INTRODUCTION

The Human Sciences Research Council (HSRC) and the Institute for Security Studies (ISS) recently completed their second collaborative nation-wide public opinion survey on South African attitudes on security-related aspects. The previous survey was undertaken in May 1995 and the results distributed in August 1995. The fieldwork for the latest survey was done in October 1996.

At the time, the 1995 survey was the first national public survey to investigate and report on South African attitudes on security issues since the 1994 elections, and it received extensive media coverage. This second survey follows and builds upon the results of the 1995 survey, although a number of questions have been refined and changed where circumstances deemed such changes to be appropriate.

This article presents only the results of public attitudes regarding two issues from the 1996 survey results, namely, undocumented migration and policing/crime issues, and is the second article in the African Security Review on the survey results.¹

METHODOLOGY AND RESEARCH DESIGN

The HSRC Omnibus survey is done quarterly and its purpose is to give clients an opportunity to participate in a national survey at a low cost. The questions relating to the security services which were included in this survey were the result of two workshops between staff members of the HSRC and ISS which were held during August/September 1996. The core questions were originally the same as those which were used during 1995, but amended where appropriate.

Two separate questionnaires were administered during this survey to two probability samples of 2 200 respondents each. Every effort was made to ensure that the data was representative of the South African population.

The fieldwork for the survey was undertaken nationally from 7 to 28 October 1996. The data was collected by specially trained interviewers using structured questionnaires during some 2 200 personal, face-to-face interviews. Interviews were often done after hours to ensure that the respondent that had been drawn, was available. If the specific individual was not home during the first visit, but was available during the fieldwork period, an appointment was made and the person concerned, revisited. If nobody in the household qualified, or was available during the fieldwork period, the household was substituted.²

Interviews were furthermore conducted in the respondents' choice of language.³

Trained co-ordinators, appointed by MarkData, were allocated to different regions in order to co-ordinate the fieldwork (data collection).⁴ Each co-ordinator was responsible for a group of interviewers who were recruited under close supervision and subsequently briefed according to specific instructions given by MarkData.⁵ The sample design was initiated by Prof Laurence Schlemmer, Ms Tertia van der Walt and Dr Mathilda du Toit and formalised by Mr B Vukasovic. The capturing, processing and verification of the data was done by the Computer Centre of the HSRC.⁶

The sample allocation for the survey was done proportional to the adjusted 1991 population census figures with a few exceptions.⁷ The visiting points were drawn by means of multiple stage cluster probability sample design. The household and respondents were selected with a random grid from qualifying household members.

The universe for this sample design was all South African residents of 18 years and older. All nine provinces were included in the survey, with care taken to include both rural and urban areas. The socio-economic classification of the respondents was represented as follows: tribal, traditional, rural in former self-governing areas and TBVC states; squatter areas, urban (metropolitan and non-metropolitan); hostels, hotels, boarding schools, etc., in urban areas (metropolitan and non-metropolitan); former townships for so-called Coloureds (metropolitan and non-metropolitan); former townships for Asians (metropolitan and non-metropolitan); former townships for Blacks (metropolitan and non-metropolitan); towns and cities (non-metropolitan); towns and cities including flats (metropolitan); and rural areas (excluding the former self-governing areas and TBVC states).

An additional sub-sample was introduced for live-in domestic and other workers, based on the incidence of households employing domestic workers as found in previous Omnibus surveys. Domestic workers who formed part of the main sample (e.g. those living at home) were interviewed as such.

Disproportion was introduced to give a minimum number of 120 respondents per province. The minimum number of Asians in the overall sample was fixed at 120.⁸

The sample realisation for this Omnibus survey was one hundred per cent and no major obstacles were encountered during the fieldwork.

UNDOCUMENTED MIGRATION

Background

Despite strenuous efforts by South African academics,⁹ mass population movements generally, and the specific issue of undocumented migration, is still shrouded in myth as opposed to hard facts. Part of the problem, of course, is that clandestine migration, by its very nature, is not very open to academic inquiry, especially in quantifiable terms. The data collected as part of the HSRC/ISS survey debunks several myths pertaining to anti-foreigner sentiment among South Africa's body politics, as it relates to illegal aliens. This, in turn, holds several domestic and regional implications.

