Carjacking in Australia: recording issues and future directions

Foreword | Carjacking literature is limited, and perceptions vary about the level of violence involved, diverse scenarios and the motivations of offenders. The media tends to overrepresent carjackings involving weapons and violence, although these are relatively rare incidents. Motivations range from instrumental triggers (where the car is used in some other crime) to acquisition for onselling the car or its parts. Similarly, methods vary from opportunistic to organised theft involving support. This paper examines carjacking in Australia, the United Kingdom, the United States and South Africa reported through the literature. Victimisation surveys currently may be the most appropriate source for collecting carjacking estimates. Offence definitions and recording practices vary between Australian jurisdictions, making accurate estimates problematic.

Toni Makkai

The perception of carjacking as an increasingly frequent and violent form of crime is common within both law enforcement and the Australian community. This paper examines reports of increased carjacking in light of existing recording practices, international data and limited local information sources. Local socioeconomic and cultural factors, such as firearm availability, are also discussed in the context of probable future trends in Australia.

The term carjacking is commonly understood as vehicle theft involving threat, violence or intimidation. However, this has yet to be clarified within most legislative or offence definitions.

Carjacking literature is limited. However, US researchers compared media reports with police records in Louisiana to demonstrate that media coverage tends to focus on the more dramatic end of the spectrum (Cherbonneau & Copes 2003). Cases involving multiple offenders and weapons such as handguns and knives, were overrepresented in published media reports. Similarly, several Australian media reports in 2006 made reference to serious injuries. One incident that sparked particular public interest involved the carjacking of a prominent Sydney obstetrician on his way to deliver a baby. The doctor was locked in the boot of his prestige car, and freed by his captors after the vehicle was set alight (Sydney Morning Herald 19 March 2007).

Currently, there is no standardised approach to recording carjacking incidents in Australia. It is both a crime against the person and a form of property crime, also referred to as violent property crime (Indermaur 1995). Carjacking appears to comprise only a small percentage of vehicle thefts in Australia. While incidents are often attributed to organised offenders driven by financial gain via vehicle disposal, it is probable that carjacking motivations are more diverse. These are likely to include a need for transportation, fraudulent vehicle disposal and the intention to commit another crime, such as armed robbery.

In the recent past, structured text searches of law enforcement theft records and insurance claim records suggested that the majority of offenders were unarmed and the majority of incidents did not involve injury. The notion of carjacking as a violent crime appears to be driven by a media preference for those relatively rare incidents that involve significant victimisation.

Claims that increased carjacking is the result of improved vehicle security are common (for example, Davis 2003). These reports tend to point to the introduction of immobilisers, which are mandatory in all vehicles produced for the Australian market from September 2001 onwards. The targeting of Impreza WRX vehicles in 2001 and 2002 led Subaru Australia to incorporate an access code and keypad into their standard immobiliser. Anticipating a possible increase in carjacking, the manufacturer also fitted all WRXs from late 2002 with an anti-hijacking device triggered by forcible removal of the driver.

Such commentary linking security devices and carjacking generally fails to note that carjacking is possibly the most serious and highly visible subset of key theft. Other forms of key theft include residential burglary and robbery with the intention of accessing keys, or using keys left in a vehicle. Compared with armed carjacking, all offer relatively low risk to an offender.

Carjacking scenarios

The term carjacking encompasses a diverse range of scenarios in which vehicles and/or keys are stolen via threats, violence and/or intimidation. These include:

  • targeting of moving vehicles - by tailing and stopping a vehicle, or bumping (intentionally causing a minor rear-end collision)
  • targeting of stationary vehicles - at locations where keys may be accessible (such as petrol stations, automatic teller machines, traffic lights or car washes)
  • targeting vehicle owners - not in the immediate vicinity of their vehicle
  • home invasions - while rare, some incidents have involved securing a vehicle from the homeowner.

