Australian Institute of Criminology

Skip to content

Examining adult-onset offending: A case for adult cautioning

Trends & issues in crime and criminal justice no. 488

Carleen Thompson, Anna Stewart, Troy Allard, April Chrzanowski, Chelsea Luker and Jerneja Sveticic
ISSN 1836-2206
Canberra: Australian Institute of Criminology, October 2014

Foreword | Very little is known about adult-onset offenders. This makes it difficult to know the most effective way for the criminal justice system to respond to these offenders. This project examined the nature of adult-onset offending in the 1983–84 Queensland Longitudinal Data Cohort and explored whether adult cautioning may be a suitable and cost-effective alternative to current court processing. Half of all offenders in this cohort started offending in adulthood (between 18 and 25 years), however, most adult-onset offenders had just one or two relatively less serious officially recorded offences. The authors argue that extending formal police cautioning to include first-time, less serious adult-onset offenders is a cost-effective strategy that would enable scarce criminal justice resources to be redirected to provide evidence-based interventions for more serious and prolific offenders who present an ongoing risk of offending.

Adam Tomison

Criminologists have traditionally considered adult-onset offending to be a rare phenomenon (Eggleston & Laub 2002). Consequently, little criminological theory, research or policy has focused on adult-onset offending. However, an emerging body of research suggests that a substantial number of offenders have their first contact with the criminal justice system (CJS) at 18 years of age or older (Delisi & Piquero 2011). Despite increasing interest in adult-onset offenders, the nature of adult-onset offending is still poorly understood. Moreover, it is unclear whether traditional criminal justice responses for adult offenders are appropriate for adult-onset offenders. In this study, the extent, nature and costs of adult-onset offending are examined, alongside the appropriateness and cost-effectiveness of current criminal justice responses.

The limited research examining adult-onset offenders indicates that these offenders have lower rates of reconviction, commit far fewer crimes and perpetrate less serious offences than early-onset offenders (eg Carrington, Matarazzo & deSouza 2005; Kratzer & Hodgins 1999). However, in some cases, the criminal careers of adult-onset offenders are extensive and serious (Delisi & Piquero 2011). Recent research in two population-based Queensland offender cohorts identified a clear and prevalent low-rate, adult-onset offender trajectory (Allard, Chrzanowski & Stewart 2012; Allard et al. 2014). Although a high-rate, adult-onset offender trajectory was not identified, a small late-onset chronic offender trajectory was identified that included offenders with an onset at 18 years or older.

Together, this research suggests that for many adult-onset offenders, their criminal career may be brief and less serious. However, for some adult-onset offenders, their criminal career may be both chronic and serious. To date, research has not disaggregated adult-onset offenders across severity or chronicity. If both low-rate/less serious and chronic/serious groups of adult-onset offenders can be identified, this has important implications for responding to these offenders.

According to best practice principles of offender rehabilitation, sanctions and interventions should be commensurate with the level of risk posed by an offender (Andrews & Dowden 2006). Intensive interventions should be reserved for chronic offenders who pose an ongoing risk. For low-risk offenders, CJS interventions should be minimised or even avoided, as such interventions may unintentionally increase the likelihood of reoffending (Andrews & Dowden 2006). For these offenders, diversion, such as formal police cautioning, may be a more appropriate, efficient and cost-effective response than current practices of court processing.

Although formal police cautioning for adults is not legislated in Queensland, police policy enables cautions to be used for minor offences perpetrated by individuals over 65 years or with intellectual disabilities (QPS 2012). Formal adult cautioning is also used for limited offences in other jurisdictions in Australia such as for minor drug offences and shoplifting in Victoria (Victoria Police 2012) and for possessing cannabis in New South Wales and Tasmania (NCPIC 2013). Furthermore, broader adult cautioning schemes operate overseas for predominantly less serious and first-time offences (eg England and Wales; Ministry of Justice 2013). Formal police cautioning is also routinely used for youths across a broad range of offences in all jurisdictions in Australia (Little & Allard 2011).

Evaluations of formal police cautioning schemes support the effectiveness of cautioning for recidivism and cost savings, particularly for low-risk offenders (eg Allard et al. 2010). If most adult-onset offenders are low-rate, low-risk offenders, diversion may be a more appropriate response to most adult-onset offending rather than processing these individuals through the adult courts.