Current South African policy on migration is vague and ranges from strategies of control (for instance enforced repatriation, greater police and army personnel on the border and lethal electric border fences) to strategies of accommodation (for example, the recent Cabinet decision to legalise the presence of illegal immigrants who have resided in the country for longer than five years, who have been gainfully employed, with no criminal record, or who are married to a South African spouse).¹⁰

The reasons for this ambiguity are not hard to find. In part it reflects the dichotomy in policy circles between the African National Congress (ANC) and the Inkatha Freedom Party (IFP). Generally, the IFP is seen as more restrictive in its approach towards illegal aliens and its leader, Dr Mangosuthu Buthelezi, in his capacity as Minister of Home Affairs, has repeatedly called for stronger control measures. The ANC, on the other hand, is seen to be more liberal. ANC members often argue that illegal immigration can only be stopped if the root causes for such forced population displacements are addressed. Hence they argue that there is a need for a regional Reconstruction and Development Programme for Southern Africa which would address poverty as one of the root causes of such population movements. This policy is probably informed by the ANC's own experiences as a liberation movement in exile in neighbouring states. But, it is also true that the ANC as the Government of South Africa has shown some hardening of its attitudes on the question of illegal aliens. Thus South Africa was one of the Southern African Development Community (SADC) countries opposing the 1995 SADC Protocol on the Free Movement of People. It is our contention that the strong anti-illegal immigrant sentiments revealed in the survey would have an impact on the policy-making process and would result in the further hardening of attitudes among policy-makers.

Survey Results

Xenophobia is viewed world-wide as a sentiment generally confined to individuals at the lower end of the socio-economic and the educational spectrum.¹¹ Various reasons have been put forward to account for this phenomenon. It has been argued that individuals at the lower end of the spectrum have less of a world view due to the fact that international travel opportunities are limited. In addition, the development of a world view will also be proscribed by a lack of access to literature on exotic places and foreign cultures.¹² Others have argued that anti-illegal immigrant sentiments among this section of the population can be accounted for on economic grounds. It is argued that illegal migrants – who are hired at low wages, work long hours and are resistant to unionisation – are a threat to the job security of unskilled or low-skilled nationals in any country.¹³ These nationals then respond to this economic threat by adopting a xenophobic posture.

While this analysis might be true for Jean-Marie le Pen's National Front in France, Germany's neo-Nazis, Britain's skinheads or the United States' right-wing militias; this survey has cast doubt on its utility in South Africa. As Graph 1 indicates, anti-illegal immigrant sentiments increase as educational qualifications increase.

Graph 1: Whether the influx of illegal immigrants is a good or a bad thing analysed by educational level

Two reasons could possibly account for this inversion. In the first instance, it could be argued that as people are better educated, they are more conscious of the perceived threat that illegal immigration holds for the country. Secondly, it could be argued that whites, because of the legacy of apartheid, are better educated than their non-white counterparts. As such, they are better represented among the more educated sector of South African society. Therefore, the survey would largely represent their xenophobic attitude towards illegal aliens. As Graph 2 illustrates, 93 per cent of all whites believe that the influx of illegal immigrants is a bad thing.

Graph 2 Whether the influx of illegal immigrants is a good or a bad thing analysed by race

It has often been assumed that Zulu speakers are more xenophobic than other groups, while Xhosa speakers are seen to be less xenophobic. This perception is closely related to the fact that Zulu speakers are seen to be largely associated with the IFP, whose hard-line attitude towards illegal immigrants was illustrated in September 1994, when the Inkatha Youth Brigade threatened that if the Government failed to take strong action against illegal aliens, it would do so itself.¹⁴ Conversely, Xhosa speakers are seen to be largely associated with the ANC, who are perceived to be more liberal in terms of its approach. This is perhaps best encapsulated by the ANC's former Deputy Minister of Home Affairs, Mr Penuel Maduna, who was quoted as saying: "History has shown us time and time again that hunger and fear are driving forces which are much stronger than even the most sophisticated alien control measures. South Africa has become the country of survival for many."¹⁵

However, the survey results have challenged these traditional viewpoints. Graph 3 indicates that Xhosa speakers are more xenophobic than Zulu speakers – 70 per cent of the Xhosa speakers thought illegal immigration is a bad thing, as opposed 50 per cent of the Zulu speakers.

