Chapter 1

Introduction

During the latter part of 2000 (August-October), more than 150 ‘experts’ were interviewed as part of a survey by the Institute for Security Studies (ISS) on the causes of and control measures to fight corruption in South Africa. This telephone survey was based on a comprehensive questionnaire administered by Markinor and funded by the European Union.

The purpose of the survey was primarily to collect new data, albeit qualitative, on corruption issues in South Africa. There is a general feeling of frustration among researchers working in this field over the fact that data about corruption is widely dispersed in more general questionnaires, such as the Idasa and Afrobarometer public opinion surveys, the Human Sciences Research Council (HSRC) household surveys and the national victim survey. A unique questionnaire was therefore compiled with the assistance of Idasa after having scoured the worldwide web and other resources for examples of international corruption surveys. The final survey instrument consisted of 28 questions dealing with all aspects of the corruption phenomenon and, for the first time in South Africa, people’s opinions were asked on the causes of corruption.

A second reason for conducting the survey was to stimulate the policy impasse that had resulted after the April 1999 anti-corruption summit. There have been some developments in anti-corruption policy initiatives since then, including the signing of a two-year programme with the United Nations Office for Drug Control and Crime Prevention (UNODCCP) in March 2001, the establishment of the Public Service Task Team mandated to produce a national anti-corruption strategy for the public sector, and the launch of the National Anti-Corruption Forum in June 2001. However, it was felt that a survey of this nature may assist in prioritising action on resolutions adopted by the April 1999 summit (see appendix 3 for a copy of the 1999 summit resolutions). On the basis of expert opinion and research, policy makers may be able to prioritise practical interventions that have to be made with limited resources, rather than to continue drawing up ‘pie in the sky’ lists of ways to fight corruption.

Who are these ‘experts’ and why should their opinions carry any weight? Firstly, they are individuals who attended one or more of the major anti-corruption conferences held over the past few years in South Africa. It can be assumed that people attending such conferences are at least professionally interested in the subject of corruption, and at best in a position to influence policy debate around the issue. Secondly, they are a highly educated group, with over half of the 154 experts interviewed holding post-graduate qualifications (27 honours, 42 masters and 12 doctors degrees among them). Slightly left of centre ideologically, and representing all races and sectors, it can be assumed that their opinions on corruption may make a useful contribution to the policy debate about corruption, its causes and how to control it raging, in South Africa.

Methodology

In 2000, after an open tender procedure, the ISS commissioned Markinor to conduct a survey among experts in the field of corruption.

Dr Bob Mattes, at the time of Idasa Public Opinion Surveys, and Lala Camerer of the Organised Crime and Corruption Programme of the ISS designed the questionnaire based on the following corruption surveys:

Huberts International Expert Panel: Public Corruption and Fraud, 19961


Idasa: Southern Africa Democracy Barometer, 1999-2000


World Bank Institute: Anti-Corruption Household Survey


International Crime Victim Survey, 1999


Seligon & Diaz-Briquets: Nicaragua Corruption Surveys, 1996 & 1998


Independent Commission Against Corruption: New South Wales Survey of Views on Reporting Workplace Corruption


Independent Commission Against Corruption: Annual Hong Kong Household Survey, 1998


Hong Kong Federation of Youth Groups: Youth Poll on Corruption


Gallup/Magyar: Gallup Pilot Study on Corruption

The draft questionnaire was circulated for comment by electronic mail to a number of researchers (both nationally and internationally) working on anti-corruption issues.

The sample of potential respondents was provided by the ISS and was compiled from a database of sources to obtain a universe from which to target the ‘experts’ to be interviewed in the survey. Sources included:

Fighting corruption: Strategies for prevention, published by the Public Service Commission, which included a list of the names of individuals and organisations which participated in the Public Sector Anti-Corruption Summit (Cape Town, 10-11 November 1998).



Fighting corruption: Towards a national integrity strategy, published by the Public Service Commission, which included a list of the names of individuals and organisations who participated in the National Anti-Corruption Summit (Cape Town, 14-15 April 1999).



A list of conference delegates who attended the 9th International Anti-Corruption Conference (Durban, 10-15 October 1999), supplied by the conference organisers. Only delegates on the list based locally were added to the expert database.



The initial sample was boosted by names of persons who attended a Transparency SA National Civil Society Anti-Corruption workshop (Kempton Park, 25 August 2000).

Using a database computer programme, a list of names was compiled from the above sources. Markinor received an initial sample of 672 names and telephone numbers of experts in the field of corruption. In total, 154 interviews were completed.

All potential respondents received a faxed notification letter to introduce the survey to them and to alert them that they would be contacted. A script of the questionnaire was created and administered by means of a computer-assisted telephone interviewing system (CATI) (see appendix 1).

Fieldwork was a real challenge as interviewers were often required to track down respondents. The CATI script was developed in such a way that Markinor could make multiple calls before the point was reached when a potential respondent was deemed to be unavailable.

In total, 3 789 contacts were made to obtain 154 successful contacts. In other words, approximately 24 calls were required to secure one completed interview. Of those successfully contacted, 191 (27%) persons refused to participate in the survey. Unfortunately, their reasons were not recorded. Fieldwork was conducted in English through telephone administered executive interviews. Fieldworkers were specially briefed and trained to deal with the sensitive subject matter of corruption.