A small number of vehicle sellers have also been victimised. Offenders pose as interested buyers but steal vehicles during test drives. While this approach involves a level of interaction between the offender and victim, only incidents involving violence, threats or intimidation are considered carjacking.

Another variation involves misrepresentation during a vehicle 'purchase', for example the use of a fraudulent cheque and false personal identification. This is characterised by deception and therefore is not considered carjacking.

From a criminological perspective, the examination of several elements of carjacking incidents would help to clarify offender motivations and trends over time. This would help shape a standardised approach to monitoring and an evidence-based approach to the formulation of public policy. The elements requiring more detailed examination are:

  • violent approaches (involving assault, threat or intimidation) versus non-violent approaches (involving fraud or deception)
  • weapon use - absence versus presence, and weapon type
  • offenders - single versus multiple offenders
  • vehicle characteristics - recovery and damage status, security features and market value.

While it would be easy to suggest that organised vehicle crime offenders' methods would be characterised by violence and weapon use, this may not be the case. The analysis of insurance claim records suggests that the forceful removal of vehicle owners by 'interested buyers' while on a test drive is associated with high-value models fitted with good security and exceptionally low recovery rates. Despite the organised approach, weapons are rarely used.

Motivations for carjacking

The elements outlined above focus on the characteristics of a carjacking incident, such as the presence of a weapon or the number of offenders. These centre on the methods by which the carjackings proceed. Offender motivations seem equally important because these inform the methods ultimately used.

International research has sought to better understand why carjackers choose certain targets and most importantly, why they choose to carjack. Carjacking seldom generates cash for offenders (unlike conventional robbery) yet involves a highly risky confrontation that is not part of conventional motor vehicle theft or other forms of key theft. However, there are benefits that can appeal to some offenders. Unlike burglary or robbery, gains in the form of a vehicle are known in advance. While the process can be violent, it is faster than conventional motor vehicle theft (especially for vehicles with anti-theft devices) and the stolen vehicle is easier to resell because it comes with a key rather than a broken steering column (Jacobs, Topalli & Wright 2003).

As with research into other offences, carjackers' motivations have been examined using interview research, either with incarcerated carjackers (for example, in South Africa; Davis 2003) or active offenders (for example, in the US; Jacobs, Topalli & Wright 2003). As in other offences, there is no single set of motivating factors, but rather a continuum, in this case from emotional triggers (like revenge) where the motor vehicle theft is only instrumental in achieving some other goal, to pure acquisitiveness. These correspond to some extent with the methods used to conduct carjacking, from spontaneous and relatively unplanned, to professional and almost businesslike. This correspondence is only partial, though, since instrumental carjackings can be well-organised and professionally executed (for example, carjacking team, backup vehicles, excessive high-powered weaponry, using a bump-and-rob technique). Similarly, offences motivated by the need to acquire can be seemingly opportunistic and amateurish in their execution (single offender, no weaponry).

It has been argued that carjacking is a unique type of offence because an object, the vehicle, is targeted rather than a subject, a person (Jacobs, Topalli & Wright 2003), and all carjacking is acquisitive to a degree. Interviews with South African offenders suggest that most carjackings they committed were motivated wholly by the need to acquire certain makes and models of vehicles, often to cater to a known market or fulfil a specific 'order' (Davis 2003). However, interview research also shows that some carjackers in a US city targeted certain individuals' vehicles to pay them back for flaunting their vehicles and for their arrogant behaviour (Jacobs, Topalli & Wright 2003). Some black South African offenders chose to target certain vehicles only after the (white) drivers locked their doors when they saw the offenders, even though no offence had yet occurred (Davis 2003). Power was also a motivating factor. South African hijackers referred to the thrill of forcing drivers - wealthy men in particular - to accede to violence, threats and intimidation (Davis 2003). Similarly, interviews with offenders operating in the United States concluded that participation in carjacking was 'activated, mediated and shaped by participation in urban street culture' (Jacobs, Topalli & Wright 2003: 673).

