevertheless unpopular) assessments which context “corruption” into specific histories. Michael (2004a, p. 1) for example treats “corruption” in Eastern Europe as specific to issues of transition from command to liberal economies. This is definitely not a history shared with Africa. Other literatures also dispute Riley's uniformity of “corruption” thesis (de Sardan, 1999, p. 25).

Notwithstanding methodological spats in the literature, the western consensus about the meaning of “corruption” is reproduced in broad epistemological agreements about how to measure it. The strength of the western approach to “corruption” measurement lies in the way it projects an aura of scientism.

The scientific imperative

A post-colonial critique of (western) science proceeds from the assumption that science, like child care and comedy, is a cultural construct that forms and is formed by political, economic, and social forces (Dubow, 2000). Seen this way science loses its universal position and competes in a marketplace of ideas. Post-colonial theory explains how science has been elevated from a local knowledge (localised in the institutions and values of the west) to its current presentations as the only knowledge worth having, via colonial and neo-colonial power relations (Escobar, 1995; Briggs and Sharp, 2004, p. 662). A central tenant in the ontology and epistemology of science is measurement.

Accurate measurement, as Philps says, depends on variation being measured against a constant ( p. 50). This is a scientific imperative. In its absence, conviction, subjectivism and bias rule. Only the most audacious researcher would claim a scientific measure of “corruption” is possible. Yet the pressure is on administrators to justify the allocation of millions of good governance dollars to Africa on something more then a whim. Enter the CPI, a business-biased calculation with a whiff of science.

The worldwide interest generated by the CPI is part of modern era optimism that everything can be measured (Babbie, 1995). The next section critiques this position. It examines assumptions buried in the CPI and considers its epistemological failings, particularly the reliance on proxy measurements, the dependence on perceptions of “corruption”, a construction of “corruption” through the prism of business and the falsely portrayed uni-modal shape of “corruption”.

“Corruption” measurement: lexicons of domination

Public administration and measurement are now tightly coupled through the logic of the management by measurement movement. The rise of new public management (Dibben et al., 2004) and the audit society (Power, 1999; Grey and Jenkins, 2004) indicates to some commentators that the “world of public management has become, first and foremost, a world of measurement” (Noordegraaf and Abma, 2003, p. 853; Razafindrakoto and Roubaud, 2006, p. 5). This can be quickly assessed by noting the current profusion of “good governance” indicators[9]. The difficulty is that many aspects of public management (such as “corruption”) are hard (impossible?) to measure since “many performances are invisible, unintended or contradictory” (Noordegraaf and Abma, 2003, p. 853).

Bevan and Hood (2006, p. 517) tell us that administration by targets and measured performance indicators is a form of control necessary for the governance of any complex system. The tight coupling between the CPI and anti-corruption programming can be visualised as such a complexity, moving through a four-step process:

zero tolerance for “corruption” is specified as the standard in advance (by western interests);
this standard is expressed in measurable form (data from CPI sources referred to below);
TI manages a system of measuring country-based performance against that specification (CPI); and
measured performance is linked to donor activity.

Embedded in this process are a number of inherited malfunctions germane to all scale-based measures of human activity (such as the CPI). Three are considered; target tunnel vision, surreptitious patterns of accolades and condemnations, and the measurement of epiphenomena when the real thing is beyond calculation.

Hitting the target-missing the point

The evidence is clear that “corruption” index positioning is beginning to influence the flow of anti-corruption resources. While TI says that its CPI should not be the basis of aid allocation, there is the strong possibility that donors are using it anyway as an allocation rationale (Galtung, 2006, p. 123). Only one organisation has so far revealed that it relies on “corruption” index positioning when determining aid eligibility. This is the US Congress funded Millennium Challenge Corporation (MCC). To receive MCC (2005) funding countries must perform above the medium within their peer group on the World Bank's Control of Corruption Index[10].

As the begging dog will embrace unnatural behavior (begging) for food reward, African governments have found themselves in some vast operant conditioning exercise whereby public policy deviations towards west endorsed anti-corruption strategies achieve the reward of aid inflows. So it is reasonable to expect that the politics of CPI targeting will narrow government behavior towards the reductionist fallacy that “good governance = no corruption” and thereby reduce performance where targets do not apply (e.g. public health). Bevan and Hood see this as “hitting the target and missing the point” ( p. 521). Future studies are needed to see how deep this service distortion runs in African governance. Distortion is one problem. The annual “corruption” Academy Awards is another.

