Foreword | Although there are numerous arguments for and against the use of suspended sentences, improving our knowledge of how this sentencing disposition is applied in practice will help inform the debate. This paper provides an overview of the use of suspended sentences in the Supreme Court of Tasmania, as well as an analysis of reconviction and breach rates for those placed on such an order. It was found that those offenders serving suspended sentences had the lowest reconviction rates compared to those who received non-custodial and unsuspended sentences and this held true irrespective of prior criminal history. Young offenders had particularly low reconviction rates post a suspended sentence, but the paper warns of the risk that those offenders who are reconvicted may attract inappropriate, more severe sentences, as those who had previously been given a suspended sentence were found to be more likely to receive a subsequent (and more serious) unsuspended sentence, regardless of offence severity. An additional finding was that only five to six percent of offenders who were in breach of a suspended sentence were returned to court for a breach action. The paper stresses the need to improve the management of breaches and for readily accessible and up-to-date sentencing information, as well as arguing that offence- or length-based restrictions on suspended sentences is unnecessary.
General Manager, Research
A suspended sentence is a prison sentence which is partly or wholly suspended on certain conditions. It has been described as a Sword of Damocles hanging over an offender's head (eg R v Locke and Paterson (1973) 6 SASR 298). As set out by the High Court in Dinsdale v The Queen (2000) 202 CLR 321, imposing a suspended sentence involves two key steps, namely, imposing a fixed term of imprisonment and then ordering that all or part of the term be held in suspense for a specified period ('the operational period'), subject to certain conditions. Suspended sentences are currently available in all Australian jurisdictions, although there have been moves to abolish them in Victoria (Bartels 2007).
The principal arguments in favour of suspended sentences are that they are an effective form of denunciation and deterrence; they are a valuable tool for those handing down sentences; they enable offenders to avoid prison, especially for short sentences; and they reduce the size of the prison population.
However, there are also compelling arguments against suspended sentences, namely, that they do not amount to real punishment at law and are regarded as a 'let-off' by the public and offenders; there are difficulties with the process for imposing the sentence and dealing with breaches; they cause net-widening and violate the proportionality principle; and they favour middle-class offenders.
As noted by the VSAC in its extensive review on this issue:
The philosophical differences between those who accept that a suspended sentence is more severe than other non-custodial orders and who believe it to be an appropriate substitute for immediate prison time, and those who question the internal logic, position and continued need for such an order are fundamental and unlikely ever to be satisfactorily resolved (2006: vii).
This paper does not endeavour to reconcile these arguments. Instead, it reports on some key findings of recently completed research on the use of suspended sentences in Tasmania in an attempt to promote a greater understanding of the use of this controversial sentencing disposition.
This paper presents key findings of a quantitative analysis of offenders sentenced in the Tasmanian Supreme Court and reconviction and breach analyses of those offenders.
There were three datasets for the sentencing, reconviction and breach analyses.
The data for the original dataset (DS1) were obtained by examining all cases where an offender was sentenced at first instance in the Supreme Court of Tasmania between 1 July 2002 and 30 June 2004 (n=838). Cases where the offender was dealt with for a breach, life redetermination or pursuant to a dangerous-offender application were excluded (n=24).
In this study, offenders' records were coded as nil, minor or significant on the basis of the sentencing judge's comment, which was in turn based on an assessment of the number and seriousness of prior offences. It is acknowledged that this approach is a subjective assessment and future research should seek to refine this measure.
The data for the reconviction dataset (DS2; n=588) were collected in January 2007 from the Tasmania Police Intrepid Centralised Enquiry (ICE) database of offenders' criminal records. The Department of Correctional Services provided information on the dates of release of offenders who served their sentence in custody. Offenders were included in DS2 if:
- they were 18 years or over at the time of their original sentence
- their original sentence was a non-custodial order or a custodial sentence of two years or less
- they had been released from custody by 1 July 2004.
Offenders were followed up for two years from the date of sentence (for wholly suspended sentences or non-custodial orders) or release from custody (unsuspended and partly suspended sentences). Pseudo-reconvictions were excluded from the analysis (see Bartels 2009a).