In this study, the extent, nature and costs of adult-onset offending is investigated, as well as potential variability in the chronicity of adult-onset offending. Analyses compare adult-onset offenders and early-onset offenders to determine if and how these offenders differ. Finally, given the frequent use of cautioning with less serious youth offending and the introduction of cautioning (of varying levels of inclusiveness) for adults nationally and internationally, the cost implications associated with cautioning low-rate, less serious adult-onset offenders is investigated.


Data sources

Data from the 1983/1984 Queensland Longitudinal Dataset (83/84 QLD) were used in this study. The 83/84 QLD includes data about all offences committed in Queensland by individuals born in 1983 or 1984, between 10 and 25 years of age, that resulted in formal (youth) police cautions, youth justice conferences, youth court finalisations or adult court finalisations. This offence-level data was obtained by linking data from the Queensland Police Service, Department of Communities and Department of Justice and Attorney-General. Data across these sources were linked and aggregated at the individual level to create a population-based Queensland offender cohort with complete official offence histories to age 25 years (N=54,598 individuals). See Allard et al. (2014) for the data linkage process.

In this study, the following offences were excluded from the 83/84 QLD:

(a) offences that resulted in not guilty verdicts;

(b) breaches of justice orders, as these offences are often technical breaches and any offending behaviour that led to breaches is already recorded in the dataset; and

(c) minor traffic offences classified under the Australian and New Zealand Offence Classification (ANZSOC; ABS 2011) division 14, as these offences are offences against the Traffic Act rather than Criminal Code. For most Traffic Act offences, only individuals who contest their State Penalties Enforcement Registry tickets are processed in the courts.

Offenders were also excluded if their usual residence was interstate or overseas because their complete offending histories were not available. Given these exclusions, the total sample was 40,523 offenders (25.9% female; 8.9% Indigenous Australian). These individuals were responsible for 206,857 offences.

Classifying adult-onset offenders

Definitions of adulthood differ across studies (eg 18, 21, 25 years; Krohn, Gibson & Thornberry 2013). However, offenders are treated as adults from 18 years in most legal systems (although not in Queensland; Chrzanowski & Wallis 2011). Developmental psychologists also identify 18 years of age as the start of a new developmental period in contemporary societies – ‘emerging adulthood’ – that markedly differs from adolescence (eg 18 year olds have new rights and responsibilities, have gone through puberty and have typically completed secondary schooling; Arnett 2000). Therefore, individuals were classified as adult-onset offenders in this study if their first official criminal justice contact was for an offence perpetrated at 18 years or older. All other offenders were classified as early-onset offenders.

As individuals in Queensland are processed in the adult courts for offences allegedly perpetrated from 17 years of age (not 18 years like other states), adult-onset offenders were identified using their age of onset. Age of onset was calculated using date of birth and the date that individuals perpetrated their first offence recorded in the 83/84 QLD (applying exclusion criteria). When the date of the offence was not available, the date of lodgement was used for court contacts. All youth cautioning, youth conferencing and youth court contacts pertained to offences perpetrated in youth and were classified as early-onset offending.

Table 1 Estimated costs of offending to the criminal justice system and wider society ($)
Criminal justice system processing costs $ Wider social and economic costs $
Youth caution 1,275 Homicide and related offences 2,329,919
Adult caution 1,103 Sexual assault and related offences 9,123
Conference 5,519 Property damage and environmental pollution 4,084
Children’s court finalisation 4,373 Unlawful entry with intent 3,490
Magistrates court finalisation 3,090 Robbery, extortion and related offences 2,798
District court finalisation 9,352 Acts intended to cause injury 2,062
Supreme court finalisation 10,663 Theft and related offences 1,510
Youth detention (per day) 567 Abduction/harassment/other offences against the person 1,000
Adult incarceration (per day) 289 Fraud, deception and related offences 517
Youth community-based supervision (per day) 35 Prohibited/regulated weapons and explosives offences and illicit drug offences 500
Adult community-based supervision (per day) 12 Other offence types 250