And the winner is … ?

In 2006, the CPI was distilled from surveys in 163 countries[11]. The ultimate “winner” that year was Finland and the ultimate “loser” was Haiti (respectively perceived as the least and most corrupt countries in the world for 2006) (TI, 2006). The way the CPI runs from 0 – highly corrupt to 10 – highly clean fosters images of countries beyond salvation, or so flawless that no further improvements are needed. TI says that the CPI is not intended to “brand” any one country. This seems a rather naïve understanding about how people interpret indexed information. A more honest (but still flawed) approach is used by Freedom House. Its 2006 survey of “oppressive” countries is called The Worst of the Worst. The World's Most Repressive Countries 2006 (Freedom House, 2006a, b, c).

The CPI can be read in two ways; a naming and proclaiming process for the “clean” countries such as Finland, Iceland and New Zealand and a naming and shaming process for the “corrupt” countries such as Haiti, Myanmar and Iraq. The CPI puts upward pressure on those performing below the target level to do better, but also provides what Bevan and Hood call “a perverse incentive” for those doing better than the target to allow their performance to deteriorate to the standard ( p. 521).

Proxy surveys – close enough?

The CPI is widely misunderstood to measure “corruption” per se. It does not do this. Rather, it measures perceptions – a proxy version of the real thing. The question becomes “Are the measures of “corruption” proxies as good as measures of the substance of “corruption”? (Dreher et al., 2007, p. 444). Some recent studies suggest they are not. Mocan (2004) has found that the correlation between perceived corruption and actual corruption is low once other relevant factors are controlled for. In his Brazilian study, Weber Abramo has shown that perceived corruption is not related to bribery. Rather, perception-based indices reflect the quality of a country's institutions not its actual degree of corruption (Weber Abramo, 2005). Andvig (2005) has reported the same finding.

In one of the most forceful rebuttals of CPI type measures of “corruption” to date Razafindrakoto and Roubaud compared the results of direct surveys for six sub-Saharan African countries with perception-based indices. The first type of survey covered a sample of over 35,000 people and objectively measured the incidence of petty bureaucratic “corruption”. The researchers then polled 350 experts for their perceptions of “corruption” in the six countries. They found that experts were an unreliable source, systematically overestimating the incidence of “corruption” and overestimating the tolerance of local people to “corruption”. Furthermore, the researchers found clear evidence of ideological bias with experts rating countries according to their own political preferences and personal experiences (Razafindrakoto and Roubaud, 2006)[12].

When we look at the expert-reliant CPI we find that none of the nine survey sources used by TI to construct the annual index measure the phenomenon of “corruption” directly. Rather these sources measure what they separately consider to be epiphenomenal occurrences that in their personal view point to the existence of “corruption”. In this flawed reasoning, a conclusion that “corruption” exists is necessitated by, or reached from, these previously known epiphenomenal facts. The overconfident expectation is that what is measured is bound to be the secondary manifestation of “corruption”. For example, the World Bank (IDA) (2005, p. 41) survey used as a CPI source (Table I) purports to measure the proxy values of accountability and transparency in African Governments because a “high degree of accountability and transparency discourages corruption”. Thus, according to this reasoning, a detection of low transparency is, ipso facto a detection of “corruption”.

This is a dangerous course to follow. Babbie calls it reification; thinking of a common manifestation of a concept as if it was the concept itself (pp. 116-118). Take the example of anxiety. A well-known symptom (proxy) is elevated heart beat. But this is also a well-know symptom of mitral valve prolapse. So the proxy list has to be expanded until agreement is reached that the concurrence of a unique set of symptoms allows for a separate and specific diagnosis of anxiety. Such concurrence has not been established in any “corruption” research or polling to date as the following scan of TI's 2006 data sources shows.

From this scan of CPI sources we note the following. Of the nine data sources used by TI only six collect data in Africa[13]. Of these six, four are international risk management and business survey organizations. They are reporting on much the same thing; business environment intelligence. For example, TI uses the Institute for Management Development's World Competitiveness Year Book as well as the World Economic Forum's Global Competitiveness Report. This monochromatic flow of data significantly trims down informational diversity and determines the CPI as a guide for capital exposure, notwithstanding its masquerade as a measurement of “corruption”. It also explains why these polls show a high degree of correspondence with other. This fact is usually presented in support of the high reliability of the polls and a trans-national consensus about what constitutes “corruption” (Galtung, 2006, p. 112; Ng, 2006, p. 823; Treisman, 2000) when it is just an artifact of sampling sameness.

