Presented on February 26 1997 at a seminar held by the Centre for the Study of Violence and Reconciliation Lisa Vetten, The Sexual Harassment Education Project Introduction Reporting a rape to the South African Police Services (SAPS) is merely the start of a long and arduous struggle through the South African criminal justice system. Is the victim's trudge through this system ultimately justified? The general public would think not. Recent media reports highlighted an incident where a group of women allegedly castrated a man accused of rape. A spokesperson for the group said that they "were sick and tired of watching by (sic) as sex-related crimes escalated in their community."1 Their actions echo recent calls for the reimposition of the death penalty, or demands that castration be introduced as a form of sentencing for sex offenders.2 Such recommendations suggest a general perception that, although the law convicts sex offenders, it is far too lenient in sentencing them. There is also an assumption that harsh sentencing can prevent an offender from repeating his behaviour, and act as a frightful warning to any man contemplating sexual violence. Understandable as calls for vengeance are (particularly when made by rape survivors and their families), they are unlikely to address the problem of sexual violence, or prevent its continuation. A focus on sentencing as the cure for rape also neatly side-steps other important considerations - such as the ability of the courts to convict in the first place, as well as an exploration of those factors that reduce or increase the likelihood of a conviction. Rape in South Africa Rape is very much an unknown quantity and subject in South Africa - a situation hampering attempts both to prevent and address the problem in any meaningful way. The little that is known seems confined to information about the number of rapes reported annually to the SAPS - figures acknowledged to substantially underestimate the size of the problem. An unofficial estimate put out by the National Institute for Crime Prevention and Rehabilitation of Offenders (NICRO) suggests quite a different picture. They have calculated that only 1 in 20 rapes is ever reported to the police, leading the organisation to speculate that a rape occurs in South Africa every 83 seconds. The SAPS' estimate is even gloomier. They conjecture that only 1 in 36 rapes is reported, so implying that a rape occurs every 35 seconds (unfortunately it is not clear how NICRO or the SAPS arrived at these figures). Regardless of the 'real' figure, the current number of reported rapes alone is sufficiently dire for rape to be regarded as a priority crime warranting immediate action. Aim of the Study This study attempts to provide more information about rape by exploring two questions: How many rapists, nationally, are ultimately prosecuted and then convicted of their crimes? How do these rates compare to those of other violent crimes? Answers to these questions may provide some clues to a further question: what hampers the courts' ability to prosecute and convict in an effective, just manner? Previous Research in this Area Limited information is available on prosecution, conviction and sentencing rates in South Africa. It has been estimated that between fifty to fifty-five percent of rape prosecutions lead to convictions, while another source speculates only about 1 in 11 rapists is ever convicted for his crime.3 From an analysis of 159 rape convictions in the Durban regional courts in 1983, it was found that on average, sentences for coloured rapists tended to be more severe than those meted out to African and Asian rapists, while white rapists received the least severe sentences. Unsurprisingly, it was also found that sentences were most severe when the victim was white.4 Another study of 118 rape convictions returned by the Durban and Coast Local Division of the South African Supreme court during 1970 - 1979 found that the average length of imprisonment was just over 7 years, while the median was 5. The death sentence was handed down on two occasions.5 Methodology This study relies upon quantitative methods of data collection and analysis and, as such, it cannot provide complete explanations for particular findings. Nonetheless, quantitative data is very useful for tracking trends, estimating the magnitude and extent of problems, as well as pointing to areas where further qualitative research is needed. Rape and Assault Perhaps the crimes most comparable to rape are those included within the various categories of assault. Assaults, like rape, are not property crimes (although they may accompany such crimes)but are directed primarily at the body with the aim of causing hurt and injury. Similarly, rape and assaults often occur between people known to each other. A legal text book6 defines these crimes as follows: Assault consists in unlawfully and intentionally: (1) applying force to the person of another, or (2)inspiring a belief in that other that force is immediately to be applied to him. Assaults are then further categorised as 'common', 'indecent' or 'with intent to cause grievous bodily harm'. Rape has been defined as intentional, unlawful sexual intercourse with a woman without her consent. To expand on this definition, 'sexual intercourse' is described as partial or full penetration of the vagina by the penis. Any other penetrative act such as sodomy, oral penetration, and penetration with foreign objects, is termed an 'indecent assault'. In general, common assaults and indecent assaults are considered less serious than rape and assault with intent to cause grievous bodily harm; sentences for the former are typically lighter than those of the latter. Figures for rape were thus compared with figures for indecent assault, common assault and assault with intent to cause grievous bodily harm (GBH). Sources of Information This study makes use of data drawn from two sources: the SAPS and the Central Statistical Service (CSS). The CSS, in turn, compiles its reports based on information received from the Departments of Justice and Correctional Services. The period between 1984/85 - 1993/94 was selected for comparison between prosecution and conviction rates. When reporting rates were also compared, the eight-year period between 1986 - 1994 was adopted as figures for the years preceding 1986 could not be obtained from the SAPS. Figures for indecent assault and common assault prior to 1990 could also not be obtained from the SAPS so their comparison is restricted to a three-year period only. Problems Encountered with the Use of this Information The statistics supplied by both the SAPS and the CSS for thi>study do not include figures from the former TBVC states and self-governing territories. CSS statistics are also not adjusted to include the outcome of appeals. In other words, it is entirely possible that the number of convictions recorded will be even lower once the outcomes of appeals are included. There is also the fact of competent verdicts to be considered. If the initial charge of rape cannot be proved in court, it may be reduced to a lesser charge - such as indecent assault - which can be proved. This is termed a competent verdict. There is no information available on how many cases initially charged as rape were subsequently reduced to lesser charges. Data collection periods for these two sources also do not coincide. The CSS compiles its reports based on the 12-month period between 1 July to 30 June of the following year. The police, on the other hand compile their reports based on the 12-month period between January and December of each year. These two different sets of figures coincide for six month only. However, this lack of overlap does not necessarily invalidate the data completely. In relation to this point, consider the amount of time that elapses between reporting and concluding a case. Once the case is reported, a docket is opened and an investigation begun - which takes time. After the investigation is completed, the docket is sent to the Attorney-General for a decision on whether to prosecute or not. If it is decided that there is sufficient evidence to warrant a prosecution, the matter is set down for trial in either the Regional or Supreme Courts. Trials are interrupted by various postponements (witnesses are not available, courts are not available, the defence lawyer requires more time or is stricken by mysterious illness). As this description suggests, it is not unusual for rape cases to take well over a year to complete. Given this time lag, projecting prosecution and conviction rates six months into the following year does not seem unreasonable. The 1986 SAPS year will be compared with CSS data for 1986/87, for example. Limitations of Findings The difficulties described above highlight some of the risks of using information derived from other sources - one has no idea of the accuracy of the original information. Any inaccuracies or defects contained in the primary information are obviously carried over into subsequent uses of the information. The findings in this study are only as accurate as SAPS and CSS records permit them to be. Findings Discussion of Table 1 Table 1 illustrates how prosecution and conviction rates generally have dropped steadily over the last eight years. Some of the explanations for this trend may be fairly straightforward. Prosecutors' posts have been frozen since the late 1980's; people leaving the courts are generally not replaced. As a result, a diminishing pool of people is expected to prosecute a greater number of crimes than ever before. There is also the misallocation of prosecutors to be considered. A prosecutor based at one of the regional courts provided the following example of this practice. A pool of prosecutors has been retained to work exclusively on fraud cases in the courts where she is based. Despite there having been a significant reduction in the number of fraud cases heard in the courts over the last few years, additional prosecutors have recently been assigned to this group. No more prosecutors have been made available to work on rape cases in spite of the very significant increases in this type of crime.7 The SAPS suffers similar difficulties. A hold has been placed on recruitment, which has been simultaneously accompanied by an exodus from the SAPS. The time put into investigations as well as the quality of investigations are jeopardised by inexperience and sheer overload. Poor investigations lead to poor evidence-gathering which weakens the State's case even further. The assertion that rape has one of the lowest conviction rates of all serious crimes seems well-founded, for this general drop is most marked in rape cases. Table 1: Comparison of reporting, prosecution and conviction rates for the period 1986 - 1994 (prosecution and conviction rates expressed as a percentage of the total reported)

