Review R W Johnson and Lawrence Schlemmer (eds) Launching Democracy in South Africa: the first open election, April 1994, New Haven and London: Yale University Press, 1996. Charles Crothers The 1994 South African election, the first 'democratic' election in this troubled nation's history, was a major dramatic show attended closely by a wide-ranging audience. It is appropriate that a major review of the experience of this election has been developed by a group of political analysts and very handsomely (if expensively) published by Yale University Press. The study was commissioned by the Institute for Multi-Party Democracy, co-sponsored by that Institute and a slew of overseas governmental and local capitalist financial supporters and carried out through a network of locally-based scholars. The book is a further stage of this project. It is organised in 13 chapters which provide scene-setting, historical, contemporaneous and prospective chapters together with regional treatments (Western Cape, KwaZulu-Natal - three chapters - and Gauteng - two chapters). Although a range of perspectives is deployed through the network of scholars, the two editors ensure some considerable consistency by contributing to eight out of 13 chapters and by writing all the national-level contributions themselves (apart from Chris de Kock's useful retrospective on trends in opinion and party support). The book portrays itself as an objective, social science-based record of the election and the fairly immediate electoral context within which it was set. It is difficult to detect any clear orientating question around which the book is organised, except for the broad interest in how this 'founding election' was carried out and the extent to which it was shaped by pre-existing social structures, and had in turn a subsequent impact on them. The tome carries no particularly discernible theoretical or ideological freight, despite a vigorous condemnation of the book for anti-ANC bias by Shuia Marks in her review for Higher Education Review (which in turn drew an even more vigorous defensive and counter- punching response from the editors in the following week's edition). Since Shula Marks failed to indicate examples of passages which drew her wrath, it is difficult to sustain her attack since the text carries few obvious markers of the ideological assumptions of its writers. 96 TRANSFORMATION 31 (1996) CROTHERS REVIEW The study is based on multiple sources, including electoral results, media accounts, observer reports and in particular a most impressive battery of surveys. The main weight of data is based on a couple of nation-wide political surveys, backed up by several region-specific studies and then a couple of post-election surveys which allow some tracing of post-election effects. The results suggest that despite considerable difficulties, it was possible for there to be reasonably fair and free elections in South Africa: several areas where marked improvement is called for are in terms of institutional development and also the building up of an electoral culture amongst the newly-enfranchsied citizenry of South Africa. It is argued that a new political sociology in South Africa has emerged which is little more than a straight 'racial census'. Indeed, of the major parties only the NP draws any praise for having secured a reasonably broad base across different race groups. It is noted that there is a 'class inversion' in the South African political landscape (at least amongst Black voters) with the ANC drawing more middle-class support for a more radical and socialist programme, whereas IFP supporters voted for that party supposedly in support of a 'robust capitalism' (oddly) coupled with highly traditional values. The book is well-written and carefully edited. There are some superb passages, in particular Johnson's vivid set of vignettes on the election scene in KwaZulu-Natal (chapter ten). Chris de Kock's time-series lay out a useful time-pattern in the development of socio-political attitudes in South Africa from the late-1980s through the early-1990s, and this over-time pattern is examined in relation to changing party support in the next chapter of the book. The difficulty of the volume is that with the enormity of the task it seeks to tackle, and without a clearcut theoretical framework to secure coherence, it sprawls and splutters. While a feast of the material needed to understand the election is laid out throughout the book, there is little guidance as to how the smorgasbord is best imbibed. Let us begin a critique with the sources of finance for the study, and how these are presented. The acknowledgments are remarkably coy about the role of the HSRC. In most nations, government social research insitutions stay well clear of overtly political issues, or at best treat them gingerly. A common distinction is between electoral and political issues: governments feel free to encourage voting and research to enhance tum-out and voter education, but stay clear of partisan politics. This book is based on a very solid base of HSRC survey research work into the arena of partisan politics, as well as covering more broadly electoral matters. Clearly, some of the outside funding received paid for aspects of these surveys, but there is no doubt that the (potentially) public cupboard of survey data has been thoroughly rifled by chapter authors to substantiate their analyses. 97 TRANSFORMATION 31 (1996) REVIEW CROTHERS In particular, de Kock's chapter mobilising an extraordinary pantry-full of politically-relevant HSRC studies. In itself, this central role of the HSRC is not problematic, especially as the data-sets involved are being gradually released for secondary analysis through SADA. However, this level of governmental social research involvement in political matters is quite remarkable in a comparative context, as is the apparent unconcern of HSRC at this large-scale scouring of its data-bases for partially private, albeit scholarly, purposes. (The book's co-editor, Lawrence Schlemmer was Vice-President of HSRC through the period during which most data was being collected, and this establishes a very clear link.) Despite the amazing data-base of survey material collected for the study, methodological standards retain the apparently standard South African social science methodological unconcern for quality-standards in relation to data. There is (almost) no reporting of survey designs, response-rates, response-biases, sampling errors or significance-testing (despite often quite small sample sizes), let alone the useful insights which can often be garnered from the debriefing of field staff. Moreover, the standard of data-analysis is very basic: cross-tabulations, although regression analyses are occasionally hinted at. Again, in itself, low methodological standards may not be a major problem (and publishers