These continuums provide an imperfect but useful way of considering how motivations and methods might combine to result in a certain type of carjacking. Figure 1 illustrates this. For example, an organised-acquisitive incident might involve a carjacking team motivated by the need to acquire a certain luxury vehicle. The team arranges for a test drive, stops and threatens the salesperson with violence, and removes the vehicle. At the opposite ends of these dimensions, an opportunistic-instrumental incident might involve a pair of unarmed offenders forcibly removing the driver of an unlocked vehicle stopped at traffic lights, to joyride for the night.

Figure 1: Conceptualising offender motivation and examples of carjacking types

carjacking types

Unfortunately, it is not clear just how appropriate it is to consider local carjackings in this way, since very little is known about this offence in Australia. Recently reported incidents highlight that there is no single type of Australian carjacking. In August 2006, for example, two pairs of carjackings involving firearms and high-powered vehicles took place in south-western Sydney (NSW Police Force Media 14, 19, 23 & 27 August 2006). A four-wheel drive and a sedan were carjacked from carwashes. Each of these vehicles was then used to carjack a Subaru WRX. The first WRX was abandoned shortly after the theft and the second WRX theft was not completed although the driver was shot. All four carjackings involved multiple armed offenders. These incidents also illustrate that the relationship between criminal motivation and the use of violence is often not straightforward.

Media reports suggest an increase in frequency and geographical spread of violent carjacking in the capital cities but even the precise number of attacks per year in Australia is not known.

International trends in carjacking

United Kingdom

More than 1.1 million vehicles were stolen in England and Wales between 1998 and 2001. An examination of theft circumstances in a sample of over 8,000 late model (first registration in or after 1997) vehicles suggests the incidence of carjacking is not high.

Where modus operandi was known, key theft accounted for 85 percent of incidents and most commonly involved burglary (37%) or keys being left in the vehicle (18%). Robbery - including carjacking but also theft of keys from bags and coats - comprised only four percent of key thefts in 2001 (Levesley et al. 2004). Extrapolation of Levesley et al.'s figures suggests that around one percent of all vehicle thefts in the United Kingdom are the result of carjacking.

Despite the relatively small number of vehicle robberies examined (n=201), the researchers observed that this represented the doubling of key theft robberies since 1998. It also included a number of violent incidents. The nature of these incidents and type of weapon used were not specified. Levesley and colleagues (2004) also noted that key theft via burglary appeared to be rising, with close to half of all key thefts in 2001 being conducted this way.

United States of America

A 10-year victimisation survey conducted by the US Department of Justice estimated that an average of 38,000 carjackings occurred each year between 1993 and 2002 (Klaus 2004). Using a base estimate of 1,200,000 vehicle thefts, carjacking accounted for approximately three percent of all incidents in the United States.

Almost three-quarters (74%) of all offenders were armed - 45 percent with a firearm - and the presence of a weapon was associated with a higher probability of successful vehicle acquisition. Approximately 15 murders involving vehicle theft occurred each year, many during carjackings. Klaus (2004) further observed that the rate of US carjackings has declined over time, falling from an average of 2.1 incidents (1993 to 1997) to 1.3 incidents per 10,000 persons (1998 to 2002). The reasons for this decline are unclear.

South Africa

In South Africa, carjacking - or 'vehicle hijacking' - was recorded as early as 1976. Reported incidents began to increase in the late 1980s, and by 1991, were common enough to require a unique recording code. This predated the introduction of immobilisation systems in 1992 (Davis 2003). The number of hijackers armed with a handgun increased by more than 55 percent between 1995 and 1998. The majority of incidents involved multiple offenders, and tended to occur at night in the inner city or central business districts (Hennop, Potgieter & Jefferson 2001). Since 2001, recorded hijacking incidents have fallen by approximately 21 percent (South African Police Service 2005).