A multiple-logistic regression model was constructed to control for the main variables affecting reconviction (see Hosmer & Lemeshow 2000; O'Connell 2006). The model included the following explanatory variables: gender, age, prior record, offence type and seriousness, and sentencing judge and disposition. All of these variables except gender and sentencing judge indicated statistically significant bi-variate relationships with reconviction; the latter two variables were retained in the regression model because of their policy significance.
The data for the breach dataset (DS3; n=310) comprised offenders from DS2 who had received a partly suspended sentence or wholly suspended sentence.
To date, most breach analyses have only examined court action taken in respect of alleged breaches (eg Tait 1995), not instances where the suspended sentence was apparently breached but no court action was taken. In the present study, a breach was taken to have occurred where an offender had been convicted of an offence punishable by imprisonment that had been committed during the operational period of the sentence.
Offenders were followed up for between two and a half and four and a half years.
Results and discussion
Unsuspended sentences were the most commonly imposed disposition (44% of all sentences), followed by wholly suspended sentence (29%). Partly suspended sentence and non-custodial orders accounted for 13 percent and 14 percent of all sentences respectively.
Unlike most Australian jurisdictions, there is no legislative maximum in Tasmania on the length of sentence which can be suspended (Bartels 2007). It has been suggested that placing a limit on the length of sentence that can be suspended would prevent the use of such sentences 'in the most controversial cases' (TLRI 2008: [3.3.17]).
On the basis of the data obtained in this study, there appears to be little need to alter the current position: fewer than seven percent of wholly suspended sentence were longer than 12 months and there were no such sentences exceeding two years, with the same results for the suspended portion of partly suspended sentence. More recent data indicate that there were no wholly suspended sentence in the Supreme Court exceeding two years between 2001 and July 2007, with the suspended portion of a partly suspended sentence only exceeding two years in two cases (TLRI 2008:[3.3.27]). Based on the present findings, the TLRI opposed the introduction of length-based restrictions on the use of suspended sentences (2008: Rec 10).
Many Australian jurisdictions have legislative restrictions on the kinds of offences for which suspended sentences are available. In 2006, the Victorian Government enacted legislation to provide that a court may not impose a wholly suspended sentence for a serious offence unless satisfied that doing so is 'appropriate because of the existence of exceptional circumstances in the interests of justice' (Sentencing (Suspended Sentences) Act 2006 (Vic), s. 4(2)).
Between 1 July 2002 and 30 June 2004, there was one suspended sentence imposed in the Supreme Court of Tasmania for manslaughter and another for assisting another to commit suicide. There were also three partly suspended sentence and two wholly suspended sentence imposed for rape. The fact that suspended sentences appear to be only rarely imposed for very serious offences suggests there is no need for offence-based restrictions on the circumstances in which such sentences may be imposed. The TLRI reached a similar conclusion, stating that 'there is no evidence that wholly suspended sentences are being used inappropriately for serious crimes in Tasmania' (2008: [3.3.29]) and accordingly opposed the introduction of offence-based restrictions on the power to order suspended sentences (TLRI 2008: Rec 10).
Prior criminal record
Prior criminal record is generally regarded as a relevant factor in sentencing (see eg ALRC 2006: [6.51]). There were statistically significant differences in sentencing disposition on the basis of the seriousness of the offender's prior record (even after controlling for the offender's age and gender, the seriousness of the offence and the sentencing judge and the seriousness of the original offence (p<0.01)).
The likelihood of receiving a suspended sentence, especially a wholly suspended sentence, is inversely correlated with previous convictions, while unsuspended sentences are most likely for offenders with significant previous offending. Whereas only nine percent of offenders with a significant prior record received a wholly suspended sentence, this rose to 39 percent of first offenders and 41 percent of those with a minor record. Somewhat unexpectedly, however, first offenders were much more likely to receive a wholly suspended sentence (39%) than a non-custodial order (24%), which they were, in turn, only slightly more likely to receive than an unsuspended sentence (23%).
This finding does not appear to conform with the parsimony principle, which
operates to prevent the imposition of a sentence that is more severe than is necessary to achieve the purpose or purposes of the sentence (ALRC 2006: [5.9]).