Source: Allard et al. 2014

Table 2 Rate and nature of offending across high and low-rate, adult-onset offenders
Rate of offending Proportion of adult-onset offenders Total offences Number of offencesa Number of eventsb Ever committed serious offencec Ever received supervised orderd Ever received suspended sentencee Ever been imprisoned as adultf
N % N % M SD M SD % % % %
Low rate 19,814 93.4 35,495 67.7 1.8 1.3 1.4 0.9 8.3 6.1 1.7 1.5
High rate 1,399 6.6 16,923 32.3 12.1 9.9 4.6 3.5 38.7 49.0 16.8 19.9

a: t(1401.4)=-38.9,p=.001,d=-1.46

b: t(1410.2)=-33.1,p=.001,d=-1.25

c: χ2(1, N=21,213)=1313.5,p=.001,φ=.25

d: χ2(1, N=21,213)=2960.8,p=.001,φ=.37

e: χ2(1, N=21,213)=1135.1,p=.001,φ=.23

f: χ2(1, N=21,213)=1656.8,p=.001,φ=.28

Rate and nature of offending

The rate and nature of offending was examined using five variables. First, the rate of offending per individual was calculated based on the number of offences for each individual (between 10 and 25 years of age). Second, the number of events per individual was calculated based on the number of formal (youth) police cautions, youth justice conferences, youth court finalisations and adult court finalisations for each individual. Third, offence types were categorised using the ANZSOC (ABS 2011) system (excluding division 14, traffic offences) that classifies offences into 16 divisions. Fourth, seriousness of offending was measured using the National Offence Index (NOI; ABS 2009) that ranks the ANZSOC multi-digit codes by their level of seriousness. The following categories of seriousness were applied to each offence in the database—serious (NOI 1–30), moderate (NOI 31–93), or minor (NOI 94–157; see Thompson et al. 2013). Fifth, outcomes of offending were categorised as—diverted from formal order, non-supervised order, community supervision, suspended sentence, detention or prison.

Costs of offending

The costs of offending were assessed using the cost estimates developed by Allard and colleagues (2014). These costings include both CJS costs and the wider social and economic costs of crime (see Table 1). The CJS costing methodology used the Transactional and Institutional Cost Analysis to assess the average opportunity costs of CJS events (eg caution, Magistrates Court), taking into account police, court and supervision costs. For example, the cost of a caution only involved police resources, whereas the cost of a Magistrates Court finalisation involved both police ($2,696) and court ($394) resources. Wider social and economic costs were assessed based on offence type. These costs were assessed using a bottom-up costing approach that involved updating Rollings’ (2008) original assessment and mapping costs to ANZSOC codes.


What is the extent and nature of adult-onset offending?

Half of offenders initiated official offending at 18 years of age or older (n=21,213; 52.3%). Approximately 80 percent of adult-onset offenders were male and 6.3 percent were Indigenous Australian. By age 25, adult-onset offenders were responsible for 25.3 percent of all offences and 32.3 percent of all events in the dataset. Most adult-onset offenders perpetrated one offence (56.6%) or one or two offences (75.1%), although 8.8 percent perpetrated five or more offences (M=2.5, SD=3.8, max=118). Over two-thirds (70.4%) of adult-onset offenders had just one finalisation (M=1.6 finalised events, SD=1.5, range=1 to 44 events). The most serious offence ever perpetrated was minor in nature for 42.7 percent of adult-onset offenders, moderate for 47.0 percent of adult-onset offenders and serious for 10.3 percent of adult-onset offenders. For 87.6 percent of adult-onset offenders, the most serious outcome recorded in the dataset was a non-supervised order. However, 7.6 percent received at least one community supervision order, 2.1 percent received at least one suspended sentence and 2.7 percent received at least one prison sentence.

To differentiate between low-rate and high-rate adult-onset offenders, the offender trajectories for the 83/84 QLD identified in Allard and colleagues’ (2014) research were used (ie adolescent onset–low, adult onset–low, adolescent onset–moderate, adolescent onset–chronic, early onset–chronic).

Adult-onset offenders were considered low-rate offenders if they were classified on a low-rate trajectory. Adult-onset offenders were considered high-rate offenders if they were classified on a moderate or chronic trajectory (see Thompson et al. 2013). Using these criteria, 93.4 percent of adult-onset offenders were classified as low-rate offenders. Just 6.6 percent of adult-onset offenders were classified as high-rate offenders. These high-rate, adult-onset offenders were responsible for many more offences, much more serious offences and received more severe sentences than the low-rate, adult-onset offenders (see Table 2).

How do low-rate, adult-onset offenders differ from low-rate, early-onset offenders?

