Judging statisticsSENTENCING STATISTICS HELP COURTS ACHIEVE CONSISTENCY AND KEEP UP WITH COMMUNITY EXPECTATIONS, SO IT IS IMPORTANT TO KNOW WHERE TO FIND THEM AND HOW TO USE THEM. BY OCTAVIAN SIMU AND PAUL MCGORRERY

38 LAW INSTITUTE JOURNAL AUGUST 2021

featuresSentencing

HIGHLIGHTS

• Courts must consider current sentencing practices when deciding an appropriate sentence.

• Current sentencing practices are gauged from a consideration of both relevant sentencing statistics and sentences imposed in comparable cases.

• The Sentencing Advisory Council’s Sentencing Snapshots and SACStat products help practitioners and courts discharge this statutory obligation, each with different counting rules.

When sentencing, judges and magistrates must weigh up myriad often competing principles and considerations. One such principle is consistency, the obligation to try and ensure that relatively similar offenders receive relatively similar sentences.1 It is towards that goal that the Sentencing Act 1991 (Vic) requires courts to “have regard to . . . current sentencing practices” when sentencing an offender,2 though they must not become a “controlling factor”.3 The term “current sentencing practices” sometimes means sentencing principles generally (eg, that an offender’s deportation will usually be a mitigating factor).4 More often, though, the term refers to the types (and lengths) of sentences imposed for a certain offence.5

The question is – where are courts supposed to "nd information about how courts are currently sentencing a particular offence?

Comparable cases The answer to that question lies in a comment of the Court of Appeal in 2009:

“The identi"cation of current sentencing practices for an offence will usually require consideration both of relevant sentencing statistics for the offence and of sentencing decisions in comparable cases”.6

It is through this pairing of quantitative and qualitative information that courts can gain the most meaningful insight into current sentencing practices for an offence. On their own, sentencing statistics can be criticised for lacking in context.7 Similarly, cases submitted by the prosecution or defence can be criticised as either not being comparable or as cherry-picked8 (even if a table of cases is limited to those involving offending of a particular nature).9 But, together, statistics and case examples give a reliable but general indication of the range of sentences imposed, and where some cases have previously fallen along that range.10

Statistics are an aid to sentencersConsistency in sentencing does not necessarily mean achieving “numerical equivalence” between sentences;11 the quest for consistency must be balanced with (and give way to) individualised justice.12 In that sense, current sentencing practices do not indicate the “correct” sentence (nor does such a creature exist),13 nor do they set the “metes and bounds” of sentencing. Similarly, sentences imposed in other cases (no matter how similar) are not precedents that must be followed.14 Instead, statistics about recent sentencing practices operate, as the Court of Appeal has said on

multiple occasions, as a “rough cross-check”.15

For example, sentencing statistics might indicate:• that non-custodial sentences for rape –

when not attracting Category 1 offence classi"cation – are rare and only imposed in extraordinary circumstances16

• that the most common prison sentence lengths for armed robbery are between three and four years17

• what an “outlier” sentence would be in cases that do not call for an extraordinarily stern sentence18

• whether the sentencing limits in the Magistrates’ Court are suf"cient to justify the matter being heard and determined summarily.19

These insights from sentencing statistics arm judges with a useful guide or “yardstick” against which an appropriate sentence may be measured.20

Statistics can be especially useful when they distinguish certain key factors, such as the manner in which the offence was committed, or whether the offender pleaded guilty.21 While an appropriate sentence is rarely determined by any one factor, a breakdown of sentences for drug traf"cking by the quantity of drug traf"cked, for example, can certainly inform the question of offence seriousness, and in turn, be an “aid in sentencing”.22

Navigating the statisticsThe three primary sources of sentencing statistics in Victoria are Sentencing Snapshots, SACStat and in-depth research reports, all of which are produced by the Sentencing Advisory Council (SAC).