The common law requirement that the court is only to impose a suspended sentence where imprisonment would be appropriate in absence of the power to suspend should be borne in mind in this context.
It would appear on the basis of these data that judges are treating wholly suspended sentence as a low-level form of non-custodial order. While judicial officers may perceive their actions as lenient, the result is a nominal imprisonment rate of 76 percent for first-time offenders. This would appear to represent little genuine attempt to use imprisonment as a sanction of last resort and suggests that net-widening may be occurring. This would not be a concern if the offending of first offenders was more serious than that of more seasoned offenders, but this was not the case: 74 percent of offences committed by first offenders were serious, compared with 87 percent for offenders with a minor record and 90 percent for offenders with a significant prior record.
Table 1 reveals another important aspect to this discussion. The offender's prior most severe sentence was classified as nil, non-custodial order, suspended sentence or custody in DS1. When a judge noted that an offender had previously served time in custody, it was not generally stated whether this was an unsuspended sentence or partly suspended sentence. References to a previous suspended sentence were, unless stated otherwise, assumed to be a wholly suspended sentence.
|Prior sentencing disposition|
|Custodial (n=256)||Suspended sentence (n=62)||Non-custodial order (n=55)||Nil (n=323)|
|Note: percentages may not sum to 100 due to rounding|
|Source: Bartels 2008|
|Full-time imprisonment (n=316)||79 (n=202)||53 (n=33)||13 (n=7)||23 (n=74)|
|Partly suspended sentence (n=86)||8 (n=20)||21 (n=13)||15 (n=8)||14 (n=45)|
|Wholly suspended sentence (n=193)||10 (n=26)||23 (n=14)||49 (n=27)||39 (n=126)|
|Non-custodial order (n=101)||3 (n=8)||3 (n=2)||24 (n=13)||24 (n=78)|
When sentences were analysed by the previous most severe sentence, there was a highly significant difference in the subsequent sentence received (p<.001). Put simply, a previous suspended sentence is a strong predictor of a subsequent unsuspended sentence. This makes sense in terms of an offender receiving an increasingly severe penalty to reflect the fact that they have failed to respond to previous more lenient dispositions. However, when coupled with the information above about wholly suspended sentence being imposed very widely as a penalty for first offenders, it leads to the disquieting inference that offenders may find themselves receiving a custodial sentence even for a minor offence on their second conviction (for comment see Roberts 2004: 29; Victorian Sentencing Advisory Council 2005: 21).
In particular, it is relevant to note that the likelihood of receiving an unsuspended sentence is more than four times higher for someone who has previously had a suspended sentence than someone with a previous non-custodial order (55% vs 13%). In addition, the chance of receiving a subsequent non-custodial order is equal for offenders who have previously served time in prison as for those who have had a suspended sentence (3%). Ironically, the chance of spending some time in prison, either by way of an unsuspended sentence or partly suspended sentence, is in fact higher for first offenders (37%) than for an offender who has previously received a non-custodial order (27%), while 74 percent of offenders before the court who are reported to have previously received a suspended sentence were subsequently sent to prison on an unsuspended sentence or partly suspended sentence.
These findings suggest that there may be an inappropriate approach being taken in respect of prior criminality. Offenders who have previously been given a suspended sentence may be treated as much more serious offenders than those with a previous non-custodial order, even when the former received a suspended sentence for their first offence. Caution should be exercised to ensure that offenders are not artificially elevated to the second-highest rung on the sentencing ladder, especially for minor or moderate offences.
Numerous studies have examined the relationship between sentencing disposition and reoffending (eg Spier 2001; Spohn & Holleran 2002; Tait 2001). In the present study, there was a highly significant difference in reconviction rates on the basis of sentencing disposition (χ2=17.4, df=3, p<.001). Somewhat unexpectedly, suspended sentences had the lowest reconviction rates: 42 percent of offenders on a wholly suspended sentence and 44 percent of those on a partly suspended sentence were reconvicted, compared with 52 percent of offenders on a non-custodial sentence and 62 percent of those in receipt of an unsuspended sentence (see Bartels 2009a). These statistically significant differences were maintained after controlling for gender, age, prior record, offence type and seriousness, and sentencing judge and disposition (p<.05) and suggest that sentencing disposition is a significant predictor of reconviction. Furthermore, of the offenders who were reconvicted, those on a wholly suspended sentence were the group most likely to be reconvicted only of a minor offence.