Using Allard and colleagues’ (2014) trajectories, 14,149 offenders were classified as low-rate, early-onset offenders. These offenders were compared with the 19,814 low-rate adult-onset offenders identified above. The results indicated that while Indigenous status did not vary across the offender groups, females were significantly more likely to be early-onset, low-rate offenders than adult-onset, low-rate offenders (see Table 3). On average, adult-onset offenders perpetrated slightly fewer offences and were responsible for slightly fewer finalisations than early-onset offenders. This may be because adult-onset offenders had less time to reoffend due to a later onset of offending.

The nature of offending differed between low-rate early-onset and adult-onset offenders. Although serious offences were infrequent for both groups of offenders, early-onset offenders were more likely than adult-onset offenders to have perpetrated serious offences. Additionally, early-onset offenders were much more likely than adult-onset offenders to perpetrate property offences (see Table 4), with nearly 70 percent of the early-onset offenders having at least one property offence (predominantly theft from retail premises but also unlawful entry and property damage) compared with one-quarter of adult-onset offenders. Early-onset offenders were also more likely to perpetrate acts intended to cause injury than adult-onset offenders. By contrast, adult-onset offenders were more likely than early-onset offenders to perpetrate public order offences (primarily offensive behaviour and disorderly conduct), dangerous or negligent acts endangering persons (14.5% dangerous or negligent operation of a vehicle; 10% dangerous/negligent driving under the influence of alcohol or other substances) and offences against justice procedures (predominantly resist or hinder a police officer or justice official). Early-onset and adult-onset, low-rate offenders also had similarly low rates of imprisonment/detention and supervised orders. Not surprisingly, given the differences in CJS responses available in the juvenile and adult justice systems, early-onset offenders were much more likely to have been diverted and adult-onset offenders were much more likely to have received a non-supervised order.

The average cost of offenders in the two low-rate groups were compared in terms of their wider social and economic costs and criminal justice processing costs (see Table 5). On average, each early-onset offender cost more socially and economically than each adult-onset offender. This reflects the more serious and costly offence types committed by early-onset offenders, as well as somewhat higher rates of offending. In terms of CJS costs, on average, each early-onset offender cost less than each adult-onset offender. Given that it generally costs more to process youth through to court than adults and that youth supervision costs are substantially more than adult supervision costs, this reflects the fact that most youth are diverted to police cautioning, which has much lower cost implications for the CJS.

What are the cost-implications associated with formally cautioning low-rate, less serious adult-onset offenders?

Given the nature and extent of the low-rate, adult-onset offending identified, the cost savings that could have accrued if police cautioning had been an available CJS response for adult-onset offenders in Queensland was assessed. The cautioning estimates were calculated in three steps. First, as cautions are typically reserved for first-time offenders and less serious offences (Allard et al. 2010), the costs of cautions were only estimated for low-rate, adult-onset offenders’ (a) first events (b) that were finalised in the Magistrates Court and (c) received a maximum penalty of a non-supervised order (eg convicted but not punished, fine). Using these criteria, 18,646 low-rate, adult-onset offenders (94.1%) could have been candidates for cautioning at their onset event. Second, the estimated cost of a caution (ie $1,103; Allard et al. 2014) was compared with the justice system costs ($3,090) that were actually incurred for these cases (total saving=$37m). Third, as 82 percent (n=15,286) of the low-rate offenders ‘eligible’ for cautioning were fined, the potential revenue that would have otherwise accrued from court ordered fines was deducted. This totalled $4.5m (after adjusting for administrative and enforcement costs of fines and unpaid fines). After these adjustments, the cost savings for the CJS for formally cautioning these first-time less serious adult-onset offenders for their first event, rather than processing them through the adult courts, was estimated at $32.5m.


This research examined the extent and nature of adult-onset offending and investigated whether formal police cautioning could be a viable and cost-effective alternative to current court processing for adult-onset offenders. The research generated four key findings. First, adult-onset offenders were prevalent. Second, the vast majority of adult-onset offenders were low-rate, less serious offenders. Third, low-rate, adult-onset offenders and low-rate, early-onset offenders had similarly low rates and less serious patterns of offending, even though they perpetrated different types of offences. Fourth, cautioning low-rate, less serious adult-onset offenders would produce substantial cost savings.

