ISS Africa
Home / 01 Dec 2008: ISS Paper 177: The Dynamics of Crime, Comparing the Results from the 1998, 2003 and 2007 National Crime and Victimisation Surveys, Robyn
01 Dec 2008: ISS Paper 177: The Dynamics of Crime, Comparing the Results from the 1998, 2003 and 2007 National Crime and Victimisation Surveys, Robyn
1 December 2008

The Dynamics of Crime
Comparing the Results from the 1998, 2003 and 2007
National Crime and Victimisation Surveys


Robyn Pharoah

Occasional Paper 177 , December 2008





Victim surveys present a valuable opportunity to examine the dynamics surrounding different crime types, such as where and when crimes occur, the use of weapons, the extent and nature of violence and other factors. This paper, one of a series of papers, discusses the dynamics surrounding key crimes identified in the National Crime and Victimisation Survey conducted by the ISS during October and November 2007. As discussed in the paper titled ‘2007 National victims of crime survey: overview of key findings` (ISS Paper 175), this survey followed on from two earlier surveys, one conducted by Statistics South Africa in 1998 and another conducted by the ISS in 2003, which sought to gather information on South Africans` perceptions and experiences of crime and the criminal justice system.

The paper presents the key findings for the six most common crimes identified in the survey: housebreaking, theft of personal property, theft out of vehicles, robbery, assault and corruption. Wherever possible, it compares the 2007 findings to those of the 1998 and 2003 surveys, but the 2003 and 2007 surveys were far more comprehensive than the earlier survey, limiting the bulk of the analysis to the 2003 and 2007 surveys. These surveys not only asked the same questions, but adopted comparable methodologies and were weighted according to the same criteria.

The paper begins by briefly discussing the most commonly reported crimes, before going on to examine each crime in detail. It focuses on the site, timing and other features of each crime, rather than the demographic or socio-economic features influencing victimisation.

The most common and serious crimes in South Africa

The most common types of crime have remained fairly stable across all three victimisation surveys. As shown in Table 1, housebreaking was by far the most frequently experienced crime in 1998, 2003 and 2007. Property theft, corruption, robbery and theft out of vehicles were, in order, the next most common crimes recorded in 2007. Corruption was the second most recorded crime in 2003, while levels of theft out of vehicles ranked slightly higher than in 2007. Property theft was the second most frequently reported crime in 1998, followed by assault, theft out of vehicles and robbery. Although it ranked fifth in 2003, assault was not amongst the top five crimes in 2007, but is included in the analysis for comparative purposes and because it is considered one of the more serious crimes.



Definition: Housebreaking occurs when someone breaks into a dwelling without permission and steals or attempts to steal something.


The incidence of housebreakings in each province has remained fairly stable across the 1998, 2003 and 2007 surveys (Table 2). Gauteng and KwaZulu-Natal consis­tently show the highest levels of housebreaking, while the Northern Cape shows the lowest. Those living in Gauteng reported the largest number of housebreakings in 1998 and the second highest in 2003 and 2007. KwaZulu-Natal saw the highest number of housebreakings in 2003 and 2007, with a little less than one in three housebreakings occurring in the province on both 2003 and 2007. Only 2 per cent of housebreakings occurred in the Northern Cape in 1998, dropping to 1 per cent in 2003 and 2007.

As shown in Table 2, KwaZulu-Natal shows the largest increase in the number of housebreakings reported since 1998, with the number of respondents experiencing burglaries increasing by 9 percentage points. The Eastern Cape, on the other hand, shows the biggest declines, with the proportion of people experiencing a burglary in the province dropping from 15 per cent in 1998 to 10 per cent in 2007.


The timing of housebreakings appears to have changed between surveys. More than two out of every three (69 per cent) of the burglaries reported in 2003 occurred at night, with the remaining break-ins occurring during the day. Those reported in the 2007 occurred almost exclusively at night, with only one in ten (9 per cent) said to have occurred during the day.