Graph 3: Whether the influx of illegal immigrants is a good or a bad thing analysed by language

In addition, Graph 4 also underlines this truism: ANC supporters are far more xenophobic than their IFP counterparts, with 60 per cent of ANC supporters believing illegal immigration to be a bad thing as opposed to the IFP's 49 per cent. This clearly illustrates that the leadership in both parties are not in tune with the aspirations of grass roots supporters.

Graph 4: Whether the influx of illegal immigrants is a good or a bad thing analysed by party support

How does one account for the difference in perception among Zulu and Xhosa speakers on the issue of illegal immigration? It seems that the reasons are to be found in the patterns of migration in Southern Africa over the past 125 000 years. Centripetal and centrifugal patterns of migration have been the norm over the millennia in the region. The arrival of Dutch colonists served only to reinforce this trend. The early nineteenth century witnessed the rise of a militarist Zulu kingdom under Shaka. This was accompanied by a period of tremendous upheaval known by the Sotho word Lifaqane which conveys the notion of forced removal. Thus, the rise of Shaka's kingdom saw the fleeing of other tribes away from the militarist regional hegemon: Rolong, Pedi Hlubi, Kora, Ndebele, Tlokwa, Mpondo, Swazi, Ngoni, Griqua, Sotho and Tswana were all part of these great population movements. These centrifugal pressures were further strengthened by the 'Great Trek': when thousands of Boer families decided, from the 1830s, to leave their British colonial masters in the Cape Colony and to trek into the interior to establish their own Afrikaner homelands.¹⁶

The significance of this legacy of ethnic diaspora for present day Southern Africa cannot be overstated. Consider the number of Swazi inside South Africa as opposed to in Swaziland; the number of Tswana speakers inside South Africa as opposed to in Botswana; the number of Basothos in South Africa as opposed to in Lesotho. There are various indications to suggest that ties of clan, lineage or tribe take precedence over the ties of citizenship generated by the state in Africa, and that these serve as facilitating factors to encourage illegal immigration. For instance, part of the reason for large concentrations of Mozambicans in the former homeland of Gazankulu (now part of the Northern Province) is that, being Shangaan, they share a certain cultural similarity with the indigenous population.

This transnational ethnic consciousness is clearly to be seen in Graph 5, with those provinces more isolated in regional terms, such as the Eastern, Western and Northern Cape being more xenophobic than those provinces who share borders with the country's regional neighbours, and whose host populations share a common cultural heritage.

Graph 5: Whether the influx of illegal immigrants is a good or a bad thing analysed by province

This could also possibly explain why Zulu respondents are less xenophobic than their Xhosa counterparts. The Zulu, after all, share a common identity with the Swazi in Swaziland, and with the Ndebele in Southern Zimbabwe. This provides them with such a transnational ethnic consciousness which sees illegal immigrants more as people than aliens. The same cannot be said of the Xhosa, largely concentrated in their Eastern Cape hinterland with the Indian Ocean at their backs. Their geographic position prevents such a regional consciousness from developing. Coupled with this is the fact that the Eastern Cape is one of the poorest provinces in the country, further resulting in fears over the 'alien bogeyman' stealing jobs from South African workers.

Conclusion

On the issue of undocumented migration these survey results have debunked certain myths surrounding illegal aliens and xenophobia. At other levels, it has confirmed that South Africans are, generally, parochial and look upon undocumented migration with suspicion, if not open hostility. This is underscored by the fact that almost two-thirds of respondents (65 per cent) indicated that illegal immigration was a 'bad' or 'very bad' thing for the country. In addition, 80 per cent of respondents are in favour of the government trying to curb a further influx of illegal migrants by strengthening border patrols, 65 per cent are in favour of enforced repatriation and 73 per cent are in favour of penalising employers who hire illegal immigrants.

