C rime statistics the world over are a source of much debate. In South Africa the issue receives special attention because of the prevalence of violent crime and the resulting need to monitor and explain the trends. As expected, the release of the South African Police Service’s (SAPS) 2003/04 annual report – containing the most recent crime statistics – roused significant public and media attention. While it contained mostly good news, much still needs to be done to ensure a continuation in the downward trends. Another positive sign is that South Africans today have access to more information about the country’s crime situation than ever before. Coinciding with the release of the most recent annual report was the posting of the long-awaited police station level crime statistics on the SAPS website.1 Previously, the data was available at national, provincial and area level only. This is encouraging because crime is, after all, best understood at local level. This article provides an overview of the latest crime statistics, highlighting some of the most important changes in the trends both nationally and provincially.2 Interpreting crime statistics When analysing official crime statistics it must be noted that real crime levels will be undercounted, because offences that are not reported to police are not reflected in the data. For a crime to appear on the official police crime records, two things need to happen: victims or witnesses must report it to the police,3 and the police must actually record the crime in their database.4 Therefore, official crime statistics will never reflect the true crime rate. Speculation abounds as to the actual levels of underreporting in South Africa. In truth, we will probably never know for sure. And while it is safe to assume that certain crimes are better recorded than others,5 real levels of interpersonal violence, child abuse and sexual offences will remain unknown. It is also important to bear in mind that the definitions of what constitutes a crime are constantly shifting. New categories of crime are created, current definitions are tweaked slightly, and certain activities become decriminalised. This changing legal framework and definitional reengineering can impact on the reliability of statistical comparisons of crime over time, whether based on police records or SA CRIME QUARTERLY No 12 JUNE 2005 1 Anton du Plessis and Antoinette Louw Institute for Security Studies anton@iss.org.za ant@iss.org.za THE TIDE IS TURNING The 2003/04 SAPS crime statistics The latest official crime statistics, for the financial year 2003/04, are encouraging because they show a decrease in levels of most serious crimes, including murder, car hijacking, burglary and farm attacks. But the good news has yet to hit home for many South Africans. In fact, most people are more afraid of crime today than they were in 1998. It is likely that only sustained decreases in violent crime over several years, coupled with better service delivery, will make a difference. the other key statistical source on crime, namely victimisation surveys. Despite these caveats, South Africa’s official crime statistics are widely regarded as comprehensive and certainly as the most detailed and reliable on the continent. National trends since 1994 Despite a brief dip in levels of recorded crime per capita between 1995 and 1997, overall crime rates rose steadily until 2001. Since then, levels have remained fairly constant, and have actually decreased in 2003/04 (Figure 1). While this is good news, it must be remembered that violent crime rates are nevertheless extraordinarily high: during the 12-month period of 2003/04, approximately 20,000 murders, 53,000 rapes, 107,000 aggravated robberies and 266,000 serious assaults were recorded by the police. Since 1994 violent crime, as a proportion of all crime, increased slightly from 31% to 33%. That is, during 2003/04 one in three crimes recorded involved violence or the threat of violence. All other crime categories declined marginally as a proportion of the overall number. Violent crime trends As highlighted above, violent crime is South Africa’s main problem. The SAPS 2003/04 annual report classifies the following offences as violent crime (or SA CRIME QUARTERLY No 12 JUNE 20052 DU PLESSIS AND LOUW what they call “contact crimes”): murder, attempted murder, rape, assault with the intent to cause grievous bodily harm (GBH), common assault, robbery with aggravating circumstances, and other robbery. As can be seen in Figure 2, most categories of violent crime, with the exception of aggravated robbery, have decreased – in some cases substantially – from the previous year’s figures. Murder Murder has decreased steadily over the past 10 years, from 67 murders per 100,000 of the population in 1994/95 to a rate of 43 per 100,000 in Source: SAPS Annual Report 2003/04 4,000 4,400 4,800 5,200 5,600 6,000 94/95 95/96 96/97 97/98 98/99 99/00 00/01 01/02 02/03 03/04 Figure 2: Change in recorded violent crime rates, 2002/03-2003/04 Source: SAPS Annual