Burglaries increasingly occur when people are at home. The number of people reporting being at home when the housebreaking occurred increased over the last decade, and particularly since 2003 (Figure 1). While 44 per cent of respondents reported they were present in 1998, 48 per cent were at home in 2003, and virtually all were home in 2007. This is concerning, as increased interaction between burglars and victims creates the potential for violence.

The greater potential for interaction does not appear to have coincided with increased use of weapons by perpetrators – in fact weapon-use may be declining. As shown in Figure 1, comparisons of the data show that slightly fewer than one third (31 per cent) of respondents reported that the perpetrators had weapons in 1998. This dropped to one quarter (26 per cent) in 2003. The 2007 questionnaire did not ask victims of burglary present at the time whether weapons were used, but one fifth of those reporting a housebreaking also reported experiencing a robbery at home where weapons were used.


Figure 1: Weapon-use in burglaries where victims were at home

Police statistics on house robbery


These findings are interesting given the increasing number of robberies in residences – house robberies –  reported to the police since these statistics were first collected in 2002/03 (Box 1). While the victim survey data supports the notion that victims are encountering burglars more than ever before, and that these encounters may be associated with force and violence, it suggests that this is not necessarily linked to more weapon-use.

Most housebreakings are successful, although fewer respondents reported that they had had items stolen from their home in 2007 than in previous years. While two thirds (66 per cent) of those that had been burgled reported that property was stolen in 2007, well over four fifths (87 per cent) had actually lost items in 2003 and 1998. This shift may be explained by the trend towards burglaries when people are home, which is likely to increase the chance of perpetrators being interrupted before they can take items. Perpetrators were most likely to take electronic equipment in 2003 (59 per cent) and 2007 (55 per cent), followed by personal items (26 per cent) in 2003 and clothing and money (25 per cent) in 2007.

In many instances respondents knew who broke into their home. Just over one third (35 per cent) reported knowing who the burglars were in 2007. This is more than in 2003, when one quarter (27 per cent) knew who was responsible. How respondents knew, however, remains similar across the two surveys. In most cases, respondents (51 per cent in both 2003 and 2007) recognised the perpetrators or knew them by name. Less often, they were known to the community (22 per cent) or by others who witnessed the theft (23 per cent in 2003 and 16 per cent in 2007).

Most housebreakings were reported to the police (59 per cent, 57 per cent and 60 per cent, in 1998, 2003 and 2007 respectively). Such cases were more likely to result in arrests in 2007 than in 2003. While one in ten (12 per cent) respondents reported that arrests had been in made in their case in 1998 and 2003, one in five reported that there had been arrests (20 per cent) in 2007.

Property theft


Definition: Theft of property consists of the unlawful and intentional appropriation of another`s property. This includes bag-snatching and pick-pocketing.

The provincial distribution of property theft followed the pattern for housebreaking. As with housebreaking, levels of property theft were highest in Gauteng and KwaZulu-Natal, with Gauteng leading in 1998 and 2007 and KwaZulu-Natal leading in 2003. The Northern Cape again recorded theft levels of less than 2 percent in all three surveys (Table 3).


Gauteng saw the largest increases in theft levels between 1998 and 2007, where the proportion of people reporting thefts almost doubled from 24 to 42 per cent. It also saw the largest increase between 2003 and 2007. The Eastern Cape again experienced the largest decline between 1998 and 2007, but the decrease between the 2003 and 2007 surveys was greatest in KwaZulu-Natal, where victimisation levels dropped from 30 per cent to 23 per cent.


As with housebreaking, the dynamics surrounding property theft appear to have shifted slightly between the 2003 and 2007 surveys. Respondents were most likely to have reported that property was stolen at home (34 per cent) in 2003, while in 2007 they were most likely to have had property stolen when they were outside in the street. As shown in Figure 2, while the proportion of thefts occurring at other locations dropped between 2003 and 2007, respondents were substantially more likely to have had belongings taken in the street in 2007 than in 2003.


Figure 2: Place where thefts of property occurred in 2003 and 2007

In most cases, victims knew immediately that items had been stolen. Victims were slightly more likely to report knowing immediately that items had been stolen in 2007 (83 per cent) than in 2003 (78 per cent). This is most likely because street-thefts often involve bag-snatching and pick-pocketing, which may be more immediately noticeable than the pilfering of items from the home.