As the CPI data are from private polling companies it comes at a price. For example, MIG's data from its Grey Area Dynamics (country × country risk analyses) costs something in the order of US$3,000 per year to registered clients. Similarly the World Economic Forum's Executive Opinion Survey, which is part of their Global Competitiveness Report, cost US$300 per copy. These price structures would restrict the use of the data to governments, companies and cashed-up NGOs. The data are confidential to the client-purchaser and very little methodological information is posted on the sources' web sites[14]. As very limited public access to the polling instruments and data manipulations exists no public scrutiny is possible. This raises an accountability issue for TI, a champion of transparency, as it continues to use secret polling to construct its CPI.

Given that international commerce's big anxiety is the hard to predict behavior of African governments towards investment, it is not surprising that CPI data sources are mainly interested in the corruption of public figures. Very little data are forthcoming about foreign and domestic business “corruption”. For example, of the 137 questions in the World Economic Forum's Executive Opinion Survey only eight pertain to business “corruption”[15]. Similarly, only a small amount of the poll data applies to “corruption”. The Executive Opinion Survey devotes only 10 per cent of its questions to “corruption”[16].

The polled “corruption” data are usually made up of ticks on scales. Question 9.09 of the Executive Opinion Survey asks “In your country, how common are un-documented extra payments or bribes from one private firm to another to secure business”? The interviewee has seven options from “common” to “never occurs.” Respondents do not have to directly witness “corruption”. Answers based on personal impressions and hearsay are usually harvested. The Executive Opinion Survey mentioned above does not check whether responses come from direct experience, general impressions or hearsay from colleagues in the field. CPI's other data sources also do not go into these matters. This has major implications on the validity, reliability and the precision of the final results that are taken at face value by CPI consumers.

There are numerous other problems with perception dependent measures of “corruption”. One concerns the amplification and abbreviation of impressions about “corruption”. The former occurs when perceptions of “corruption” by business people respondents are untrustworthy because they may have been temporally elevated in the lead up to the completion of the survey simply by being exposed to media reportage on local “corruption” stories. CPI data methodologies do not attempt to control for this effect. The opposite effect, of under-perceiving “corruption” also happens in conditions of media oppression, and when “corruption” is conducted in a highly secretive manner and remote from the perceiver (Kalnins, 2005, p. 7; Galtung, 2006, p. 101).

A related failing is the impact on CPI poll data of self-interested decisions by business people not to report “corruption” for fear of losing favour with government and indigenous business contacts.

Johnston reports other subtle erosions of CPI validity. Business respondents, particularly those naïve to the workings of countries they are reporting on, could mistake a faction driven scandal as an outbreak of “corruption” (Johnston, 2006, p. 13). Johnston also gives credence to the “culture shock” factor. For example, a young executive from Finland (CPI low level “corruption”) newly arrived in Uganda answering poll questions about his or her perceptions of Ugandan “corruption” may talk up their “awareness”. However, a long-term expatriate resident from Hungary (CPI middle level “corruption”) may talk down their perceptions of Ugandan “corruption”, even though both may have had the same matters in mind when they responded to the polls. These observations underscore deeply troubled methodologies harvesting biased business judgments of African “corruption”. This point is developed next.

CPI and international business

There are a number of frameworks available to develop our thinking about corruption. They organize around different key concepts; institutional “corruption” (LaPolombara, 1994), democratic “corruption” (Drury et al., 2006), donor “corruption” (Doig, 2006) and human rights “corruption” (Akcay, 2006). A literature review suggests that academic interest in the relationship between human rights and “corruption” is very weak and interest in the relationship between business and “corruption” very strong. I would argue that the injury “corruption” can do to business (particularly the western transnational variety) is now the dominant discourse in anti-corruption program planning. This bias is reflected in the way questions are posed in the polls used by TI. For example, the Political & Economic Risk Consultancy (2005) asks “To what extent does corruption exist in the country in which you are posted in a way that detracts from the business environment for foreign companies?” (Lambsdorff, 1998, emphasis added).