Report 2003/04 -20 -15 -10 -5 0 5 Attempted murder % -18 Murder -10 Other robbery -8 Assault GBH -4 Common assault -3 Rape -1 Aggravated robbery 3 N o . p er 1 0 0 ,0 0 0 p eo p le Figure 1: Total recorded crime rate in SA, 1994/95-2003/04 2003/04 (Figure 3). The drop between 2002/03 and 2003/04 represents a decrease of 10%. This is reassuring because murder is the one violent crime that does not suffer from under-reporting, and is thus the most reliable indicator of the real violent crime situation. However, at 43 murders per 100,000 people in 2003/04, or 20,000 per year, the situation is still dire. But the decrease since 1994 has meant that South Africa has fallen far behind the main contenders for the unsavoury title of ‘murder capital of the world’. According to the most recent Interpol statistics,6 the average murder rate for the 111 countries referred to on their website is eight per 100,000. Colombia and Jamaica now lead the pack with murder rates of about 78 and 59 per 100,000 respectively. Attempted murder While murder continues to decrease, attempted murder has moved in the opposite direction, and has, until 2003/04, increased (Figure 3). (The rate of 69 per 100,000 people in 1994/95 increased to 79 per 100,000 in 2002/03.) It would make sense if the trends for murder and attempted murder tracked each other closely. The fact that they don’t suggests either that murderers are becoming less effective, or that recording peculiarities are to blame. The former is unlikely, which suggests that this counter-intuitive trend is the SA CRIME QUARTERLY No 12 JUNE 2005 3DU PLESSIS AND LOUW result of more people reporting attempted murders, rather than an increase in the real incidence. However, 2003/04 saw an end to this trend, with attempted murders showing a substantial drop of 18% (to a rate of 65 per 100,000) from the previous year. This sudden decrease is difficult to explain, but could be linked to a change in the classification of attempted murders in the official database. Robbery By far the most troubling national trend is the increase in levels of aggravated robbery. Between 1994/95 and 2003/04 the aggravated robbery rate increased by 31%. After an increase of 3% in the most recent financial year, the rate per 100,000 had reached 288. According to the SAPS statistics, aggravated robbery includes incidents ranging from street level muggings at knifepoint, to car hijackings and major cash-in- transit (CIT) heists involving millions of rands. Fortunately, separate statistics are provided for hijacking (vehicular and truck) and CIT heists (see discussion below), even though these crimes are still included in the aggravated robbery category. Leaving aside these special categories of aggravated robbery, it is concerning that the overall rate of aggravated robbery continues to increase. Although this is a crime that is well known for being under- reported (only about 29% of victims nationally report robbery to the police),7 the increase is too Source: SAPS Annual Report 2003/04 Figure 3: Trends for murder and attempted murder in SA 0 10 20 30 40 50 60 70 80 90 100 94/95 95/96 96/97 97/98 98/99 99/00 00/01 01/02 02/03 03/04 N o . p er 1 0 0 ,0 0 0 p eo p le Murder Attempted murder SA CRIME QUARTERLY No 12 JUNE 20054 DU PLESSIS AND LOUW drastic to be attributed simply to higher rates of reporting. Aggravated robbery seems to have genuinely worsened. A possible explanation is the increase in legitimate as well as fraudulent reported cell phone robberies. Stolen or lost cell phone sim cards can only be replaced if owners present a police case number. Cell phone providers estimate that there are over 20 million cell phone users in South Africa. A conservative estimate might be that one in every 30 cell phones are lost or stolen each year, which translates into between 600,000 and 700,000 potential sim cards that need to be replaced. When figures of this magnitude are factored into the equation, it becomes plausible that the increase in aggravated robbery is linked to the proliferation of cell phone related robberies. Rape and assault Trends in other categories of violent crime are more difficult to explain. Unlike murder, which is highly reported and detected by police, crimes like rape, child abuse, assault GBH and common assault are seldom reported and are thus undercounted in the official records. It is therefore difficult to interpret the latest statistics that show decreases of 1% in recorded rapes, 4% for assault GBH, and 3% for common assault between 2002/03 and 2003/04 (Figure 2). Nevertheless, as in the case of murder, what can be said is that the rates for these violent crimes remain high: in 2003/04 there were 114 recorded rapes per 100,000 people in the country, 561 assaults GBH, and 606 common assaults. There has been a lot of speculation about the number of rapes that go unreported. Without engaging in what amounts to a fruitless debate, the most that can be said about these trends is that increases in recorded levels (rather than decreases as is the