Other aspects of the crime remain similar between the surveys. In most cases, victims were alone when they had items stolen (57 and 60 per cent in 2003 and 2007 respectively). They also most often lost cell phones (69 and 45 per cent in 2007 and 2003 respectively), money (43 and 37 per cent) or their purse or wallet (23 and 17 per cent).

Victims often knew who stole their things, although less so than with housebreaking. One out of every five (18 per cent) reported that the thief was known to them in 2003, while one out of every four (25 per cent) reported accordingly in 2007. The majority of respondents in both 2003 and 2007 recognised the perpetrator or knew them by name, although respondents were more likely to know the perpetrator in 2007 than in 2003 (77 per cent compared to 70 per cent).

The extent to which victims know the perpetrator varies according to where people live. Those living in urban areas were substantially more likely to know the person, with this increasing from 80 per cent in 2003 to 91 per cent in 2007. The 2007 survey made the distinction between formal and informal urban settlement areas. The findings show that levels of familiarity are much higher in formal urban areas (80 per cent) than in informal settlements (12 per cent), where populations tend to be more transitory and less settled. Respondents living in rural areas were least likely to know the perpetrator, particularly in areas outside the jurisdiction of traditional authorities (4 per cent and 3 per cent compared to 15 per cent and 6 per cent in 2003 and 2007 respectively).


Levels of reporting have dropped off dramatically since 1998. In 2007, a little more than two out of every five (44 per cent) respondents reported the theft to the police, slightly more than the 41 per cent who reported their goods stolen in 2003. The 2003 and 2007 figures are both substantially lower than those recorded in 1998, when three out of every five (59 per cent) said they reported the crime to the police.

The trend is less clear regarding the arrests resulting from reports. While the number of arrests were similar in the 1998 and 2007 surveys (6 per cent and 7 per cent respectively), they were almost double this number in 2003. More longitudinal data is needed to establish whether this pattern is due to major changes in police performance between 2003 and 2007 or whether the 2003 spike is simply due to anomalies in the data.

Theft out of vehicles

Definition: Theft from a motor vehicle consists of the unlawful and intentional removal of parts, accessories or equipment, that form part of a motor vehicle or articles in or on the vehicle from the vehicle. This includes theft of car radios or goods left in the car or parts of the car such as a car mirror or spare tyre.


Figure 3: When cases of theft out of vehicles occur, 2003 and 2007


As shown in Table 4, respondents in Gauteng reported the highest rates of theft of out vehicles in all three of the surveys, with the number of respondents reporting thefts increasing year-on-year in 1998, 2003 and 2007. This increase was spread fairly evenly over the period; the number of respondents reporting thefts from their vehicles increasing by 14 percentage points from 27 per cent to 41 per cent between 1998 and 2007, and to 6 percentage points between 2003 and 2007. KwaZulu-Natal saw the second highest levels of theft. North West province shows the largest declines in this crime type over both the nine year period between 1998 and 2007 and between 2003 and 2007, where it appears alongside Limpopo province.


The majority of thefts out of vehicles occurred when the vehicle was parked at home. The proportion of victims reporting that the car was broken into at their home increased slightly between the 2003 and 2007 surveys, from 57 per cent in 2003 to 63 per cent in 2007. The frequency with which other locations were mentioned remained fairly similar in both surveys, although those spoken to in 2003 were a little more likely to have their car broken into when it was parked outside on the street in a residential area (15 per cent compared to 9 per cent), while respondents in 2007 were more likely to have had items stolen when the car was parked at work or at the shops (12 per cent and 5 per cent respectively).

Thefts were most likely to occur at night, and to lesser extent, in the evening. In 2003, two out of every five thefts occurred at night – defined in the survey as between 10pm and 4am – rising to just under half in 2007. As shown in Figure 3, the number of thefts in the evening – between 7pm to 10pm – dropped slightly in 2007, from 24 to 18 per cent.