This over-weighted interest in business welfare in the polls reflects a strong theme in the literature. “Corruption” is seen to draw talent away from entrepreneurship and innovation (Murphy et al., 1991), increase borrowing costs for governments and businesses in emerging economies (Ciocchini et al., 2003), retard business growth more then taxation (Fisman and Svensson, 2007), foster allocative inefficiencies (Lien, 1990), maintain social and political inequality (Enderwick, 2005) and lead to lower levels of investment (Mauro, 1998; Wei, 1997). It is with these results in mind that TI's chairman and founder, ex-World bank employee Peter Eigen, recently defended the prominence given in the CPI to the business voice when he said:

The best available method of compiling comparative data (about “corruption”), therefore, is to build on the experience and perceptions of (business people) who are most directly confronted with the realities of corruption in a country (Bistrich, 2004).

Galtung (2006, p. 112) expresses the same view offering no proof that business perception is any more valid then other perceptions, e.g. victims of “corruption”.

The CPI then is central intelligence to foreign businesses scoping new opportunities in Africa. The strength of the business paradigm in the CPI detected in Table I can also be seen in Table II which simply records the business focus in the mission orientations of the data sources, from statement culled directly from their separate web sites.

The current engagement between Africa and the West testifies to the resilience of the neo-colonial paradigm (de Maria, 2007, 2008a). Within that engagement a flawed measure of “corruption” has been heavily promoted using a highly sectionalized definition of “corruption”. The methodologically troubled pro-business CPI is part of a multilateral campaign to “expand and solidify the institutional foundations for a global market economy” (Bukovansky, 2006, p. 182). The final section of the paper considers a response to this issue.

The African Union: “Corruption” fighter?

Authors who have written in the post-colonial mould such as Said and Fanon have drawn our attention to the way western knowledge and procedure have achieved dominance through views propagated since colonial times. These views treat the West as rational, objective and progressive and Africa as irrational, subjective and archaically traditional (Frankel and Shenhav, 2003, p. 1537). From the headwaters of this ethnocentrism flow, as we have noted, western prerogatives to take action in Africa as it sees fit. Salbu, writing in a recent issue of the Yale Journal of International Law, offers a strong response to this matter. He argues that the intervention of the West in African “corruption” could be justified through the normative conception of the global village. However, he sees nothing other then cultural heterogeneity, which, he says, continues to disrupt efforts to address world problems as a single community (Salbu, 1999, p. 223, emphasis added). Western intervention into the affairs of Africa creates what Salbu calls “moral panic”, by which he means the dangers of intrusiveness, paternalism, imperialism and disrespect that comes about when a powerful state imposes its will on another state. How can the initiative about “corruption” be secured by Africa?

Perhaps, the answer lies with the AU; a bold, continent wide experiment gingerly unfolding from its headquarters in Addis Ababa. Africa got is first true continental structure in 2002[17]. The Union aspires to a common market, a pan-African federalism, a single currency, a single defence force as well as other federal institutions of state. One of AU's objectives is “To promote and defend African common positions on issues of interest to the continent and its peoples” (Constitutive Act of the African Union, 2000, Article 3(d)). This, along with the AU's new Convention on Combating Corruption and its even newer African Charter on Democracy, Elections and Governance suggests that the mandate is emerging for the AU to attend to “corruption” issues that threaten the broad AU mission.

The AU Convention on Combating Corruption came into force in August 2006. Objective 3 say the AU is to “Coordinate and harmonize the policies and legislation between State Parties for the purposes of prevention, detection, punishment and eradication of corruption on the continent”. However, as presently constructed the AU Convention focuses on prescribing anti-corruption responsibilities to its members rather then to itself. There is provision within the Convention for the establishment of an Advisory Board on Corruption to promote anti-corruption on the Continent, research corruption and promoter harmonisation of public sector codes (among others). The African Charter on Democracy, Elections and Governance was adopted on 30 January 2007. Article 2(9) commits member states of the AU to “Promote the fight against corruption”.

However, there are enormous problems to overcome if the AU is to be the agency of first instance with respect to the measurement and management of “corruption”. The first concerns how successful will be the biggest public administration project that Africa has ever known-the development of full political and economic integration leading to the United States of Africa.

The main conclusion reached at the AU's Executive Council 9th Extraordinary Session held in November 2006 in Addis Ababa, was that:

[…] all Member States accept the United States of Africa as a common and desirable goal (but) differences exist over the modalities and time frame for achieving this goal and the appropriate pace of integration.

The AU Assembly in Accra, Ghana in July 2007 was devoted to what was called a “Grand Debate on the Union Government.”

There are problems of course, in getting the AU to assume “corruption” measurement and management responsibilities. The first concerns the high level of present conflict in Africa[18]. This suggests that arriving at a continent-wide consensus about the “corruption” is, and will remain, a thorny issue.