Just over half of all offenders in the cohort initiated (official) offending in adulthood. The sheer magnitude of adult-onset offending provides a compelling argument for investigating the nature of this phenomenon. It is considered that, this study is the first to disaggregate adult-onset offenders by their chronicity of offending. However, consistent with previous research (Carrington, Matarazzo & deSouza 2005; Kratzer & Hodgins 1999), adult-onset offenders were predominantly less serious, low-rate offenders. In most cases, adult-onset offenders perpetrated just one or two offences that were minor or moderate in nature and resulted in non-supervised orders.

Just 6.6 percent of adult-onset offenders had a moderate or chronic pattern of offending that was more serious in nature. Similar to research with early-onset chronic offenders (Piquero 2008; Wolfgang, Figlio & Sellin 1972), high-rate adult-onset offenders in this study were responsible for a disproportionate amount of offences in general and serious offences in particular. Therefore, while adult-onset offenders are usually treated as one homogenous group, a criminal onset in adulthood can denote the beginning of markedly different criminal careers.

Given that the vast majority of adult-onset offenders were low-rate offenders, these offenders (currently processed in the adult court system) were compared with low-rate, early-onset offenders (typically diverted from the CJS). The results indicated that low-rate, adult-onset offenders and low-rate, early-onset offenders had similarly low rates of (predominantly less serious) offending. While there were differences in the types of crimes commonly perpetrated by adult-onset and early-onset, low-rate offenders, these differences seem to reflect the social behaviour, culture and developmental ‘struggles’ of each developmental period. For the younger low-rate offenders, their offences mirrored those typically reported for ‘adolescence-limited offenders’, including shoplifting and other property offences, public order offences and drug offences. For the adult-onset, low-rate offenders, their offences were often associated with fairly common social behaviour in emerging adulthood, including offences related to drinking and other substances and disturbances to public order, as well as resisting/hindering police and dangerous/negligent driving. Many of these offences are not surprising in a culture among a lot of young adults of binge drinking, nightclubbing and generally ‘partying’ (Druginfo 2009), as well as other forms of risk-taking behaviours (eg related to driving; Arnett 2000). Together, these findings suggest that similar to adolescence-limited patterns of offending, low-rate, adult-onset offending may be associated with psychosocial factors relevant to this developmental period. Interventions that target these psychosocial factors, such as risky drinking and driving behaviours, and difficulties transitioning to adulthood, may also reduce low-rate, adult-onset offending.

Despite slightly lower individual rates of offending, perpetrating somewhat less serious offences and costing less socially and economically, each adult-onset offender was more costly to the CJS than each early-onset offender. This is primarily due to the frequent use of cautioning with low-rate, early-onset offenders which is not available to most adult offenders in Queensland. If first-time, low-rate, less serious adult-onset offenders were formally cautioned, it would have saved $32.5m in police and court costs. This represents a 23.4 percent cost reduction in processing this group through the CJS and a 4.3 percent reduction in the cost of processing all members of the 83/84 cohort through the CJS (Allard et al. 2014). The reduced length of time that police spend on each caution (4.5 hours) compared with preparing for each court appearance (11 hours), would have saved police 121,199 hours or approximately eight full-time policing positions. Reducing the number of Magistrates Court appearances by 18,646 would have also reduced the annual workload of the Magistrates Court by approximately 1.2 percent (Magistrates Court of Queensland 2012). However, these figures only represent the reduced workloads associated with diverting less serious, first-time offences by low-rate, adult-onset offenders in this single cohort. The workload reductions would be substantially more if considered on a cross-sectional basis.

While it is questionable whether police cautioning could be used to respond to all of these first-time, adult-onset offenders, evidence does indicate that 85 percent of youth are cautioned the first time they have contact with the system (Allard et al. 2009). Therefore, considerable cost savings would result even if similar cautioning rates could be achieved for first-time, adult-onset offenders.