These findings hold deeper significance than just on the issue of undocumented migrants. They add an element of caution on grandiose designs of regional integration, as South Africans are thinking either nationally, subnationally or in ethnic terms, rather than regionally. These high levels of xenophobia are also filtering into government policy on the vexing issue of illegal immigration. Recently there has been a further hardening of attitudes towards illegal aliens. This can be seen in Dr Buthelezi's recent call on government service departments to "request the identity documents or passports of all foreigners requesting services by the government and in this way ensure that they do not gain access to services in short supply to our own people."¹⁷ While it is true that the influx of illegal aliens holds a socio-economic threat to South Africa's security and that some control measures are necessary, it is also true that if this is done unilaterally and without regards to the human rights regime, Pretoria will continue to live 'against' as opposed to 'with' its neighbours. This, in turn, does not bode well for SADC and South Africa's own diplomatic efforts for securing a viable security regime in the region.

Clearly, solutions to the 'problem' of undocumented migration should ideally be focused on regional, national, provincial and local reconstruction, job creation and economic growth rather than bureaucratic and technical control over undocumented migration. Yet, in the interim, it may be electric border fences rather than the economies of the Southern African countries that may be switched on.

POLICING/CRIME

Introduction

Crime and policing are emerging as the key policy issues in post-apartheid South Africa. While some categories of property crime (where the wealthy are largely the victims) began to stabilise during 1996, reported levels of violent crimes, such as rape and serious assault (in which poorer communities are generally victimised), continue to rise. Trends throughout the country have not been uniform – provincial crime totals suggest that Gauteng has the highest level of property crime per head of population while the Northern Cape has the highest level of violent crime.

Government responses to the issue of growing criminality have been centred on the National Crime Prevention Strategy (NCPS). The NCPS is a detailed plan designed to implement crime prevention measures across four key departments in the criminal justice system – Safety and Security, Justice, Correctional Services and Welfare. The plan has been slow to get off the ground. One year after it was first announced, little actual implementation has taken place and the NCPS (as predicted at its launch) has been plagued by its overly bureaucratic committee structure and the difficulties of working across departments.

The Department of Safety and Security and the South African Police Service (SAPS) in particular have increasingly come under both public and political pressure to improve their performance. The SAPS has been characterised by poor management and over-centralised organisation and has inherited poorly trained and equipped 'bantustan' police agencies. Declining morale within the SAPS has corresponded with declining public perceptions of its ability to fight crime. Given the poor level of policing service in the townships prior to 1994 some improvements have been made here – most notably is the more equitable distribution of personnel resources between the townships and the suburbs. Important structural challenges – most notable is the devolution of policing powers – are the key to the introduction of more effective and service-oriented policing. While these reforms are in the pipeline, they will have in themselves a further disruptive influence on the ability of the SAPS to deliver on the ground.

Survey Results

Respondents were asked whether the police were more effective in combating crime than two years ago. Their answers are reflected in Graph 6. The overall response in this respect was that the respondents thought the police were less effective than they were two or three years ago, with 42 per cent of them stating that the police were now 'less effective' or 'much less effective' than two or three years ago. A total of 32 per cent of respondents contended that the police were 'more effective' or 'much more effective' than two or three years ago.

Graph 6: Whether the police are now more effective in combating crime than two years ago

Broken down by race, white respondents are the most inclined to state that the police are either 'much less effective' or 'less effective' in combating crime than two years ago, with 67 per cent of them expressing that view, followed by coloured respondents (53 per cent) and Asian respondents (36 per cent). The only respondents who were inclined to the view that the police were more effective now than two or three years ago, was the black category. Altogether 38 per cent of them stated that the police were now 'much more effective' or 'more effective', while 36 per cent stated the opposite view.

The response regarding effectiveness of the police was also analysed according to language (see Graph 7). The figure reveals that the following language groups have indicated the police to be less effective now than two or three years ago: Afrikaans (61 per cent), English (53 per cent), both Afrikaans and English (50 per cent), Zulus (41 per cent) and Tswanas (35 per cent). Those language groups indicating that the Government is now more effective in combating crime than two or three years ago, are: Southern Sothos (45 per cent), Tswanas (44 per cent), Northern Sothos (42 per cent), 'other' (37 per cent) and Xhosa speakers (35 per cent).

Graph 7: Whether the police are now more effective in combating crime than two years ago by language group

A number of questions were also asked relating to the willingness of South Africans to assist in combating crime. The responses to these questions are briefly presented in Graphs 8 to 11.