Figure 4: Day of the week on which theft out of vehicles occur, 2007

The majority also occurred during the week rather than over the weekend (71 per cent and 67 per cent in 20033 and 2007 respectively). The 2007 survey asked respondents on what day of the week items were stolen. As shown in Figure 4, the number of thefts peaked on Friday, when almost one out of every three cars were stolen, dropping slightly on Saturday, before declining to much lower levels as the new week begins.

The thieves were most likely to steal the car radio (66 per cent in 2003 and 49 per cent in 2007), and to a lesser extent, money, purses or wallets that may have been in the vehicle (19 per cent and 20 per cent in 2003 and 2007 respectively). The theft of cell phones decreased between 2003 and 2007, from 17 to 12 per cent. The theft of car parts other than the tyres rose over this period, from 13 to 19 per cent. These findings suggest that perpetrators often target cars for their components rather than valuables belonging to the driver or passengers – although these clearly motivate many.

Relatively few thefts appear to be ‘smash-and-grabs` with the potential to escalate into more serious, violent crimes. Only the 2007 survey asked whether the respondent was in the car at the time of the theft, but these results show that only one in ten (10 per cent) were in the vehicle. This suggests that while smash-and-grabs are a major public concern in South Africa, they are not the primary crime type.

Arrest rates are consistently low. Comparison across the three surveys shows that only 5 per cent of the cases reported in 1998 resulted in arrests, while only 4 and 6 per cent of reports led to arrests in 2003 and 2007 respectively. This may be one reason why reporting rates for this crime have declined since the 1998 survey.

Unsurprisingly, given the timing and context of the thefts, very few victims (10 per cent in 2003 and 12 per cent in 2007) knew who had taken the items. Those who did know, most often recognised the perpetrator. This was particularly so in 2007, when almost three quarters (74 per cent) knew the thieves by name or face (compared to 57 per cent in 2003).



Definition: Robbery involves forcefully taking property from a person or the threat of force, for example, pointing a knife at someone and demanding that they hand over their jewellery.


The provincial distribution of robberies in 1998, 2003 and 2007 varies somewhat from the pattern for housebreaking and theft. Levels of robbery in the 12 months preceding the survey were highest in Gauteng in all three surveys. They were second highest in KwaZulu-Natal in 1998 and 2003, but the Western Cape ranked second in 2007, with KwaZulu-Natal fourth after the Eastern Cape (Table 5). As previously, the Northern Cape saw the lowest robbery rates.

The largest change between the 1998 and 2007 surveys occurred in the Western Cape, where robbery rates increased by 14 percentage points, and in KwaZulu-Natal which saw the number of people reporting robberies decline by 12 percentage points (Table 5). The Western Cape also saw the largest change between the 2003 and 2007 surveys, where the number of those reporting being robbed more than doubled from 10 per cent to 24 percent. Gauteng saw the steepest decline, with victimisation levels dropping from 46 per cent to 31 per cent over this period.



As in 2003, most of the victims interviewed in 2007 were robbed in the street in a residential area (55 per cent and 58 per cent). They were next most likely to have been robbed at a train, bus station or taxi or bus stop (8 per cent), or in some other outdoor area (7 per cent). After the street, respondents in the 2003 survey were most likely to have been robbed at home (13 per cent) or when using public transport, but only 5 per cent of respondents reported being robbed at home in 2007. It is possible that the 2007 figures under-estimate the number of robberies occurring in people`s homes. The findings show that 6 per cent of those reporting being robbed at home also reported experiencing a housebreaking, suggesting that some respondents may have categorised robberies as housebreakings. If these housebreakings were in fact robberies, then the difference between the 2003 and 2007 results is small.

The majority of respondents were alone when they were robbed. Almost four out of every five (78 per cent) victims reported being alone when they were accosted, more than in 2003, when just over half (51 per cent) were alone.

The most frequently stolen items were the same in both 2003 and 2007, although their relative ranking changed. In 2007, respondents were most likely to have lost their cell phone (73 per cent), handbag or wallet (45 per cent) or money (43 per cent). In 2003 they were most likely to have had their handbag or wallet stolen (39 per cent), followed by cash (32 per cent) and their cell-phone (25 per cent).