can be loath to have tomes gummed up with reams of methodological detail) but the failure of authors to mobilise data to support crucial arguments is a definite failing (as I will show with several examples). The political sociology of developed countries has developed an impressive array of models to trace the often weak social links between social background variables, policy interests, electorate and party institutional characteristics, party loyalties, leadership pulls and pushes and party vote choice. The theoretical model deployed by the authors to handle the South African situation is a most truncated and blunt portion of this apparatus. On the social background side, the work is driven only by race, and provincial location; no other social background characteristic is more than briefly mentioned (this is a striking example of an approach castigated by Taylor and Orkin, 1995). Only whether or not someone voted, and their party choice, was of interest to the authors, since policy choices are not generally considered (except perhaps as bundles indicated by party stances - although see pages 357-60 for a description of citizens' views). In general, the authors are uniformly unconcerned with tracing weak social effects on electoral involvement and voting choices, and strongly concerned with establishing to what extent 'strong' social effects could be sufficiently decoupled from politics to allow reasonable individual political choices to be achieved. This is the abiding and trenchant interest of the book. There is one exception to my strictures: the chapter on the Western Cape explores the interrelationships 98 TRANSFORMATION 31 (1996) CROTHERS REVIEW between social structure and politics in an interesting way. It is a pity that the approach of this chapter was not taken up more broadly in other parts of this study. Examining the 1994 election as a 'founding election' provides something of a broader and comparative framework within which some points made in the book can be set. As the editors point out, other instances of 'founding elections' have often been attended by subsequent civil war, and often they succeed in installing strong elites effectively detached from their bases of political support, offset only by a weak political culture. In contrast, several features of the South African situation seem to auger well, or at least better. On the other hand, however muffled in language of respect for minorities, the authors unfailingly see the overwhelming electoral support for one major party as a result redolent with political vulnerabilities. There is a strong moralistic sentiment throughout many passages that South Africa is now too vulnerable to the power which a one-party state is able to wield, unbalanced by countervailing political forces. The lack of theoretical and methodological sophistication undermines some of the book's concluding arguments. The key results reported above do not even receive the benefit of direct empirical testing from the volume of survey data readily available! The 'racial census' and 'inverted class support' arguments are drawn after a somewhat cursory review of electoral results (which at best allow only an indirect test). Most of the book works steadily through the long time-sequence of the run-up to the election, followed by the lead-in, the election period itself and its aftermath. Provincial peculiarities and intersting points are tracked down. One example of one of these side-shows are the several pages of the book which are, correctly, spent on probing into one of the more embarrassing issues in the election: the contrast between survey data (which showed a strong majority for ANC) and the KwaZulu-Natal political settlement (in which IFP was accorded a narrow majority). Johnson takes the latter as 'correct' (although there are no strong grounds for so doing) and suggests, through some highly speculative 'detective' work into some of his survey data, that the very large discrepancy (incidentally, surprisingly confirmed also by post-election surveys) is due not just to IFP-supporting respondents not declaring their support, but actually concealing their support by falsely claiming to support the ANC. The stimulus for this difficulty experienced by crypto-IFP supporters is seen as arising from the characteristics of the interviewers used in the study (interviewers clearly had much the same socio-demographic characteristics as ANC cadres). This argument rests on the empirically unsupported claim that IFP supporters were more scared of potential ANC harassment were their preference TRANSFORMATION 31 (1996) 99 REVIEW CROTHERS to be known than the reverse. Again, there is the curious reluctance to actually display the data on which these speculations are based. We have to be content with inferences and simulations. There may well be some (or even much) truth in Johnson's speculations, although given the very large gap between the election results and the survey data to be explained away they fail to be sufficiently convincing. Oddly, though, Johnson overlooks a useful point (subsequently confirmed by Durban metropolitan social research) that he makes only a few pages earlier: that the urban (and therefore differentially ANC-supporting) component of the KwaZulu-Natal population had been considerably overestimated. Appropriate re-weighting of the results might be a more easily acceptable explanation of the difference. With the benefit of hindsight afforded by the subsequent May 1996 KwaZulu-Natal local elections, it can also be suggested that the reality of political support between JJFP and ANC fell somewhere between the survey results and the official settlement. This volume is able to put a massively detailed survey-data flesh onto the bare-bone outlines of earlier-published treatments of the South Africa's founding election (eg the Reynold's collection: 1994). It is extremely useful to have such a compendium of factual material. But the limited deployment of theoretical and methodological expertise and sophistication renders the volume inadequate to the task it set itself. Some crucial questions are not posed, and the answers to others remain problematic. Although the range of information employed has been artificially foreshortened by keeping a limited focus, an understanding of the election has been smothered by information rather than illuminated by it. Extensive and more deeply analytical reworking of this material is urgently needed to ensure that understanding of its role in South African society is extended. REFERENCES Reynolds, Andrew (ed) (1994) Election '94: the campaigns, results andfuture prospects, London and Cape Town: James Currey and David Philip. Taylor, Rupert and Mark Orion (1995) 'The racialisation of social scientific research on South Afiica, South African Sociological Review 7(2). 1 °° TRANSFORMATION 31 (1996)