Table 3 Offending profiles of low-rate offenders – early-onset versus adult-onset offenders
Age of onset Malea Indigenous Australianb Age of onset Number of offencesc Number of eventsd Ever committed serious offencee Ever been diverted from the CJS Ever received non-supervised orderf Ever received supervised orderg Ever received suspended sentenceh Ever been imprisoned as an adulti
% % M SD M SD M SD % % % % % %
Low rate, early-onset 65.4 5.8 15.3 1.8 2.4 1.7 1.8 1.3 14.0 74.8 50.5 7.0 0.8 0.7
Low rate, adult-onset 77.7 5.9 21.5 2.2 1.8 1.3 1.4 0.9 8.3 0.0 95.9 6.1 1.7 1.5

a: χ2(1,N=33,871)=659.3,p<.001,φ=-.1, missing n = 92

b: χ2(1,N=33,963)=0.8,p=.778,φ=.00

c: t(25,229.1)=33.5,p=.001,d=0.40

d: t(23,588.1)=33.7,p=.001,d=0.36

e: χ2(1,N=33,963)=287.4,p=.001,φ=-.09

f: χ2(1,N=33,963)=9,622,p=.001,φ=.53

g: χ2(1,N=33,963)=12.0,p=.001,φ=-.02

h: χ2(1,N=33,963)=57.1,p=.001,φ=.04

i: χ2(1,N=33,963)=51.9,p=.001,φ=.04

Note: Diversions were primarily cautions. Conferencing was available in limited jurisdictions for youths in this cohort because conferencing was only operating in pilot mode until 2003. Time to offend varied across early-onset (10–25 years) and adult-onset offenders (18–25 years). No low rate offenders were sentenced to detention

Table 4 Percentage of offenders who had ever perpetrated each ANZSOC offence type across low-rate offenders – early-onset versus adult-onset offenders
Offence type ANZSOC codes Early onset, low-rate (%) Adult onset, low-rate (%) χ2 (df=1)a φ
Personal offences Homicide and related offences 0.1 0.1 0.0 .00
Acts intended to cause injury 10.0 5.9 193.4* -.08
Sexual assault and related offences 0.9 0.5 24.9* -.03
Dangerous or negligent acts endangering persons 12.0 24.3 800.2* .15
Abduction and related offences 0.3 0.3 0.1 .00
Robbery, extortion and related offences 0.7 0.2 39.3* -.03
Any personal offence 22.3 30.2 261.1* .09
Property offences Unlawful entry with intent/burglary, break and enter 11.2 1.8 1,348.8* -.20
Theft and related offences 53.3 15.5 5,505.3* -.40
Fraud, deception and related offences 3.7 4.8 25.2* .03
Property damage and environmental pollution 15.1 6.1 745.8* -.15
Any property offence 68.9 25.2 6,422.7* -.44
Drug offences Illicit drug offences 21.2 15.6 172.0* -.07
Public order offences Public order offences 24.5 40.4 926.2* .17
Other offences Offences against justice procedures 12.9 20.2 310.7* .10
Weapons and explosives offences 3.5 2.5 31.1* -.03
Miscellaneous offences 3.1 1.0 199.4* -.08

* p≤.001

a: Given high power for analyses, readers should interpret the results in conjunction with effect sizes. Traffic and breach offences excluded

Table 5 Cost of low rate offenders – early-onset versus adult-onset offenders
Age of onset Proportion of all low-rate offenders CJS costs Economic & social costs Total costs
N % M Group Costs ($mil) % M Group costs ($mil) % M Group costs ($mil) %
Early onset 14,149 41.7 5,808 82.2 37.2 4,642 65.7 55.8 10,449 147.8 43.7
Adult onset 19,814 58.3 6,993 138.6 62.8 2,628 52.1 44.2 9,621 190.6 56.3
Total 33,963 100.0 6,499 220.7 100.0 3,467 117.8 100.0 9,966 338.4 100.0

Additionally, cautioning first-time, low-rate, adult-onset offenders is consistent with the dominant model of offender rehabilitation (ie risk-needs-responsivity principles) and is likely to reduce recidivism, thereby further reducing costs. Given the brief and less serious nature of most adult-onset offending and since CJS intervention may actually increase the likelihood of low-risk offenders reoffending (Andrews & Dowden 2006), formal cautioning may be more commensurate with the risks and needs of the vast majority of adult-onset offenders.

The use of cautioning may be particularly relevant, or arguably necessary, for young adult offenders. Farrington, Loeber and Howell (2012) argue that processing young offenders (ie under 25 years of age) in the adult courts, which is more punitive than the youth justice system, increases their likelihood of reoffending. They assert that young adults should be dealt with by extending the youth justice system. In Queensland and other jurisdictions in Australia, this would entail extending formal police cautioning to young adult offenders.