In the higher courts, the most favoured statistical products are Snapshots. The SAC currently publishes Snapshots on 25 of the most serious offences sentenced in the higher courts. A new Snapshot for each of these is published every two years. They include graphs, "gures and narratives about how many people were sentenced for each offence and how they were sentenced, including sentence types, charge-level prison sentence lengths, case-level prison sentence lengths and non-parole period lengths. Judicial references to Snapshots tend to range from a general indication that the judge has seen the relevant statistics (eg, “I have considered the statistics and current sentencing practices”)23 to detailed recitation of the statistics contained within.24

In the Magistrates’ Court, or for offences not covered in Snapshots, sentencing statistics are available via the SAC’s online statistical

AUGUST 2021 LAW INSTITUTE JOURNAL 39

Sentencingfeatures

ADO

BE S

TOCK

database, SACStat.25 This resource currently receives more than 60,000 views each year. SACStat presents charge-level data on sentence types and lengths for each offence (by jurisdiction), as well as case-level data on sentence types, sentence lengths and offender demographics in relation to every offence with 40 or more charges (Magistrates’ Court) or 10 or more charges (higher courts) in the most recent three or "ve year period.

Counting rules matterIt is important, however, to understand not just how to "nd sentencing statistics, but also how to interpret them. In some contexts (eg, Snapshots) the SAC uses what is known as the “principal proven offence” counting rule. This means that data presented in Snapshots illustrates how offenders were sentenced when the relevant offence received the most serious penalty in the case (typically, the longest period of imprisonment). This counting rule prevents the sentencing data from being skewed by more serious offences in the case. In other contexts (eg, SACStat) the SAC uses what is known as the “all proven offence” counting rule, whereby the data illustrates how every charge of an offence was sentenced.

As an example, the most recent Snapshot for theft shows that there were 132 people sentenced in the higher courts for theft as their principal offence in the "ve years to 2019–20, that 73.5 per cent received imprisonment, and that the median charge-level prison sentence length was one year and six months. In contrast, the most recent data on SACStat shows that there were a total of 3437 charges of theft sentenced in the higher courts in the "ve years to 2018–19, that 81.6 per cent received a prison sentence, and that the median charge-level prison sentence length was six months. That is, using the all proven offence counting rule in SACStat instead of the principal proven offence counting rule in Snapshots yields 26 times as many charges, a higher imprisonment rate, and much shorter prison sentence lengths. This suggests that most theft charges are heard in the higher courts because of more serious offences in the case, offences that drove both the jurisdiction and the type of sentence imposed on the case as a whole.

In effect, Snapshots are most useful for illustrating current sentencing practices in the higher courts when the relevant offence is the principal proven offence. In all other contexts, SACStat will typically be more useful.

Manifest excess/inadequacyIn addition to their role in informing the sentencing discretion at "rst instance, sentencing statistics can play a key role in manifest excess/inadequacy sentence appeals. For instance, if a particular offence receives a sentence that is a statistical outlier even though the offending does not necessarily fall into the “worst case” category, this will not be determinative of the appeal, but it can suggest that the sentence warrants closer scrutiny.26 Where the Court of Appeal is of the view that the circumstances of the case did not call for an outlier sentence, the Court will often refer to statistics to support its conclusion that the sentence was manifestly excessive.27 On the other hand, where the Court "nds that the circumstances did justify an outlier sentence, the relevance of sentencing statistics will give way to other factors.28

Beyond speci"c cases, the Court of Appeal has also utilised sentencing statistics to develop a unique line of case law in Australia whereby it declares current sentencing practices for certain offences (such as recklessly causing serious injury by glassing29 and rape by digital penetration30) to be inadequate and requiring an “uplift”. On each occasion, the Court has referred to statistics and cases to ascertain current sentencing practices before making such a declaration. For example, in Hogarth v The Queen, the Court referred to a recent SAC report on sentencing for aggravated burglary31 and said:

“We do not consider that current sentencing for confrontational aggravated burglary re-ects the objective seriousness of this form of the offence. The clustering of sentences around a median of two years shows how far current sentencing has departed from the parameters set by the maximum penalty of 25 years”.32

ConclusionSentencing statistics play a critical function for courts in Victoria (above and beyond their additional value to media, researchers, policy makers, students and the general community). They help courts ful"l their legislative obligation to consider current sentencing practices by providing them with another yardstick, such as the maximum penalty or the standard sentence, against which they can determine an appropriate sentence. To that end, practitioners should be in a position to assist courts in identifying and understanding the most relevant statistics. In turn, there is scope for greater consistency and transparency in the way

ADOBE STOCK

Sentencingfeatures

40 LAW INSTITUTE JOURNAL AUGUST 2021

statistics are relied on in discharging that obligation. Judges are not expected to be statisticians, nor are sentencing statistics ever going to be determinative, precedent-setting or controlling. Rather, as the Court of Appeal has described, they can be a rough but helpful cross-check.