The first phase of the survey was conducted between 17 August 2000 and 20 September 2000. The second phase was completed from 3-6 October 2000.

Problems encountered included:

In numerous cases, contact details were incomplete.

There was a considerable time-delay between the conferences and the survey.

There is a relatively high degree of movement among members of the sample group.

Encouragingly, 97% of those who agreed to be interviewed requested a copy of the survey results. A number of letters were also received from high-profile individuals (including a minister and a deputy minister) who, although not available to be interviewed personally, were willing to acknowledge the fact that such a survey was being undertaken, and encouraged its execution.

The survey data was reviewed in June 2001 for accuracy of interpretation by DRA-Development.

Expert surveys

In certain areas of social science research that require specialised knowledge, samples are drawn not from the general population, but rather from an allegedly knowledgeable part of the population, which serves as the universe. These so-called ‘expert’ samples have some advantages over the general public, such as concentrated knowledge or at least interest in the subject matter. However, they also have disadvantages. In particular, the small sample size limits precision.2

In a recent paper criticising the Transparency International Corruption Perception Index, Michael Johnston, an American academic who has written extensively on corruption, noted that:

"the potential of any research to produce rich and useful insights depends fundamentally upon careful design and honest application, not upon the apparent simplicity of its methods or results. The task now is to bring evidence of many sorts together into discussions of corruption that can match the comparative reach of most statistical indices with the complexity of corruption itself, and of the societies it affects most."3

As such, this type of qualitative data collection based on expert opinion also has a role to play in elucidating the understanding of corruption.

In South Africa, knowledge on corruption is incomplete and limited. While a survey among ‘experts in the field’ will not bring definitive answers to the multiple relevant questions around corruption and its control, it is nevertheless hoped that it will facilitate, as well as stimulate further discussion. Experts are likely to have more informed views on the extent of corruption and its causes and solutions. An inventory of views on the extent and seriousness of corruption in South Africa, the causes which are considered the most important and on the perceptions of the effectiveness of anti-corruption strategies which have been tried or proposed, provides new food for thought for those frustrated by an inability to offer any information-based insights into the corruption phenomenon.

Information elicited from expert opinions has to be seen in its proper context. This data reflects the opinions of respondents. By selecting specific respondents, these opinions are considered to be ‘expert’ opinions. Views say something about reality, but must not be confused with it.4 Despite the limitations of the study, the information obtained may be of some use by adding to the knowledge and contributing to the picture of the corruption question in South Africa. It presents, at least, some kind of informed reflection.

While the mere record of the perceived problems or priorities as indicated by a number of ‘experts’ should not be used as the only basis for policy interventions, these findings could be used together with other sources of information in order to promote a more informed approach to the problem of fighting corruption.

Survey demographics

The questionnaire provided the gender, age, race, home language, sectoral occupation, educational level, religion and ideological orientation of the 154 respondents (see table 1).

Gender, age, race, language and occupation

The majority of respondents (79%) were male and 21% of the sample were female (see table 1). The age profile of respondents indicates that the largest proportion of the sample was between 35 and 50 years of age, with slightly less over 50 years and the remainder spread below 35 years of age.

The majority of the sample was white (56%), with over a quarter (27%) African, 9% coloured and 8% Asian. For analytical purposes, the African, coloured and Asian categories were collapsed into one category of ‘black’ respondents, making up 44% of the sample. Where significant differences were identified within the racial categorisation, these are discussed in relation to a specific question.

English (47%) was the most common home language followed by Afrikaans (27%) and African languages (23%), including Zulu, Xhosa, Tswana, North Sotho and South Sotho. Only four respondents, mainly from the donor community, spoke other home languages.

Respondents represented a broad range of occupational sectors, including public, private and non-profit. The public sector was the most represented, accounting for almost half (43%) of the experts interviewed (figure 1). This was followed by the civil society and non-governmental organisation (NGO) sector (33%), the private business sector (21%) and other (3%).

Academic qualification, religion and ideological orientation

The majority (53%) of respondents had a post-graduate qualification, 25% had a matric qualification and 22% had a bachelors’ degree (figure 2). Among the post-graduate group, 27 respondents had an honours degree, 42 a masters degree and 12 a doctor’s degree.

The majority of experts (77%) defined themselves as Christian, 6% as other, 6% as Moslem, 4% as atheist, 3% as agnostic, 3% refused to answer and 1% as Jewish.

Politically, the experts were most likely (28%) to define themselves in the centre (5) with an overall mean of 4.36, slightly to the left of centre on a scale from 1 (left) to 10 (right).

Structure of the report

Largely following the design of the questionnaire, the report is broken down into several focus areas:

conceptual and practical understandings of corruption;

the extent, location and seriousness of corruption;

conditions for and causes of corruption;

evaluation of policy responses to fight corruption;

evaluation of anti-corruption agencies; and

evaluation of anti-corruption strategies.

Each of the focus areas documents the main findings of key questions and provides a brief analysis of the data. A number of graphs are provided in the text.

Other than a few instances where cross-tabulations by variables such as race or sector are significant, the data is largely left as collected. The small numbers make inferences about sub-groups within the data not advisable (ie not the small numbers themselves).