THE QUEEN v PHAM [2015] HCA 39

The use of appellate cases and statistics in sentencing

 

In The Queen v Pham [2015] HCA 39, the High Court allowed an appeal by the DPP (Cth). The Court set aside orders of the Victorian Court of Appeal (Pham v The Queen [2014] VSCA 204) and remitted the matter to that Court for determination. The Victorian Court of Appeal had allowed a severity appeal and reduced the sentence of the offender from 8 years 6 months (npp 6 years) to 6 years (npp 4 years). He had pleaded guilty to a charge of importing a marketable quantity of a border controlled drug (577.1 grams of heroin) contrary to s 307.2(1) Criminal Code (Cth).

The High Court (French CJ, Keane and Nettle JJ in a plurality judgment and Bell and Gageler JJ in a joint judgment) unanimously held that it is implicit in Pt IB Crimes Act 1914 (Cth) that a court sentencing for a Commonwealth offence must have regard to sentences that have been imposed in other States and Territories: [23], [41]. The Court of Appeal erred in law by deciding that the offender should be sentenced according to current Victorian sentencing practices to the exclusion of sentencing practices throughout the Commonwealth: [30], [41].

The High Court also unanimously upheld a second ground of appeal (see [38], [54]) that the Court of Appeal erred in law by “adopting an impermissible statistical analysis of comparable cases to determine the objective seriousness of the subject offence”: [3], [43]. However, there was a difference of opinion between the plurality and the joint judgment as to why the second ground should be upheld. The disagreement centred upon the use of statistics generally and the scope of the decisions of Hili v The Queen (2010) 242 CLR 520 and Barbaro v The Queen (2014) 253 CLR 58.

The Court of Appeal’s impermissible statistical analysis

Maxwell P attached to his judgment a table of 32 cases of intermediate appellate courts (Attachment “A”) for offences involving a marketable quantity of a border controlled drug where the offender was a “courier (or recipient), and no more”, had pleaded guilty and had “no (or no relevant) prior convictions”. A column in the table expressed the quantity imported as a percentage of the commercial quantity for each of the different drug types. The cases were ranked from the highest percentage to the lowest.

Maxwell P also attached a graph (Attachment “B”) plotting the head sentences of each of the 32 cases and the quantity of the drug imported as a percentage of the commercial quantity. Victorian cases were separated from all other jurisdictions. Maxwell P said the graph showed sentences in NSW, Queensland and Western Australia were substantially higher than Victoria and that the sentence imposed in this case was “well outside the range indicated by Victorian practice”: [14].

French CJ, Keane and Nettle JJ

The plurality held that “to prefer one State’s sentencing practices to sentencing practices elsewhere in the Commonwealth … is contrary to principle, tends to exacerbate inconsistency and so ultimately is unfair”: [27]. The plurality set out at [28] a non-exhaustive list of propositions concerning “the way in which the assessment of sentences in other cases is to be approached” [footnotes excluded]:

It is to be noted that no reference was made by the plurality to the statement in Barbaro at [41] that a court can synthesise raw material like statistics.

The plurality observed that intermediate appellate courts must have regard to sentencing decisions of other intermediate appellate courts in comparable cases as “yardsticks” that may serve to illustrate (although not define) the possible range available: [29]. Further, a court must have regard to such a decision unless the objective or subjective circumstances of the case make it distinguishable, or if the court thinks the outcome is manifestly inadequate or excessive: [29].

The plurality said this case illustrates the inutility of the presentation of sentences imposed on federal offenders using numerical tables, bar charts and graphs referred to in Hili (at [48]): [32], [33]. The statistical analysis was also flawed by treating the weight of drug imported as “the only variable factor affecting offence seriousness” and assuming that “courier” status was of uniform significance. These were both departures from fundamental principle and could not be used to indicate error in the first instance sentence: [37].

Bell and Gageler JJ

The joint Justices held that even if the Court of Appeal misused the table of 32 cases to determine the objective seriousness of the offence it does not demonstrate that presentation of this type of material is impermissible: [45]. Hili and Barbaro are concerned not only with the consistent application of sentencing principles but also with reasonable consistency in sentencing outcomes: [42], [46]. In Hili, the court said it is not useful to use statistical material which only refers to the lengths of sentences passed because it says nothing about why sentences were fixed: [46].

Statistical material showing the pattern of past sentences for an offence may serve as a yardstick by which the sentencer assesses a proposed sentence and the appellate court assesses a challenge of manifest inadequacy or excess: [47]. In Barbaro, the court held that judges must have regard to past cases: they may establish a range; the history stands as a yardstick to examine a sentence but it does not define the outer boundary of the permissible discretion. It was accepted that comparable cases and sentencing statistics are aids and part of the material which the sentencer must take into account: [48]. The Commonwealth Sentencing Database is a source of potentially relevant information about the pattern of sentencing for federal offences: [49].

Bell and Gageler JJ said at [49]:

Comparable cases decided by intermediate courts of appeal provide the most useful guidance to a sentencing judge. An appellate court’s reasons reveal the mix of factors that were taken into account and will usually involve consideration of the appropriateness of the sentence imposed at first instance: [50].

A deficiency in the table used in this case was the exclusion of relevant prior convictions. The offender had a number of prior convictions and had served prison sentences. The table was, therefore, a yardstick of limited utility by which to gauge the offender’s sentence: [51]. It was an error for Maxwell P to reason that in the case of couriers the only factor affecting the seriousness of the offence was the quantity of the drug imported: [52]. Osborn JA also wrongly treated the pattern of past sentences in the table as defining the boundaries of the proper exercise of the sentencing judge’s discretion: [54]. And Osborn JA’s observation that the sentence was “very heavy” fell short of a conclusion that it was manifestly excessive. This conclusion is required by House v The King (1936) 55 CLR 499 at 505 before the Court of Appeal could intervene: [56].

The sections in the Sentencing Bench Book including [10-020] Consistency, [10-022] Use of information about sentences in other cases and [10-024] Use of sentencing statistics — Hili v The Queen will be amended to reflect the contents of this Special Bulletin.