Socioeconomic circumstances quite different from those found in Australia appear to play a significant role in the popularity of carjacking in the United States and South Africa. Poverty, unemployment and the violent culture that evolved from the apartheid regime may lie behind the increase in South African carjacking.

Measuring carjacking in Australia

How carjacking is recorded currently

There are no means currently available to assess whether carjacking is occurring more often or in a wider range of locations. Difficulties in accurately capturing and counting all instances of carjacking in Australia relate to:

  • differing understandings of what the term carjacking describes
  • differing offences that could be involved in a carjacking incident
  • jurisdictional differences in legislation and police recording practices.

Media portrayals of carjacking tend to emphasise violence, highlighting any resulting injury or emotional distress to victims, although clearly one defining aspect of this crime is the actual targeting of a vehicle by thieves. If all incidents of carjacking are to be accurately captured and counted, what then becomes the critical defining component of the event that guides what is counted? An important element is the presence of threat or menace, consistent with the Australian Bureau of Statistics (ABS) definition of robbery:

The unlawful taking of property, with intent to permanently deprive the owner of the property, from the immediate possession, control, custody or care of a person or organisation, accompanied by the use, and/or threatened use of immediate force or violence (ABS 2006a: 41).

Does this mean that, for counting purposes, all carjacking incidents are actually captured in robbery statistics? The situation is unclear. National crime statistics such as ABS Recorded crime, victims report the number of victims of a subset of all possible criminal offences reported to or detected by police in all Australian jurisdictions. Because some incidents can result in multiple criminal offences under the Australian Standard Offence Classification (ASOC) (ABS 1997), a single victim of one incident might be recorded in crime statistics multiple times. A carjacking which involved threat, violence with a weapon, victim injury and the actual removal of the vehicle may result in a single victim being counted once each in the offence categories of assault, armed robbery and motor vehicle theft. If the theft was not completed (the vehicle was not stolen), victims would not be counted in motor vehicle theft statistics because this offence division does not record theft attempts. However, potentially they would still be counted as both armed robbery and assault victims. If the owner of the carjacked vehicle remained in the vehicle (for instance, the victim was ordered by offenders to proceed to an ATM and withdraw cash from their own account) they may also be counted in the kidnapping/abduction division. International research has also noted that the difficulty in accurately assessing rates of carjacking comes from this convergence of offences (see Cherbonneau & Copes 2003).

Victimisation surveys provide an alternative source of information about the prevalence of certain crimes. These surveys canvass samples of the population about their experiences of certain crimes and can provide valuable information about the extent of crime in the community. As with recorded crime statistics, the accurate capture of carjacking is dependent on a shared understanding of what is meant by the term. For instance, the USNational Crime Victimization Survey reports on experiences of carjacking, which is defined as 'completed or attempted robbery in which a car or other motor vehicle was taken or an attempt was made to take it and the offender was a stranger to the victim' (Klaus 2004). Carjacking is not specifically examined in the major national Australian crime victimisation survey, Crime and safety, Australia (ABS 2006b). However, respondents were questioned about robbery and assault, which could encompass instances of recent carjacking.

Detailed discussion of relevant legislative instruments, or argument for or against uniform national legislation, is beyond the scope of this paper. However, another obstacle to obtaining an accurate national count of carjacking is the differing criminal codes in Australian jurisdictions. A standalone offence of 'Taking motor vehicle or vessel with assault or with occupant on board' exists in NSW (Crimes Act 1900 (NSW) s. 154C). This captures the key elements of direct threat to a victim and the unlawful removal of the vehicle, and would allow counting of a subset of instances of carjacking as portrayed in the media. However, similar offences are not found in all states and territories. The circumstances of the incident may give rise to a range of differing offences depending on the state or territory where the offence occurred. Even if all jurisdictions enacted legislation to specifically address carjacking, a standalone offence emphasising threat might not capture the situation where a car with keys in the ignition is stolen from some probable location. Where an owner is momentarily absent from their vehicle, such as in a service station, this would be recorded as motor vehicle theft.