Another problem concerns the power of the AU to impose sanctions. It is noted that there are no sanctions in the AU Convention on Combating Corruption. This is obviously a very sensitive issue as it impinges on sovereignty. It seems like it will take many years for African countries to get used to the idea of up-drafting some of their sovereignty to the AU. The nine regional groupings currently active across the continent offer sovereignty sharing experiences as a dress rehearsal for effective membership in the AU[19].

A third problem concerns timing. Is it too late for the AU to exercise alternativeness with respect to the measurement and management of “corruption”? Messages coming out of AU headquarters suggest an endorsement of western approaches to “corruption”. The problems in achieving a United States of Africa and a continent-wide approach to “corruption” are real and complex. However, relying on extra-territorial direction cannot be an option for a continent wishing to move on from its colonial past. If a pan-African world is to be built it will need alternative epistemologies to tackle “corruption”.

Searching for an alternative to the CPI

Even such a prominent scholar-warrior for the western approach to “corruption”, Rose-Ackerman (2006, p. 291) has conceded that the direction of causation between poverty and “corruption” is not clear. Does poverty cause “corruption” or does “corruption” cause poverty? At this late stage in the paper the lid on this Pandora's Box here can only be slightly opened. The unpacked assumptions within the CPI logic that have been explored above suggest a strong leaning to the “corruption” causes poverty side of what is a massive controversy in the world of overseas aid and development. If a public administration slogan was needed to typify this approach it would be “Conquer corruption and all else will follow”. Hopefully, the weaknesses in this approach have been plausibly presented above.

If we change our research and policy gaze from “corruption” to poverty a number of significant things happen. First we stand to get much closed to the realities of Africa where good income outweighs good government. Secondly, economics as the paradigm controlling the “corruption” debate is supplanted by more culturally sensitive scholarships, predominantly ethnography and sociology (de Maria, 2008b). Third, we replace a fixation with measurement by a commitment to negotiation and dialogue. Fourthly, we re-consider long peripheralised constructs such as heritage, linguistics, space, religion and of course, the colonial experience for their utility in explaining behaviour commonly referred to as “corrupt”.

Research supporting the causative flow from income inequality to “corruption” is now being reported in the literature. You and Khagram (2005, p. 137) claim to have recently published the first cross-country study of the causal effect of income inequality on “corruption”[20]. They found income inequality to be a significant determinant of “corruption” and that poverty habituates norms of “corruption” as the “way things are done” ( p. 154).

Conclusion

The paper advanced the post-colonial position that “corruption” cannot be comprehended outside the experience, nor can it submit to empirical investigation as promoted by the management by measurement movement. “Corruption”, it was concluded, is essentially a complex, normative phenomenon (Bukovansky, 2002, p. 4) that plays cat and mouse even with the most sophisticated empirical designs. Research set up in scientific ways does not guarantee scientific measurement. To try and measure “corruption” misrepresents the subjective and political processes alive in its manifestation as well as writing down the potential for bias in the surveys used (You and Khagram, 2005, p. 154). The normative dimension of “corruption” is studiously avoided by the empirical researcher because it produces “noise” that could compromise the delicate “glassware” of empirical research design. The reality of cultural variance requires “corruption” to remain permanently the subject of disputation about moral standards that should apply to the bearers of public and private office.

Within that post-colonial context a central purpose emerged; to expose the business bias in the CPI and identify serious epistemological weakness in its methodology. The primacy of the western value of measurement wearing the apparel of science in African “corruption” discourse was targeted. In the prescient words of Joss (1993, p. 360); “The suspicion is … that the methodological requirements of empirical research [into “corruption”] ensures its irrelevance to the very question public administration most needs to address”. Joss is concerned that the search for a measurable “corruption” may lead empirical researchers (such as those who compile the CPI or who do secondary research relying on CPI figures) “to impoverish normatively rich concepts and hence degrade moral discourse” ( p. 360).

More attention needs to be given to providing alternatives to the western anti-“corruption” juggernaut. In the space available, the paper opened up two possibilities; a pan-Africanist drive against “corruption” through the AU and replacing the dominant economic paradigm with more culturally sensate frameworks. It is time for Africa to take charge of the measurement and management of its own “corruption”.

Table ICPI data sources: proxy surveys

Table IICPI data sources: mission statements

References

Abdi, A. (1999), "Identity formations and deformations in South Africa: a historical and contemporary overview", Journal of Black Studies, Vol. 30 pp.147-63.