While limited research has examined the impact of cautioning for adults, there is some evidence that cautioning adults may be associated with lower rates of reoffending (AONSW 2011; Office for Criminal Justice Reform 2010). For youth offenders, evidence indicates that diverting first-time youthful offenders reduces recidivism. Allard and colleagues (2009) found that young people who were cautioned for their first offence were 1.7 times less likely to have a second contact with the CJS than young people who appeared in court, after controlling for potential demographic and offence differences. If cautioning proved to be similarly effective for reducing recidivism among first-time, adult-onset offenders, there would be significant additional cost savings, as well as social benefits. For all of these reasons, broader adult cautioning schemes, or other adult diversion schemes, have been introduced overseas for less serious first offences (eg Ministry of Justice 2013). Additionally, drug diversion schemes, as well as other targeted diversion schemes have been implemented across Australia (NCPIC 2013).

Importantly, the findings from this study must be interpreted according to the limitations of this research. First, as this study relied on officially recorded offending, it is possible that adult-onset offenders had offended prior to 18 years of age but it was undetected. Although this is a major limitation to this study, these data are ideal for assessing system costs and impact. Nevertheless, findings should be replicated in studies using self-report data. Second, the 83/84 QLD does not control for attrition due to death or moving interstate that may inaccurately resemble desistence. Third, the 83/84 QLD does not control for migration into Queensland, which may have resulted in individuals with offending histories elsewhere appearing for the first time in the 83/84 QLD as adult-onset offenders.

It should be noted that data was only available to 25 years of age. There is increasing evidence that the ages between 18 and 25 years represent a unique developmental period, often called ‘emerging adulthood’ (Arnett 2000). The low-rate, adult-onset offending in this study appeared to reflect social factors associated with emerging adulthood. Therefore, research using older adult-onset offenders may produce different results. In addition, the progression of adult-onset offenders’ criminal careers beyond 25 years of age could not be assessed. Future research should examine the progression of different adult-onset offending patterns beyond 25 years, as well as the heterogeneity of offending that commences after 25 years of age.

The costs used in this study were based on a bottom-up costing approach and CJS costs were average opportunity costs. The use of average rather than marginal costs means that a particular reduction in crime may not result in the specified cost savings, because many costs are fixed. However, the cost savings are likely to result if there is increasing population and future costs are delayed or avoided.

That the CJS costs were estimated for Queensland only should also be noted. There is some variation between jurisdictions in the costs of criminal justice practices, as well as variations in responses to adult offending (eg differences in the use of infringement notices, cautions, forum sentencing, diversion programs). Additionally, offenders are processed in the adult courts from 17 years of age in Queensland. If adult cautioning was introduced, even more offenders would be eligible for cautions than was estimated in this study (ie 2,531 offenders saving an additional $4.5m; Thompson et al. 2013).

Crude eligibility criteria for cautioning were used in this study to estimate potential cost savings. In practice, more carefully selected criteria would be necessary.

Another study limitation was that as criminal trajectories cannot be determined a priori, 75 percent of high-rate, adult-onset offenders in this sample would have been ‘eligible’ for cautioning for their first CJS contact (as occurs with youths). However, there is no reason to believe that a non-supervised order (current practice) would be more effective than formal police cautioning.

Finally, it was not possible to examine the factors that lead to more serious adult-onset offending pathways in this study. Since little is known about these offenders, future research should investigate the factors associated with high-rate, adult-onset offending to inform prevention and intervention strategies.

Despite these limitations, three conclusions can be made from this study. First, there are a large proportion of offenders who do not come into contact with the CJS until 18 years or older. Second, for 95 percent of these offenders, their offending career is brief and less serious. Third, in line with best practice principles of risk-needs-responsivity, it may be more appropriate to respond to these offenders using diversionary schemes like formal adult cautioning. Doing so would save the CJS considerable resources that could be targeted towards more prolific offenders and/or used to address the social problems that lead to these forms of adult-onset offending.