Differences in police recording practices in Australian jurisdictions, especially with respect to the offence of assault, have meant that national statistics reporting victims of assault are no longer produced (see ABS 2006b). If assault becomes a critical component of a uniform definition of carjacking, differing recording practices have the potential to influence the accuracy of any counts derived from recorded statistics concerning assault.

Moving towards a system of recording carjacking in Australia

In moving towards some uniform definition of carjacking that could be used to guide counting procedures, a clear and agreed understanding of its essential characteristics is necessary. Ideally, any definition will also include clear guidelines as to what carjacking is not. This definition could then underpin an agreed counting methodology that would permit the monitoring of trends over time. In turn, this would provide an empirical basis on which to formulate appropriate law enforcement, crime prevention and judicial responses.

Key elements in overseas definitions include threatened or actual assault, with a vehicle as the primary target. These are also the key elements of the standalone carjacking offence in NSW. Greater difficulty may emerge when deciding what carjacking is not. US counts exclude incidents where the offenders are known to the victims; however, this could exclude instances where carjacking is used as revenge against a peer. Other problematic scenarios that could be considered carjacking include the theft of a momentarily unattended vehicle from a service station, and the incidental theft of a motor vehicle, where the goal is the abduction or kidnapping of a driver or passenger.

The difficulties associated with multiple criminal codes and varying jurisdictional recording practices mean that victimisation surveys may be the most appropriate way to gather national carjacking estimates in Australia at present. However, this is still contingent on a shared and easily understood definition of this crime.

Carjacking in Australia: possible future directions

Slightly fewer than 74,000 vehicle thefts were recorded in Australia in 2006 (National Motor Vehicle Theft Reduction Council 2007). Analysis of police and insurance records suggests that fewer than 300 carjackings (as defined above) occur each year. This equates to less than half of one percent of all theft incidents and broadly accords with United Kingdom findings (Levesley et al. 2004).

The same intelligence suggests that not all Australian carjackings involved vehicles fitted with immobiliser systems, and the great majority of vehicles were recovered. A small number of incidents involved the high-visibility theft of valuable vehicles that were then immediately destroyed by fire, giving rise to the possibility that the owners may have staged the incident before filing a fraudulent insurance claim.

An estimate of fewer than 300 incidents is supported by charge records relating to s. 154C of the Crimes Act 1900 (NSW). Since 2002, the greatest number of charges laid for carjacking in NSW was 63, in 2005. This fell to 38 in 2006.

Combined, these observations offer little support for carjacking as a preferred approach for Australian-based syndicates stealing for profit through vehicle re-identification, the sale of spare parts or the export of stolen cars.

Police intelligence also indicates that organised vehicle crime syndicates are adept at circumventing some types of immobilisers, making it unlikely that participants would wish to attract law enforcement attention by engaging in violent carjacking.

An estimate of 300 carjacking incidents in Australia equates to 0.15 incidents per 10,000 population (estimated using 2006 population; ABS 2006c). Compared with rates of 2.7 and 1.3 per 10,000 population in South Africa and the US respectively (South African Police Service 2005; Klaus 2004), the incidence of carjacking in Australia is relatively low. It is possible that this difference is mediated by the availability of firearms. Forty-five percent of US carjackers used a firearm. Handguns were the most common weapon type, and were strongly associated with the probability of a successful outcome for the offender (Klaus 2004). Interviews with incarcerated South African carjackers also showed a strong preference for handguns, especially pistols (although assault rifles were also sometimes used where the offence involved multiple vehicles and offenders) (Davis 2003). In the United Kingdom and Australia, firearm-related crime is less prevalent, and domestic burglary may be the preferred approach to key theft.