case here) would be encouraging in so far as they show that more victims are turning to the state for protection. In this regard, cognisance should also be taken of the impact that various pieces of legislation and improved service delivery can have on recorded levels of rape and assault. The implementation of the Domestic Violence Act,8 which came into force on the 15th of December 1999, probably increased reporting and detection of assault cases. The finalisation of the new Sexual Offences Bill, which proposes a broader definition of rape that includes male victims of rape, will also result in recorded cases going up. The establishment of sexual offences courts and rape care centres will have the same effect. Hijacking, bank robbery and cash-in-transit heists One of the crimes that dominates public concern with violent crime is car hijacking. Given this, the SAPS has, since 1996, reflected hijacking cases separately from other aggravated robberies. The same goes for cash-in-transit (CIT) heists and bank robberies. The latest statistics provide good news: hijacking (cars and trucks) is down 6% in 2003/04 from 2002/3 levels (Table 1). Better still, these figures are likely to reflect the true picture because hijackings are among the crimes that are highly reported by the public. This is largely due to the fact that cars and trucks, as valuable items, tend to be insured. The decrease in hijacking could well be the result of targeted enforcement efforts. These robberies, which are often linked to organised crime syndicates, are one type of crime that the SAPS is well equipped to deal with on a national level. Anti-hijacking task teams and dedicated hijacking courts appear to have been effective in curbing the incidence of this serious crime. CIT heists and bank robberies are also worth mentioning. Although CIT robberies fell by 49% in 2003/04, it must be remembered that this crime increased by as much as 57% the year before. The high profile nature of these heists – which often involve heavily armed attacks carried out in public with military precision – attracts significant local and international media attention, and should be monitored closely. The figures for bank robbery also show significant decreases. In 2003/04 the downward trend for this crime continued, with a drop of 58% from the previous year. The substantial investments made by the banking sector in reducing robberies, and the industry’s efforts to work closely with the police, have no doubt contributed to these declines. -8 Table 1: Number of ‘special category’ robberies in SA, 1996/97-2003/04 1996/97 1997/98 1998/99 1999/00 2000/01 2001/02 2002/03 2003/04 Carjacking 12,912 13,052 15,773 15,172 14,930 15,846 14,691 13,793 Truck hijacking 3,732 4,657 6,134 5,088 4,548 3,333 986 901 CIT robbery 359 236 223 226 196 238 374 192 Bank robbery 561 463 493 450 469 356 127 54 Source: www.saps.gov.za SA CRIME QUARTERLY No 12 JUNE 2005 5 Property crime Whereas violent crime makes up approximately 30% of all recorded crime, property crimes account for the majority (about 56%) of all offences. The most serious property crimes recorded by the SAPS all decreased between 2002/03 and 2003/04 (Figure 4). Residential burglary The decrease in rates of residential housebreaking in the past year is significant because burglary had been increasing fairly steadily since 1994. A rate of 596 housebreakings per 100,000 people was recorded in 1994/95; this increased to 704 in 2002/03, and then dropped by 8% to 645 per 100,000 in 2003/04. Whether these increases were in fact ‘real’ or the result of more victims reporting the crime to the police is debatable. The latest estimates are that only 57% of victims in South Africa report their case of housebreaking to the police.9 Burglary is not therefore one of the most reported crime types, which suggests that these trends could well say more about reporting patterns, than actual burglary rates. Among the reasons for the relatively low rates of reporting are that housebreaking has a very low detection rate, and as a result many victims simply don’t bother to notify the police. Insurance no doubt also plays a role. Most people cannot afford to insure their household property, and thus the incentive to report created by the need to make an insurance claim, would not apply in many cases of burglary. Car theft The continued decrease in car theft (this crime has declined steadily since 1998/99) is another positive sign. While murder is the most accurate indicator of violent crime, recorded car theft is a reliable indicator in the property crime category: estimates are that 97% of victims notify the police when their vehicle is stolen.10 Once again, the need to make an insurance claim no doubt explains the high reporting rates. The trend in the SAPS statistics is furthermore supported by recent data released by South Africa’s major motor insurance and security companies that show a sharp decrease in car theft related insurance claims.11 Stock theft The importance of stock theft in South Africa