Weapon-use has increased over the last decade. As shown in Figure 5, a little less than three out of every five (60 per cent) victims reported that the people who had robbed them had used a weapon in 1998. This increased to more than four out of every five (85 per cent) in 2003, with virtually all of those surveyed in 2007 reporting that the perpetrators had weapons. This off-sets the apparent trend towards fewer weapons reportedly being used in burglaries.


Figure 5: Weapon-use and injuries sustained during robberies in the 1998, 2003 and 2007 surveys


According to the 2007 and 1998 surveys, perpetrators were most often armed with knives or other sharp objects (53 and 50 per cent respectively) and to a lesser extent firearms (45 and 43 per cent), while the majority (74 per cent) used guns in the 2003 study. It is not clear whether these changes are the product of changing criminal approaches or anomalies in the data. Further data will again be required to assess trends over time.

Increased weapon-use has not led to a clear increase in the number of victims sustaining injuries, or implied violence. As Figure 5 shows, while the proportion of those reporting injuries increased between the 2003 and 2007 surveys from 14 to 20 per cent, the number of victims hurt during robberies in both these surveys was lower than in 1998, when one in three reported injuries. Slight differences in the phrasing of the question in the 2007 survey may also falsely increase the extent of the change between the 2003 and 2007. While the 1998 and 2003 studies asked only whether the respondent was hurt, the 2007 asked whether the respondent or anyone their household was injured, which may have led to a higher number of injuries being reported.

The findings on hospitalisation support the suggestion that robberies may be becoming less violent. The findings show that just over half (56 per cent) of those injured during a robbery in the year before the 1998 survey needed to be hospitalised. This dropped to slightly less than half (49 per cent) in 2003 and to a little more than two fifths (42 per cent) in 2007.

As with the other crimes, many knew the robbers. Respondents were slightly more likely to know who robbed them in 2007 (22 per cent) than in 2003 (19 per cent). In 2003, the vast majority knew the perpetrators by name or face (90 per cent). Respondents in 2007 were also most likely to recognise the perpetrator (49 per cent), but in many instances it was other witnesses who knew them (28 per cent).

The number of respondents reporting robbery to the police increased between both the 1998 and 2007 surveys and the 2003 and 2007 surveys. Two out of every five (41 per cent) reported the robbery to the police in 1998. This dropped to almost one in three (29 per cent) in 2003, but increased again to just under half (49 per cent) in 2007.
As with property theft, there is no clear trend with respect to arrests. The number of arrests increased substantially between the 1998 and 2003 surveys, from only 6 per cent to 22 per cent, but the number of arrests dropped off a little in 2007 to reach 17 per cent.



Definition: Assault includes being attacked, physically beaten or threatened by an attacker in a frightening way without the attacker taking any property from the victim. This includes domestic assault.


The incidence of assault across provinces varies from the other crimes discussed in this paper (Table 6). While Gauteng and KwaZulu-Natal saw the highest and second highest levels of assault in the 2003 and 2007 surveys, the Eastern Cape showed the most assaults in 1998, followed by the Free State and North West province. KwaZulu-Natal and Mpumalanga were tied in fourth place, followed by Gauteng. As previously, however, the Northern Cape showed the lowest levels of assault.

The steepest increase in the number of assaults between the 1998 and 2003 surveys occurred in Gauteng, where the proportion of respondents assaulted doubled. The Free State saw the greatest declines. Despite the large increase between the 1998 and 2007 surveys, the number of assaults recorded for Gauteng in 2007 was still lower than recorded in 2003, when more than one out of every three respondents had been assaulted. Limpopo showed the largest increase in levels of assault between the 2003 and 2007 surveys, increasing 5 percentage points from 8 per cent to 13 per cent.


In the 2003 survey, assaults were most likely to occur in the home, and to a much lesser extent on the street; either in a residential area or near shops and other venues. As Figure 6 shows, fewer victims reported that they had been assaulted at home in 2007, with respondents equally likely (36 per cent) to report having been attacked at home and in the street away from a residential area.