Year	Common assault			Assault: GBH			Indecent assault			Rape
	Reports	Prosecutions	Convictions	Reports	Prosecutions	Convictions	Reports	Prosecutions	Convictions	Reports	Prosecutions	Convictions
1986				109 755	66 448(61%)	46 531(42%)				15 816	9 639 (61%)	4 984 (32%)
1987				120 779	69 560(58%)	48 690(40%)				18 145	10 424 (57%)	5 233(29%)
1988				125 571	70 429(56%)	50 188(40%)				19 368	10 104 (52%)	5 160 (27%)
1989				128 887	64 709(50%)	45 897(36%)				20 458	9 873 (48%)	4 991(24%)
1990	130 291	43 073(33%)	36 381(28%)	123 639	62 464(51%)	43 461(35%)	2 304	1 186(51%)	868 (38%)	20 281	9 057 (45%)	4 661 (23%)
1991	141 196	41 509(29%)	34 707(25%)	129 241	62 490(48%)	42 222(33%)	2 639	1 233(47%)	873 (33%)	22 749	9 450 (42%)	4 841 (21%)
1992	143 224	36 746(26%)	30 928(22%)	135 991	60 368(44%)	43 291(32%)	2 812	1 152(41%)	882 (31%)	24 316	8 998 (37%)	4 753 (20%)
1993	148 855	35 572(24%)	29 185(20%)	144 475	54 739(38%)	38 263(26%)	3 127	1 058(34%)	753 (24%)	27 039	8 553(32%)	4 311 (16%)

Table 2: Comparison of Prosecution and Conviction rates for the period 1984/85 - 1993/94

Year	Common Assault			Assault: GBH			Indecent Assault			Rape
	Prosecutions	Convictions	%	Prosecutions	Convictions	%	Prosecutions	Convictions	%	Prosecutions	Convictions	%
84/85	56 397	48 053	85%	74 849	52 253	70%	597	441	74%	9 816	5 038	51%
85/86	51 006	43 657	86%	67 315	47 290	70%	587	464	80%	8 956	4 593	51%
86/87	47 991	40 918	85%	66 448	46 531	70%	634	472	74%	9 639	4 984	52%
87/88	48 715	41 422	85%	69 560	48 690	70%	729	523	72%	10 424	5 233	50%
88/89	48 167	41 325	86%	70 429	50 188	71%	908	654	72%	10 104	5 160	51%
89/90	43 787	37 438	86%	64 709	45 897	71%	1 135	820	72%	9 873	4 991	51%
90/91	43 073	36 381	84%	62 464	43 461	70%	1 186	868	73%	9 057	4 661	51%
91/92	41 509	34 707	84%	62 490	42 222	68%	1 233	873	71%	9 450	4 841	51%
92/93	36 746	30 928	84%	60 368	43 291	72%	1 152	882	77%	8 998	4 753	53%
93/94	35 572	29 185	82%	54 739	38 263	70%	1 058	753	71%	8 553	4 311	50%