Altogether 53 per cent of respondents were willing or very willing to perform military duties in their own areas, while 32 per cent were unwilling or very unwilling. The supporters of the Freedom Front (FF) were the most willing to do part-time military service in their own areas (72 per cent), followed by the supporters of the Pan-Africanist Congress (PAC) (60 per cent), the ANC (58 per cent), the IFP (49 per cent), the National Party (NP) (47 per cent) and the Democratic Party (DP) (37 per cent). (See Graph 8.)

Graph 8: Whether respondents would be willing to do part-time military service in own area analysed by party support

Black people were most likely to say they were willing to perform military service in their own areas (55 per cent). They were followed by Asians (50 per cent), coloured people (48 per cent) and whites (44 per cent). (See Graph 9.)

Graph 9: Whether respondents were willing to do part-time military service in their own area analysed by race

Supporters of the FF and the PAC indicated the greatest willingness to perform part-time military service in support of the SAPS, with 56 per cent of each expressing that view. They were followed in this view by the supporters of the ANC (54 per cent). Supporters of the DP (27 per cent) were the least likely to perform part-time military service in support of the SAPS. (See Graph 10.)

Graph 10: Whether respondents would be willing to do part-time military service in support of the SAPS analysed by party support

Analysed by race, black people were most inclined to perform part-time military service in support of the SAPS (51 per cent), followed by coloured and Asian people (43 per cent and 42 per cent), while whites (40 per cent) were the least willing to do part-time military service in support of the SAPS. (See Graph 11.)

Graph 11: Whether respondents were willing to do part-time military service in support of the SAPS analysed by race

Conclusion

Most significantly, the survey results demonstrate the degree to which public perceptions of police effectiveness have become negative. This has serious consequences: reduced confidence in the SAPS' ability to respond to high levels of crime may herald a growth in self-policing and vigilante initiatives. Indeed, there is already considerable evidence from across the country that this is the case. However, this general conclusion needs to be applied with varying degrees of emphasis to different South Africans.

White South Africans have displayed a dramatic turnaround over the last decade in their view of police effectiveness. The majority of whites (and especially Afrikaans speaking people) now believe that the police no longer effectively protect them from crime. Again, over the longer term this suggests that the 'laager culture' of the suburbs will intensify as private and physical security measures continue to be implemented. If it is accepted that uncertainties around safety and security are the single biggest cause of a loss of white confidence in the new order (as has been demonstrated in other empirical surveys) the decline in confidence in the police by white respondents is a serious cause for concern.

The finding that a growing number of black respondents (although Zulu speakers are least inclined to take this view) are more inclined to believe in the effectiveness of the police is as important and is consistent with other survey material. This suggests some success in the police transformation process in the townships – even if it points less to greater effectiveness than a decline in harassment. Despite this conclusion, the overall finding of the survey – that South Africans believe that policing effectiveness has declined – suggests that, unless rapid improvements are made, positive views of the SAPS will be undercut over time.

It is remarkable that age, level of education and gender do not provide significant differentials on the issue of police effectiveness. As expected, older people are more likely to believe that the police are not (or cannot) do their job effectively, but the divergence in opinion across other categories is surprisingly small. More educated people (or at least those with a Standard 9 or 10) are more likely to judge police effectiveness as in decline. South African women (unlike most comparative survey material elsewhere in the world) are slightly less likely to believe that police effectiveness is waning.

Race and class are the most significant determinants of whether or not South Africans view police effectiveness to have declined. This is consistent with a developing trend since 1994 – indeed, the decline of violence after 1994 has (according to recent survey evidence) increased feelings of safety and security among many black South Africans. While no breakdown of the findings according to political party has been produced, it is probable that any such results would reflect that (comparatively wealthy) urban supporters of the DP and NP would consider police effectiveness to have declined the most. It appears to be highly likely that the issue of Government's ability to end crime will be the key issue separating black and white voters in the 1999 election.

ENDNOTES

The previous article, on public attitudes towards women in the security agencies and the issue of language usage, appeared in African Security Review, 6(3), June 1997.



Substitution of the visiting point was only allowed for the following reasons:

* refusal by the selected respondent;

* empty premises, e.g. no building or occupied structure;

* nobody on the stand qualified for the survey;

* after three visits were done at different dates and times, and the respondent could still not be found at home;

* respondent could not communicate with the interviewer because of the use of a foreign language; or

* respondent was not physically/mentally able to be interviewed.