The majority of those assaulted in all three surveys were male. The proportion of male as opposed to female victims increased between the 1998 (51 per cent), 2003 (57 per cent) and 2007 surveys (59 per cent).


Figure 6: Where assault occurred, 2003 and 2007

Most assaults were carried out by people known to the victims. The data from the 1998 and the later surveys are not directly comparable, but comparison between the 2003 and 2007 results shows that respondents were most likely to have been assaulted by known community members (29 and 24 per cent in 2003 and 2007 respectively). Almost one out of every five were assaulted by a friend or acquaintance (18 per cent in 2007 and 17 per cent in 2003), one  in ten (13 per cent) by a spouse or partner, and a minority by other relatives or household members. Only one third had been assaulted by a stranger (30 and 24 per cent in 2003 and 2007) or the police (4 and 3 per cent).

The results suggest relatively high levels of domestic abuse, particularly in 2007. The majority (93 and 70 per cent) of those assaulted by their spouse or partner were women. Two out of five (39 per cent) of these occurred in the home in 2007, as did one in five (24 per cent) in 2003.

Alcohol and drugs play a role in many assaults. The 1998 survey did not ask respondents about drug and alcohol use at the time of the attack, but just under half of the victims spoken to in 2007 and 2003 (48 and 47 per cent respectively) reported that their attacker was under the influence of drugs or alcohol. The number reporting that they were under the influence at the time dropped between the 2003 and 2007 surveys, from one in four (28 per cent) to one in six (17 per cent). Men were more likely than women to report that their attacker was inebriated (70 per cent compared to 30 per cent). This supports the perception that much interpersonal violence amongst men occurs in the context of heavy drinking.

The results suggest that weapon-use is declining. While just over half of victims reported that weapons were used in the attack in 1998 (56 per cent) and 2003 (55 per cent), this dropped to just one third (30 per cent) in 2007. The majority (57 per cent) of those spoken to in 1998 reported that their attacker had used weapons such as a panga, axe, stick or club, while in 2003 30 per cent of perpetrators tended to have used such weapons; or a knife (36 per cent).

Respondents in 2003 and 2007 surveys were more likely than their counterparts in 1998 to have been injured (71 and 70 compared to 49 per cent) – although the differences in the phrasing of the question in the 2003 and 2007 surveys may over-represent the injuries reported in 2007. The majority required medical attention (82, 78 and 76 per cent in 2007, 2003 and 1998 respectively), with just under half admitted to hospital.

Reporting rates for assault have increased steadily across the three surveys. While two out every five victims (38 per cent) reported the assault in 1998, just over half (55 per cent) reported it in 2003, while three out of four (76 per cent) reported it in 2007. Men and women were almost equally likely to report (49 and 52 per cent respectively). Very few respondents reported being pressurised not to report (10 and 6 per cent in 2003 and 2007 respectively).

As with the crimes discussed already, there is no clear pattern regarding arrests. Only one in ten (11 per cent) reports resulted in arrests in 1998. This rose to one in three (32 per cent) in 2003, but decreased to a little less than one in six (14 per cent) in 2007.



Definition: The abuse of entrusted power for private benefit


The 2003 and 2007 surveys asked South Africans about their experience of corruption, primarily in the public sector. Respondents were most likely to have been asked to pay bribes in relation to policing, particularly traffic fines, with one out of every three (33 per cent) South Africans having been asked to pay a bribe in connection with a traffic fine. In both 2003 and 2007 most officials requested bribes in the form of money (77 and 84 per cent), although respondents reported more requests for presents and favours in 2007 than in 2003 (12 and 11 per cent compared to 7 and 9 per cent in 2003).


Figure 8: Changes in the solicitation of bribes by government service area between 2003 and 2007

Requests by officials for bribes have increased in several service areas since 2003. The greatest increases occurred in relation to traffic offences, correctional services and the transport authorities, with the number of people asked to pay a bribe to adjust or discard a traffic fine, when visiting a prison or obtaining a driver`s licence increasing by 5 per cent (Figure 8). The greatest decreases have occurred in relation to public sector employment, where the number of people asked to pay a bribe to obtain a job dropped by 6 percentage points.