There was no statistically significant difference in reconviction rates between partly suspended sentence and wholly suspended sentence, even though one might expect partly suspended sentence (which are a form of custodial sentence served immediately) to have a similar reconviction rate to unsuspended sentences. It is always assumed that prison exacerbates criminality but these figures may suggest otherwise, with offenders on a partly suspended sentence having similar reconviction rates to those on a wholly suspended sentence. This may suggest that the very fact of suspension may influence reconviction rates, rather than whether an offender serves time in a carceral environment. However, there may be other factors involved. For example, judges may only select the 'best' cases for partial suspension, or the same factors which compelled the judge to suspend a sentence also have a protective influence against further offending.
Prior criminal record
It is widely acknowledged in the recidivism literature that prior criminal record is strongly associated with reconviction (eg see Makkai et al 2004; Ross & Guarnieri 1996) and this was confirmed by this study. Only 35 percent of offenders with no prior convictions at the time of the original sentence were reconvicted, compared with 52 percent of those with a minor record and 76 percent of those with a significant prior record. These differences were highly significant (p<.001) and remained so in the regression model.
Table 2 sets out the proportion of offenders reconvicted by prior record and sentencing disposition. This demonstrates that for offenders with no prior record, sentences where the offender is required to serve time in custody—unsuspended sentence and partly suspended sentence—were in fact less likely to result in reconviction, at 26 percent and 25 percent respectively, than wholly suspended sentence (32%) or non-custodial orders (47%). This finding contradicts conventional wisdom that a wholly suspended sentence is particularly appropriate for first offenders, in order to keep them from the supposedly corrupting influences of prison. Indeed, these findings may support the view that prison can function as an effective deterrent for a first-time offender.
|Note: percentages may not sum to 100 due to rounding|
|Source: Bartels 2009a|
|77 (n=97)||58 (n=31)||26 (n=35)|
|Partly suspended sentence|
|82 (n=17)||45 (n=29)||25 (n=32)|
|Wholly suspended sentence|
|68 (n=25)||49 (n=80)||32 (n=116)|
|75 (n=8)||63 (n=24)||47 (n=72)|
Offenders with a minor record, by contrast, appeared to perform comparatively better on suspended sentences, with 45 percent of those on a partly suspended sentence and 49 percent on wholly suspended sentence reconvicted, compared with 63 percent for non-custodial orders and 58 percent for unsuspended sentences. There was a poor outcome for all sentencing dispositions for offenders with a significant record, but those on a wholly suspended sentence in fact performed best of all, with 68 percent reconvicted.
Perhaps even more significantly, none of the offenders aged 55 years or over at the time of the original sentence were reconvicted, compared with almost two-thirds (66%) of the 18–24 years group. The differences in these rates were highly significant (χ2=78.2, df=4, p<.001) and the regression model supports this finding (χ2=33.1, df=4, p<.001).
As set out in Table 3, offenders aged 18–24 years performed much better on wholly suspended sentence (53% reconvicted) and partly suspended sentence (55%) than on unsuspended sentences (86%) or non-custodial orders (72%). For offenders aged 25–34 years, unsuspended sentences had a worse outcome (72%) but there was little difference between the other sentencing dispositions (46–48%). There were no statistically significant differences in reconviction outcomes for offenders aged 35 years and over (p>.05) and no offenders aged 55 years and over were reconvicted.