All URLs correct at July 2014

  • Allard T, Chrzanowski A & Stewart A 2012. Targeting crime prevention to reduce offending: Identifying communities that generate chronic and costly offenders. Trends & Issues in Crime and Criminal Justice no. 445.
  • Allard T et al. 2014. The monetary cost of offender trajectories: Findings from Queensland (Australia). Australian and New Zealand Journal of Criminology 47: 81–101
  • Allard T et al. 2010. Police diversion of young offenders and Indigenous over-representation. Trends & Issues in Crime and Criminal Justice no. 390.
  • Allard T et al. 2009. The use and impact of police diversion for reducing Indigenous over-representation. Report for Criminological Research Council.
  • Andrews D & Dowden C 2006. Risk principle of case classification in correctional treatment: A meta-analytic investigation. International Journal of Offender Therapy and Comparative Criminology 50(1): 88–100
  • Arnett JJ 2000. Emerging adulthood: A theory of development from the late teens through the twenties. American Psychologist 55: 469–480
  • Australian Bureau of Statistics (ABS) 2011. Australian and New Zealand standard offence classification (ANZSOC), 3rd ed. ABS cat. no. 1234.0. Canberra: ABS
  • Australian Bureau of Statistics (ABS) 2009. National offence index. ABS cat. no. 1234.0.55.001. Canberra: ABS
  • Carrington PJ, Matarazzo A & deSouza P 2005. Court careers of a Canadian birth cohort. Crime and Justice Research Paper Series. Ottawa, Ontario: Canadian Centre for Justice Statistics, Statistics Canada
  • Chrzanowski A & Wallis R 2011. Understanding the juvenile justice system, in Stewart A, Allard T & Dennison S (eds), Evidence based policy and practice in youth justice. New South Wales: Federation Press: 7–27
  • DeLisi M & Piquero AR 2011. New frontiers in criminal careers research, 2000–2011: A state-of-the-art review. Journal of Criminal Justice 39: 289–301
  • DrugInfo 2009. All about binge drinking.
  • Eggleston EP & Laub JH 2002. The onset of adult offending: A neglected dimension of the criminal career. Journal of Criminal Justice 30: 603–622
  • Farrington DP, Loeber R & Howell JC 2012. Young adult offenders: The need for more effective legislative options and justice processing. Criminology and Public Policy 11: 729–750
  • Kratzer L & Hodgins S 1999. A typology of offenders: A test of Moffit’s theory among males and females from childhood to age 30. Criminal Behaviour and Mental Health 9: 57–73
  • Krohn MD, Gibson CL & Thornberry TP 2013. Under the protective bud the bloom awaits: A review of theory and research on adult-onset and late-blooming offenders, in Gibson CL & Krohn MD (eds), Handbook of life-course criminology. New York: Springer: 183–200
  • Little S & Allard T 2011. Responding to offending—Youth justice system responses, in Stewart A, Allard T & Dennison S (eds), Evidence based policy and practice in youth justice. NSW: Federation Press: 152–168
  • Magistrates Court of Queensland 2012. Magistrates Court of Queensland: Annual report 2011–2012. Queensland: Queensland Courts
  • Ministry of Justice 2013. Simple cautions for adult offenders. United Kingdom: Ministry of Justice.
  • NCPIC 2013. Cannabis and the law. Fact sheet no. 2.
  • Office for Criminal Justice Reform 2010. Initial findings from a review of the use of out-of-court disposals. London: Ministry of Justice
  • Piquero AR 2008. Taking stock of developmental trajectories of criminal activity over the life course, in Liberman AM (ed), The long view of crime: A synthesis of longitudinal research. New York: Springer: 23–78
  • Queensland Police Service (QPS) 2012. 2011–2012 annual statistical review. Brisbane: QPS.
  • Rollings M 2008. Counting the costs of crime in Australia: A 2005 update. Research and Public Policy series no. 91. Canberra: Australian Institute of Criminology.
  • The Audit Office for New South Wales (AONSW) 2011. The effectiveness of cautioning for minor cannabis offences. New South Wales: AONSW
  • Thompson C, Stewart A, Allard T & Chrzanowski A 2013. Understanding the extent and nature of adult-onset offending: Implications for the effective and efficient use of criminal justice and crime reduction resources. Report for Criminological Research Council
  • Victoria Police 2012. Crime statistics 2011/2012. Victoria: Victoria Police
  • Wolfgang M, Figlio RM & Sellin T 1972. Delinquency in a birth cohort. Chicago, IL: University of Chicago Press 

About the authors

Carleen Thompson, Anna Stewart, Troy Allard, April Chrzanowski, Jerneja Sveticic and Chelsea Luker are from the School of Criminology and Criminal Justice, Griffith University.


We acknowledge and thank all Queensland Government agencies who provided data for the project.