Analysis of the results by province suggests that levels of corruption have remained fairly stable in several provinces, but have increased in Gauteng. As shown in Figure 9, while one out of every three (34 per cent) respondents reported that they had been asked to pay a bribe in Gauteng in 2003, more than two out of every five (45 per cent) reported receiving such requests in 2007. Levels of corruption decreased in KwaZulu-Natal, the Free State and the Western Cape, with KwaZulu-Natal seeing the largest decline (from 21 to 14 per cent).


Figure 9: Solicitation of bribes by province in 2003 and 2007

In most cases, those who were asked to pay a bribe did so. Payment levels were highest in relation to traffic fines and customs procedures, where three out of four (76 per cent) and a little more than two out of every three (69 per cent) respondents respectively reported paying the bribe. They were least likely to pay bribes requested in relation to court services (17 per cent), schooling (31 per cent) or when visiting someone in prison (28 per cent). These findings suggest that South Africans are relatively permissive when it comes to what may be perceived as ‘minor` transgressions.

Research conducted in 2007 by Ipsos-Markinor provides a counterpoint to these results. The bi-annual Government Performance Barometer Survey, conducted amongst a nationally representative sample of 3 500 South Africans, asked respondents to rank certain actions on a scale of one to ten according to whether they thought they were always justified (10), never justified (1) or something in between. As shown in Table 7, respondents were surprisingly accepting of tax evasion, avoiding paying for transport or illegitimately accepting government grants. They were also quite open to people accepting bribes and paying a traffic officer to quash a speeding fine.


Table 7: South Africans` acceptance of illegal activities

Corruption appears increasingly opportunistic. In 2003, the majority (61 per cent) of those who refused to pay reported that the official had as a result refused to assist them, while just over one in four (27 per cent) helped them anyway. In 2007, more than half (53 per cent) reported that the official had assisted anyway, while slightly fewer than one quarter (22 per cent) reported that they refused to assist them. Of concern, however, is that the same proportion (22 per cent) stated that the official threatened them with harm (compared to 8 per cent in 2003), suggesting the potential for greater aggressiveness in the future. 

Levels of corruption in the private sector appear much lower than in the public sector in both the 2003 and 2007 surveys. The results show that only 6 per cent of respondents in 2007 had been asked to pay a bribe to speed up a job application, slightly up from 4 per cent in 2003. Only 5 per cent had been asked to pay a bribe to obtain a job in 2003 and 2007.

Summary and discussion

Several key points emerge from the findings discussed in this paper (Table 8). In addition to specific provincial dynamics, the findings highlight two particularly interesting findings.

The first is the seeming intrusion of crime into respondents` homes and personal spaces – perhaps one reason why South Africans` perceptions of crime have worsened. While more data is required to establish firmly the trends over time, this analysis shows that while much crime occurs in public areas and on the street, crimes such as robbery often occur in the home and in the street in residential areas, presumably often when victims are leaving or returning home or visiting friends and family. Perhaps less surprisingly, crimes like assault frequently occur in the home, while housebreakings increasingly occur when people are at home. These findings suggest the need for greater policing in residential areas – and strengthening the police`s ability to respond to domestic violence. They also underscore the importance of South African`s putting in place measures to protect their homes.

Table 8: Summary of key points


The second is the relatively large number of people who know who victimised them. It is remarkable that not only do many victims know who attacked them or stole their things, but that so many know them well enough to recognise them by name or face.  These levels of familiarity highlight the essential role that communities must play in addressing crime in South Africa. Rather than holding the police solely responsible for reducing crime, communities must become active, proactive partners in addressing crime.

In the short-term, this requires not only promoting and strengthening community-policing interfaces such as Community Police Forums, but also improving the accessibility of anonymous reporting hotlines, as well as the public`s awareness of such mechanisms. More needs to be done to educate the public about the SAPS Crime Stop hotline and the number people should call to tip off the police. The line should be advertised widely in the print and audio-visual media, and promotional materials should emphasise that callers can remain anonymous.