|18–24 yrs (n=242)||25–34 yrs (n=166)||35–44 yrs (n=94)||45–54 yrs (n=52)||55+ yrs (n=25)|
|Full-time imprisonment (n=170)||86 (n=65)||72 (n=53)||33 (n=30)||27 (n=11)||0 (n=11)|
|Partly suspended sentence (n=79)||55 (n=29)||48 (n=23)||38 (n=21)||0 (n=2)||0 (n=4)|
|Wholly suspended sentence (n=225)||53 (n=101)||46 (n=57)||32 (n=31)||17 (n=30)||0 (n=6)|
|Non-custodial order (n=105)||72 (n=47)||46 (n=33)||25 (n=12)||33 (n=9)||0 (n=4)|
|Note: percentages may not sum to 100 due to rounding|
|Source: Bartels 2008|
Section 27(1) of the Tasmanian Sentencing Act 1997 (the Act) provides that a suspended sentence may be breached by committing an imprisonable offence or breaching a condition of the suspended sentence during the operational period of the sentence. Where a breach appears to have occurred, the Director of Public Prosecutions (DPP) or a police or probation officer may apply to the court for an order, although, in practice, breaches of sentences imposed in the Supreme Court will be dealt with by the DPP. Section 27(4) provides that if the court is satisfied that the offender has breached the condition of the suspended sentence without reasonable excuse or committed a new offence, the court may order the suspended sentence to take effect, order a substituted sentence to take effect instead, or vary the conditions on which the execution of the sentence was suspended.
Breach rates and actioned cases
The findings reveal that 40 percent of offenders on a partly suspended sentence (32 out of 81) and 41 percent of offenders on a wholly suspended sentence (94 out of 229) breached their sentence by committing one or more offences punishable by imprisonment during the operational period of their sentence. However, breach action was taken in respect of only seven offenders: two on a partly suspended sentence and five on a wholly suspended sentence. This means that only six percent and five percent of offenders respectively who were in breach were brought back to court for breach action.
In the TLRI's Sentencing Issues Paper, it was noted that there are 'concerns that breach proceedings are neglected and many offenders breach without proceedings being initiated' (Warner 2002: 71). The data in this study confirm that there are indeed numerous instances of breached suspended sentences which are not coming to the attention of the relevant authorities.
Of the seven actioned cases, the sentence was activated in both partly suspended sentence cases and two out of the five wholly suspended sentence cases. The remaining three suspended sentences were continued. Accordingly, three percent of breached sentences were activated and only one percent of offenders in receipt of a suspended sentence were ultimately required to serve time in custody. Keeping offenders out of prison and reducing the size of the prison population are key arguments in favour of suspended sentences (Ancel 1971). Nevertheless, a failure to prosecute breached sentences not only reduces the effectiveness of the sentence for offenders (who may be encouraged to continue offending in the absence of breach action), but also potentially contributes to the negative perception of suspended sentences and may thereby undermine the criminal justice system generally.
The TLRI has described this study's findings as 'startling', noting that while
failure to initiate breach proceedings is not an inherent flaw of the suspended sentence, such a failure merely fuels the public perception that such sentences are an ineffectual slap on the wrist and contributes to a lack of confidence in sentencing (2008: [3.3.12]).
Frequency of offending
It would be understandable to see a lack of prosecutorial action on breaches if offenders committed very few offences in breach, but this is not the case. In fact, only a small proportion of offenders in breach committed one (8%) or two (5%) offences. The greatest proportion (27%) committed three to five offences, while 40 percent of offenders in breach committed 11 or more offences.
Only one of the seven actioned cases involved an offender who had committed over 20 offences. Breach action was taken against an offender who had committed 32 offences, including aggravated burglary and stealing, although the suspended sentence was ultimately continued with an extended operational period. No breach action was taken against the five most prolific offenders, who committed between 44 and 83 offences.
Prior criminal record
Unsurprisingly, there were significant differences in breach rates on the basis of the offender's prior criminal record. An offender with a significant prior record was over four times more likely to breach their partly suspended sentence than a first offender (82% vs 19%), while first offenders for both types of sentence were least likely to breach. These findings generally conform with those of Oatham and Simon (1972), who found prior offending strongly associated with breaching: 11 percent of male first offenders breached, compared with 71 percent for men with five or more previous convictions.
Only two of the offenders against whom breach action was taken had significant prior records.
This paper provides a quantitative overview of the use of suspended sentences in the Supreme Court of Tasmania.