To learn more about "nding and using sentencing statistics, the SAC has developed a free, one-hour online or in-person training session. To book, call the SAC on 1300 363 196 or email contact@sentencingcouncil.vic.gov.au ■

Octavian Simu is a legal policy officer at the Sentencing Advisory Council. Paul McGorrery is the manager of legal policy and community engagement at the Sentencing Advisory Council.

1. Wong v The Queen [2001] HCA 64, at [7]; Chris Maxwell, “Discretionary power and consistency: Is the sentencing discretion different?”, as in Jason NE Varuhas and Shona Wilson Stark (eds), The Frontiers of Public Law (Bloomsbury, 2020).

2. Sentencing Act 1991 (Vic), s5(2)(b).3. DPP v Dalgliesh (a pseudonym) [2017] HCA 41, at [9] and [68].4. Sentencing Advisory Council, Deportation and Sentencing: An Emerging

Area of Jurisprudence (2019).5. Sentencing Advisory Council, Sentencing Guidance in Victoria: Report (2016) 27.6. DPP v CPD [2009] VSCA 114, at [78] (emphasis added).7. See, eg, Kamal v The Queen [2021] VSCA 27, at [69].8. See, eg, DPP v Stevens [2020] VCC 1677, at [127].9. Nguyen v The Queen [2017] VSCA 286, at [46]-[47]; DPP (Cth) v Thomas [2016]

VSCA 237, at [171]-[191].10. See, eg, DPP v Nguyen [2021] VCC 337, at [53]-[54].11. The Queen v Pham [2015] HCA 39, at [28]; Hili v The Queen [2010] HCA 45, at [48].

12. Arie Freiberg and Sarah Krasnostein, “Statistics, damn statistics and sentencing” (2011) 21(2) Journal of Judicial Administration 73, 73-74; DPP v Dalgliesh (a pseudonym) [2017] HCA 41, at [49].

13. See, eg, DPP v Watson [2021] VCC 133, at [75].14. Hardwick (a pseudonym) v The Queen [2021] VSCA 67, at [43]; DPP v Page [2019]

VCC 618, at [49].15. Short v The Queen [2016] VSCA 210, at [59]; Russell v The Queen [2011] VSCA 147. 16. DPP v Worthington (a pseudonym) [2020] VCC 670, at [20]; Sentencing Advisory Council,

Sentencing Sex Offences in Victoria: An Analysis of Three Sentencing Reforms (2021).17. DPP v Wetzel [2020] VCC 1934, at [19].18. DPP v King [2019] VCC 1801, at [49].19. DPP v Zakariya & Anor [2020] VCC 603, [62]; Criminal Procedure Act 2009 (Vic) s29(2)(b).20. Nguyen v The Queen [2016] VSCA 198, at [83].21. Neither Snapshots nor SACStat currently present guilty plea rates, though the

SAC hopes to incorporate this data in the future.22. The Queen v Pham [2015] HCA 39, at [35]-[37] and [45].23. DPP v Cream [2021] VCC 366, at [40]; DPP v Cahill & Anor [2017] VCC 749, at [79].24. R v Lucas [2021] VSC 81, at [212].25. Sentencing Advisory Council, “SACStat: Sentencing Advisory Council Statistics”.

https://www.sentencingcouncil.vic.gov.au/sacstat/home.html.26. See, eg, Said v The Queen [2020] VSCA 178, at [75].27. See, eg, Traeger (a pseudonym) v The Queen [2019] VSCA 231, at [29]-[31].28. Said v The Queen [2020] VSCA 178, at [79].29. Winch v The Queen [2010] VSCA 141.30. Shrestha v The Queen [2017] VSCA 364.31. Sentencing Advisory Council, Aggravated Burglary: Current Sentencing Practices (2011).32. Hogarth v The Queen [2012] VSCA 302, at [58].

AUGUST 2021 LAW INSTITUTE JOURNAL 41

Sentencingfeatures