In the longer-term, it requires engaging with the public to change the often permissive attitudes that exist towards crime and criminals in many communities. This is primarily about engendering a fundamental shift in the way South Africans view crime. Government, non-governmental organisations and the private sector should take the lead in promoting discussion at the local and national levels about what is acceptable and unacceptable behaviour. They can also help by acknowledging and raising awareness about the ramifications of not reporting known criminals. It is also partly about helping enabling the public to trust the police. Government needs to do more to stamp out corruption within the police and promote and support a more positive public image of the police. 


The data from the 1998, 2003 and 2007 surveys highlights the often variable dynamics between the most frequently experienced crimes in South Africa, but as varied as these are, the data also show trends shared by different crime types. Particular points of commonality include the particular challenges faced by Gauteng and KwaZulu-Natal in addressing crime; the continued prevalence of serious domestic violence; and, the permissiveness of South African society towards crime – whether in terms of paying bribes or knowing who commits crime without necessarily reporting them. 

These findings speak to the depth of the crime problem, and the relatively long way that the country has still to go before crime is addressed effectively. They also underscore the important role that all aspects of society must play in reducing crime. The police alone can not tackle crime, nor can better functioning courts; or a more responsive criminal justice system as a whole. The public needs to take greater responsibility for addressing crime; whether through adopting measures to reduce the likelihood of victimisation; reporting suspects to the police, or demanding that the police perform their duties satisfactorily.

Finally, the findings support the need for a broad approach to addressing crime that encompasses not only making the criminal justice system work better, but also changing public attitudes towards crime and addressing the underlying socio-economic factors that feed both crime and these perceptions on the acceptability of crime. Efforts to mitigate and respond to crime must be balanced by longer-term systemic changes aimed addressing the causes of crime. Bringing about these changes goes beyond the mandate of the police. Social crime prevention must be a priority that is integrated across a range of government departments, most immediately the departments of social development and education. 


  1. Differences in the way the question was asked in 1998 and 2003 may reduce the comparability of the results between these surveys. The 1998 survey asked respondents if, ‘…any government official, for instance a customs official, police officer or inspector asked you or wanted you to pay a bribe for his/her service?` In contrast the 2003 survey asked respondents if ‘…any government or public official asked you or indicated to you that they would be receptive to the following: money, a favour, a present in return for a service that the official is legally required to perform?` As noted in Burton et al (2003) the description of a bribe was more comprehensive, including not only monetary forms of corruption but also favours and presents, which may have led to higher reporting of corruption in 2003 that do not necessarily reflect rising corruption levels.


Burton P, Du Plessis A, Louw A, Mistry D and Van Vuuren H 2003. National Victims of Crime Survey: South Africa 2003, ISS Monograph Series No. 101. Pretoria: Institute for Security Studies.

Crime Information Management 2008. Robbery at residential premises (subcategory of aggravated robbery) in the RSA per Province between April and March: 2001/2002 to 2007/2008. Pretoria: South African Police Service (SAPS). Available at [accessed November 2008].

Hammond, T and Pennington, S 2007. The Crime Statistics: Are we our own worst enemy? Johannesburg: Ipsos Markinor. Available at [accessed November 2008].


In 2007 the Institute for Security Studies conducted a national crime and victimisation survey. Th e survey followed on from two earlier victim surveys conducted by the ISS and Statistics South Africa in 1998 and 2003 respectively. Together, these three surveys provide an unprecedented snapshot of both changing victimisation levels and perceptions of crime and the criminal justice system over the last decade. Th is paper discusses the dynamics surrounding the six most common crimes identifi ed in the 2007 study. It compares the data from the 1998, 2003 and 2007 surveys to provide a picture of provincial victimisation trends, the sites and characteristics of the crimes, weapon use, victims` familiarity with the perpetrators and arrest and conviction trends.


Robyn Pharoah is an independent researcher working in Cape Town, in South Africa. She worked previously for the ISS Crime and Justice Programme in Pretoria and oversaw the running of the 2007 national victimisation survey.


This paper was made possible through the generous funding of Irish Aid.