The findings from the original dataset support the conclusion that there is no need for offence- or length-based restrictions on the availability of suspended sentences in Tasmania. The data also indicate that a first offender was more likely to receive a wholly suspended sentence than a non-custodial order. Furthermore, analysis of sentencing dispositions by the previous most severe sentence suggests that a previous suspended sentence is a strong predictor of a subsequent unsuspended sentence, making it of vital importance that first offenders are not pushed prematurely up the sentencing ladder.
The key finding of the reconviction analysis is that offenders on a wholly suspended sentence have the lowest reconviction rates of the four sentencing options studied, followed by partly suspended sentence. This suggests that for some offenders, suspended sentences do in fact 'work'. Analysis by prior record indicated that first offenders were in fact less likely to be reconvicted following a sentence requiring them to serve time in custody than on a wholly suspended sentence or non-custodial order. This finding contradicts the conventional wisdom that a wholly suspended sentence is particularly appropriate for first offenders (Bartels 2009a). Young offenders performed particularly well on a suspended sentence.
The principal finding of the breach analysis is that there is an overwhelming lack of action taken by the prosecuting authorities in Tasmania in respect of breached sentences. In fact, breach action is taken in only a very small proportion of cases, with numerous examples of repeat and serious offending going unprosecuted. This failure to deal appropriately with breached sentences not only reduces the effectiveness of the individual sentence, but may undermine confidence in the criminal justice system as a whole (Bartels 2009b). These findings prompted the TLRI to make several recommendations to improve the management of breaches in Tasmania, observing that:
Quite clearly a situation in which only five percent of breached orders result in proceedings is unacceptable. It makes a farce of the suspended sentence (2008: [3.3.40]).
The findings in this paper also highlight the need for maintaining readily accessible and up-to-date sentencing information on suspended sentences, so that judicial officers, policymakers, the media and the public are able to inform themselves about the use and utility of this controversial sentencing option.
- Ancel M 1971. Suspended sentence. London: Heinemann
- Australian Law Reform Commission (ALRC) 2006. Same crime, same time. Report 103. Canberra: ALRC
- Bartels L 2007. The use of suspended sentences in Australia: unsheathing the sword of Damocles. Criminal law journal 31: 113–131
- Bartels L 2008. Sword or feather? The use and utility of suspended sentences in Tasmania. Unpublished PhD thesis
- Bartels L 2009a. The weight of the sword of Damocles: a reconviction analysis of suspended sentences in Tasmania. Australian and New Zealand journal of criminology 42: 101–122
- Bartels L 2009b. Sword or butter knife? A breach analysis of suspended sentences in Tasmania. Current issues in criminal justice. Under revision
- Hosmer D & Lemeshow S 2000. Applied logistic regression, 2nd ed. New York: John Wiley & Sons
- Oatham E & Simon F 1972. Are suspended sentences working? New society 21: 233
- O'Connell A 2006. Logistic regression models for ordinal response variables. Thousand Oaks, CA: Sage Publications
- Spier P 2001. Recidivism patterns for people convicted in 1995. Wellington, New Zealand: Ministry of Justice
- Spohn C & Holleran D 2002. The effect of imprisonment on recidivism rates of felony offenders: a focus on drug offenders. Criminology 40: 329–358
- Tait D 1995. The invisible sanction: suspended sentences in Victoria 1985–1991. Australian and New Zealand journal of criminology 28: 143–162
- Tait D 2001. The effectiveness of criminal sanctions: a natural experiment. Report to the Criminology Research Council 33: 96–97. Canberra: Author
- Tasmania Law Reform Institute (TLRI) 2008. Sentencing. Final report no. 11. Hobart: TLRI
- Victorian Sentencing Advisory Council (VSAC) 2006. Suspended sentences final report—part 1. Melbourne: VSAC
- Victorian Sentencing Advisory Council (VSAC) 2005. Suspended sentences—discussion paper. Melbourne: VSAC
- Warner K 2002. Sentencing. Issues paper no. 2. Hobart: Tasmania Law Reform Institute
The author would like to acknowledge the University of Tasmania and the Australian Research Council in supporting the research reported in this paper.