At present, it is not possible to determine the incidence of carjacking with any degree of accuracy. However, previously conducted intelligence analyses suggest that, at most, only a few hundred vehicles are stolen by threat, force or intimidation in a year. The majority of vehicles carjacked in Australia do not appear to be stolen for profit via disposal of the stolen vehicle, but targeted by opportunistic or semi-organised groups for subsequent crime, fraudulent insurance claimants or opportunistic individuals in need of transportation.

There is little support for the proposition that Australia will experience a steady rise in carjackings as a direct result of improved vehicle security. While the introduction of immobilisers to the vehicle fleet will progressively reduce the number of available targets for opportunistic offenders with limited technical skills, only a small number would consider the high-risk tactic of carjacking. As in the United Kingdom, those who continue their involvement in vehicle crime will tend to prefer burglary as a means of accessing keys. Limited available information suggests that local offenders engaged in vehicle crime for transportation or insurance fraud are generally unarmed. In addition, the broader socioeconomic picture of Australian society is one of relative good health in terms of wealth distribution and social cohesion, generating little in the way of strong motivation for victimisation that is both personal and violent.

This need not preclude the emergence of isolated hot times or hotspots. These are crimes of the same general type that are clustered in time or space, but are not necessarily perpetrated by the same offender or group. Sydney has experienced a number of carjacking clusters since the late 1990s, each lasting around three to six months and occurring in different locations including the eastern suburbs, the inner city and the south-west. While rare, it is likely that such clusters are the work of organised and semi-organised groups working for a particular criminal purpose involving financial profit.


  • All URLs were correct at 8 February 2008
  • Australian Bureau of Statistics (ABS) 1997. Australian standard offence classification, 1997. ABS cat. No. 1234.0. Canberra: Australian Bureau of Statistics
  • Australian Bureau of Statistics (ABS) 2006a. Recorded crime, victims, Australia, 2005. ABS cat. no. 4510.0. Canberra: Australian Bureau of Statistics
  • Australian Bureau of Statistics (ABS) 2006b. Crime and safety, Australia, April 2005. ABS cat. no. 4509.0. Canberra: Australian Bureau of Statistics
  • Australian Bureau of Statistics (ABS) 2006c. Population by age and sex, Australian states and territories, June 2006. ABS cat. no. 3201.0. Canberra: Australian Bureau of Statistics
  • Cherbonneau MG & Copes H 2003. Media construction of carjacking: a content analysis of newspaper articles from 1993-2002. Journal of crime and justice 26(2): 1-21
  • National Motor Vehicle Theft Reduction Council. CARS quickstats.
  • Davis L 2003. Carjacking: insights from South Africa to a new crime problem. Australian and New Zealand journal of criminology 26(2): 173-191
  • Hennop E, Potgieter J & Jefferson C 2001. The role of firearms in crime in South Africa: a detailed analysis of police dockets. Monograph no. 55. Cape Town: Institute for Security Studies.
  • Indermaur D 1995. Violent property crime. Sydney: Federation Press
  • Jacobs BA, Topalli V & Wright R 2003. Carjacking, streetlife and offender motivation. British journal of criminology 43(4): 673-688
  • Klaus P 2004. National Crime Victimization Survey, carjacking, 1993-2002. Crime data brief. July. Washington, DC: Bureau of Justice Statistics.
  • Levesley T, Braun G, Wilkinson M & Powell C 2004. Emerging methods of car theft - theft of keys. Home Office Findings 239. London: Home Office.
  • NSW Police Force media n.d. [accessed August 2006].
  • South African Police Service 2005. Carjacking (sub category of robbery aggravating) in the RSA for the period April to March 2001/2002 to 2004/2005. Crime Information Analysis Service.

About the authors

Ms Lisa Jane Young is a lecturer in criminal intelligence at the Australian Graduate School of Policing, Charles Sturt University. Dr Maria Borzycki is research analyst in the crime monitoring program at the AIC.