is sometimes underestimated. A recent national victim survey found stock theft to be the fifth most common crime in the country, with 2.5% of respondents surveyed saying they had experienced this crime.12 The survey also found that most victims were poor – 68% earned less than R1,000 a month – which DU PLESSIS AND LOUW Figure 4: Change in property crime rates, 2002/03-2003/04 Source: SAPS Annual Report 2003/04 -15 -10 -5 0 Housebreaking (business) % Theft out of vehicles Stock theft -14 -14 Housebreaking (residential) -13 Theft of vehicles-7 means they were potentially robbed of a significant portion of their annual income. The impact of stock theft on both the subsistence and commercial farming communities is thus likely to be severe. However, using police statistics to track this crime type is difficult, because only 36% of victims say they report the theft to the authorities.13 The low reporting rates were found to be related to accessibility of the police in rural areas, use of traditional authorities to resolve the matter, and the fact that positive outcomes for reporting are rare. It is therefore difficult to interpret the figures in the official database which show that stock theft decreased by 13% between 2002/03 and 2003/04 (Figure 4). ‘Victimless’ crimes Certain types of crimes are known as ‘victimless’ crimes because there is no victim as such who will report the crime. These offences are also known by the SAPS as crimes that are “heavily dependent on police action for detection” and thus for recording in the database.14 Typical examples are possession of firearms, drug related crimes, and driving under the influence of alcohol. Because the recording of these offences is highly dependent on proactive police action, like search-and-seizure operations and roadblocks, increases in the statistics can be interpreted as a positive indicator of enforcement efforts. The SAPS figures show that cases of illegal possession of firearms increased by 4%, drug related crime increased by 14%, and driving under the influence of drugs or alcohol increased by 10%. The police deserve recognition for the increases in recorded levels of ‘victimless crimes’ particularly because these offences tend to be linked to various other forms of criminality (see the article on the 2005 gun amnesty in this issue, for example). Provincial crime trends Recorded crime rates vary greatly by province. In terms of trends over time, the 2003/04 statistics show some surprising results for certain crime categories. This article outlines the provincial patterns for three serious crime types only: murder, aggravated robbery, and farm attacks. Murder The provinces with the highest murder rates in 2003/04 – and whose rates were higher than the national average of 43 per 100,000 were: Western Cape (60), Eastern Cape (53), KwaZulu-Natal (53), Northern Cape (50), and Gauteng (45). Limpopo’s rate was lowest, with 13 murders per 100,000. In terms of trends over time, murder rates decreased in eight of the country’s nine provinces between 2002/03–2003/04 (Figure 6). While most provincial decreases were not out of kilter with the national average, two provinces stand out because their murder rates decreased by more than that of the country as a whole. In Gauteng, the murder rate dropped by 16% and in Western Cape by 25%. The latter trend is fairly surprising: not only is the decrease substantial for a single 12-month period, but the province has also had more year-on-year increases in murder than any other since 1994. Indeed, while most provinces recorded declines of between 10%–32% between 1994/95 and 2003/04, Western Cape’s murder rate increased by 4% over the 10 year period. (Gauteng’s murder rate, by comparison, dropped by 28% in the past decade.) The decrease in the Western Cape is nevertheless good news which would be reinforced by research that assesses to what extent the decline is linked to the high density policing operations in the Cape Town metropolitan area, to a reduction in or better control of firearms or alcohol consumption, or changes to patterns of social behaviour or the structures that influence it. Aggravated robbery Gauteng stands out as the province with by far the highest rate of aggravated robbery in the country, with 649 robberies per 100,000 people in 2003/04. The only other province with a rate higher than the national average of 288 per 100,000 was the Western Cape with 293 robberies per 100,000 people. KwaZulu-Natal was in third place with a rate of 284 – only just below the national figure. The lowest aggravated robbery rate was recorded in Limpopo (62). People living in Gauteng are therefore 10 times more likely to be a victim of aggravated robbery than those in Limpopo. SA CRIME QUARTERLY No 12 JUNE 20056 DU PLESSIS AND LOUW As mentioned above, the national increase in aggravated robbery remains a key concern. Rates also increased in seven of the nine provinces, with decreases recorded in only the Western Cape and Limpopo (Figure 6). Violence on farms and smallholdings