The questionnaire was printed in English and Afrikaans and translated during training into other languages used in the local area.



The co-ordinators were trained between 3 and 4 October 1996 by MarkData in the Western Cape, Northern Cape, Free State, KwaZulu/Natal, Mpumalanga, Northern Province, Gauteng and North West. Clients were invited to attend the training session and to give additional information pertaining to their questions.



Interviewers were recruited to conduct interviews as close as possible to their own residential area. Preference was given to interviewers with prior interviewing experience. Co-ordinators were responsible for the recruitment of the interviewers. All interviewers had to comply with the following minimum requirements: a matric educational qualification or equivalent; and fully bilingual or multilingual in order to speak the relevant language in a particular fieldwork area.



The co-ordinator's responsibility was to ensure that all the questionnaires sent to MarkData had already been checked for correct completion and that 20 per cent field control was done. From the office, a 20 per cent fieldwork control (back check) was done once again on each interviewer's questionnaires. The co-ordinators also supervised the fieldwork process to ensure that the correct procedure was maintained. Details of the fieldwork progress were communicated to MarkData on a regular basis.



Census enumerator areas and similar areas were used as the clusters. The number of respondents drawn per enumerator area was either 4 or 8. All clusters were drawn with probability (a serial sampling procedure was applied).



The factor weights to be applied to the captured data set were derived at the sampling stage. The aim of the factor weights was to correct the disproportions which were incorporated in the sample design. The only weighting targets were the stratification variables: 'province' and 'socio-economic' category. RIM weighting is generally used in two cases: when the purpose is to weight data according to various characteristics, but the relationship of the intersection of those characteristics is unknown; or when there are not enough respondents to fill all the possible cells (e.g. males who had passed Standard 8, 35-44 years old, employed part-time). As the RIM weighting runs, it attempts to distort each variable as little as possible, while still trying to attain all the desired proportions among the characteristics. The 'Root Mean Square' figure was used to determine how much distortion had been introduced.

The following variables were submitted to the RIM weighting procedure at various stages: age; gender; education; employment status; occupation; marital status; language; and population group.

Close examination of outputs suggested that a satisfactory solution cannot be obtained despite the increased number of iterations (too high between the weights ratio) and only 'age', 'gender', 'education' and 'population group' variables were retained. RIM weighting targets for the former TBVC states were estimated according to population characteristics drawn from the available census data. Weighted and unweighted frequencies on all biographical variables are available.



See for instance E Leistner, Migration of High-level African manpower to South Africa, Africa Insight, 23(3), 1993; M Reitzes, Alien Issues, Indicator South Africa, 12(1), 1994; H Solomon, Migration in Southern Africa: A Comparative Perspective, Indian Journal of African Studies, 5(2), 1993; H Toolo and L Bethlehem, Labour Migration to South Africa, paper presented to the Workshop on Labour Migration to South Africa, National Labour and Economic Development Institute (NALEDI), Johannesburg, 31 August 1994; P Vale and H Solomon, Migration and Global Change: Understanding the Tsunami Effect, paper presented to the seminar on Migration: Sources, Patterns, Implications, South African Institute for International Affairs, Johannesburg, 6 May 1993; C Dolan, Policy Challenges for the New South Africa, Southern African Migration: Domestic and Regional Policy Implications, Workshop Proceedings, 14, Centre for Policy Studies, Johannesburg, 1995, pp. 53-54.



H Solomon, Strategic Perspectives on Illegal Immigration into South Africa, African Security Review, 5(4), 1996, pp. 12-14.



See for instance S Collinson, Europe and International Migration, Pinter Publishers for the Royal Institute of International Affairs, London, 1994.



These views were expressed by some of the participants at the Salzburg Seminar on Involuntary Migration, Schloss Leopoldskron, Salzburg, Austria, 8-15 July 1995.



See for instance Toolo and Bethlehem, op. cit.



Solomon, 1996, op. cit., p. 11.



M Reitzes, Divided on the 'Demon': Immigration Policy since the Election, Policy Review Series, 8(9), 1995, p. 15.



T R H Davenport, South Africa: A Modern History, Macmillan, Hong Kong, 1991, pp. 12-15.



Government to Deny Services to Illegal Immigrants, Fast Facts, 4, April 1997, p.1.