Attacks against people living on farms and smallholdings have attracted significant attention since the police first started recording them as separate incidents in 1997. The senseless and often brutal nature of some of these attacks (most of which are committed during robberies)15 sparked widespread public debate that eventually led to a commission of inquiry into farm attacks.16 In 2003/04, the majority of farm attacks (‘farm’ attacks here always refers to both farms and smallholdings) were recorded in Mpumalanga, Gauteng and KwaZulu-Natal (Figure 7). Northern Cape and Western Cape were the least affected provinces. After increasing significantly between 1997/98 and 2001/02, farm attacks have since declined (Figure 8). Between 2002/03 and 2003/04, incidents dropped by 14%, and murders in the course of these attacks decreased by 15%. While this is certainly good news, the trend does not apply to all provinces: attacks increased in KwaZulu-Natal, Eastern Cape and Free State between 2002/03 and 2003/04, while murders as a result of these incidents increased in Gauteng during the same period. Conclusion Overall, the statistics indicate that most categories of crime are on the decrease. Although the reasons for this trend have not been established, possible explanations point to increased police numbers, and visible police operations that have targeted key crime hotspots, accompanied by operations focused on syndicated criminal activity. It is difficult to find fault with such responses if indeed they have contributed to the decrease in crime rates – decreases which in some cases have been substantial. However, strategies relying solely on enforcement are costly to sustain, and are unlikely to fix the underlying causes of crime – and particularly violent crime. This typically requires social and developmental interventions, which are supported by targeted police action. The good news implicit in this analysis of the latest crime figures has yet to hit home for many South Africans. In fact, most people in the country are more afraid of crime today than they were in SA CRIME QUARTERLY No 12 JUNE 2005 7DU PLESSIS AND LOUW Figure 5: Change in provincial murder rates, 2002/03-2003/04 Figure 6: Change in provincial aggravated robbery rates, 2002/03-2003/04 Source: SAPS Annual Report 2003/04 -30 -25 -20 -15 -10 -5 0 5 Western Cape % -25 Gauteng-16 SA-10 Free State-7 KwaZulu-Natal-6 North West-6 Northern Cape-5 Mpumalanga-3 Limpopo-1 Eastern Cape1 Source: SAPS Annual Report 2003/04 -20 0 20 40 60 80 100 Western Cape % -6 Limpopo-1 Gauteng1 2 3 6 7 10 32 85 KwaZulu-Natal SA Eastern Cape Mpumalanga North West Free State Northern Cape SA CRIME QUARTERLY No 12 JUNE 20058 DU PLESSIS AND LOUW 1998.17 It may well take several years of continued improvements in the crime situation – and some good public relations campaigns by government – to convince South Africans, both locally and abroad, that things are actually improving. Endnotes 1 The SAPS website is 2 Figures in this article come from the SAPS annual report, pp 34–41 which can be accessed at . Detailed statistical tables can also be drawn from the website, although readers should note that figures showing the percentage change from one year to the next differ between the website tables and the annual report. This is because the percentages in the annual report are calculated using the rates per 100,000 people in each year, while those on the website are based on the raw number of crimes recorded in each year. 3 Certain crimes are also discovered by the police without being reported by a member of the public, such as murder, and drug and firearm related crimes. 4 For the crime to be reflected in the SAPS record, the police must first open a case docket. 5 Murder, car theft, car hijacking, and burglary are usually well reported. 6 Interpol website These statistics are only available up to 1998. 7 P Burton, A du Plessis, T Leggett, A Louw, D Mistry and H van Vuuren, National Victims of Crime Survey: South Africa 2003, ISS Monograph No 101, Institute for Security Studies, Pretoria, July 2004, p 107. 8 116 of 1998. 9 P Burton et al, op cit, p 107. 10 Ibid. 11 See 12/0/05; 12 P Burton et al, op cit, p 138. 13 Ibid. 14 SAPS Annual Report, 2003/04, p 34. 15 Analyses conducted by the SAPS Crime Information Analysis Centre found that more than 90% of all farm attacks were linked to robbery, see SAPS Annual Report 2003/04. 16 Report of the Committee of Inquiry into Farm Attacks, 31 July 2003, p 44, 17 Burton et al, op cit, p 50. Figure 7: Distribution of attacks on farms and smallholdings by province, 2003/04, n=773 Figure 8: Trend for attacks and murders on farms and smallholdings, 1997/98-2003/04 0 1,200 800 400 1,000 600 200 97/98 N u m b er 490 87 98/99 827 153 99/00 823 148 00/01 908 140 01/02 1,069 140 02/03 903 103 03/04 773 88 Incidents Murders Source: SAPS Annual Report 2003/04 Source: SAPS Annual Report 2003/04 Mpumalanga 26% Gauteng 24%KZN 15% North West 9% Free State 8% Limpopo 6% Eastern Cape 6% Western Cape 5% Northern Cape 1%