respondent. Special leave to appeal granted on grounds one to six inclusive in each matter (refused on ground seven in Jones); appeals treated as instituted and heard instanter and dismissed in each matter.
Key principles
There neither is, nor should be, any judicially determined norm or starting point, whether expressed as a percentage or otherwise, for the period of imprisonment that a federal...
Consistency in federal sentencing is achieved through the consistent application of the relevant statutory provisions in Pt IB of the Crimes Act 1914 (Cth), particularly s 16A(1)...
References to a 'norm' of 60-66 per cent for the period to be served before release are apt to mislead because they distract from the applicable statutory provisions, suggest a...
Appellate intervention on the ground of manifest inadequacy is warranted only where the sentence is unreasonable or plainly unjust such that it may be inferred that there has...
Issues before the court
Whether there exists, or should exist, a judicially determined 'norm' or starting point (expressed as a percentage) for the period of imprisonment a...
Plain English Summary
Two men who ran a tax evasion scheme using fake offshore loans to claim false deductions were given short prison sentences with early release. Prosecutors successfully appealed, saying the sentences were far too light. The NSW Court of Criminal Appeal increased the sentences but said there is usually a 'norm' that offenders serve 60-66% of their sentence before release. The High Court said there is no such fixed percentage or starting point in federal sentencing law. Judges must look at all the circumstances of the particular crime and offender under the Commonwealth Crimes Act and decide what sentence is appropriate. Numerical rules or charts of past cases cannot replace that. The original sentences were indeed too lenient given the serious, planned fraud over years that hurt the whole community, so the increased sentences stood.
AI-generated legal information, not legal advice. Zoe can make mistakes — check the cited source, and for advice about your situation consult a qualified Australian lawyer.
Deep Dive
2,905 words · generated 24/04/2026
Whether the sentences imposed at first instance were manifestly inadequate and whether the Court of Criminal Appeal gave adequate reasons for so...
How consistency in the sentencing of federal offenders is to be achieved under s 68 of the Judiciary Act 1903 (Cth) and Pt IB of the Crimes Act.
Cited legislation
5 cited instruments linked from this judgment.
What happened
Anthony Joseph Luis Hili and Glyn Morgan Jones were friends who operated painting and carpentry contracting businesses. In 1997 they formed a company to undertake work for the New South Wales Department of Housing. Through their accountants in Burwood they were introduced to a tax evasion scheme that used an international round-robin of transactions. False deductible expenses were claimed by companies they controlled for the years ended 30 June 2000 to 2004. The amounts were cycled through Vanuatu-based entities and returned to the applicants disguised as loans from a foreign lender. Hili engaged in 15 such transactions between January and November 2002 and lodged false returns for two companies and himself. Jones engaged in 11 transactions between April 2001 and February 2004 and lodged false returns for his company and himself.
Hili and his companies evaded tax of $398,537.82 and became liable for over $1 million once penalties and interest were added. Jones and his companies evaded $362,925.24 and faced a similar total liability. Both made partial repayments and were realising assets at the time of sentence. Each was previously of good character, pleaded guilty at the earliest opportunity, cooperated fully with police, and gave enforceable undertakings under s 21E of the Crimes Act 1914 (Cth) to assist in prosecuting four named promoters of the scheme. Morgan DCJ allowed a 50 per cent discount for plea and assistance (12.5 per cent referable to future assistance). Hili received 18 months' imprisonment with a recognizance release order after seven months. Jones received concurrent 18-month sentences on each of three counts (defrauding the Commonwealth contrary to s 29D, obtaining a financial advantage by deception contrary to s 134.2(1) of the Criminal Code, and money laundering contrary to s 400.4(1)) with release after seven months.
The Commonwealth Director of Public Prosecutions appealed to the New South Wales Court of Criminal Appeal on the ground of manifest inadequacy. That Court (McClellan CJ at CL, Howie and Rothman JJ) allowed the appeal in respect of Hili's sentence and two of Jones's counts. It re-sentenced Hili to three years with release after 18 months and adjusted Jones's sentences to produce an effective three years with release after 18 months. No order was made on Jones's money-laundering sentence, raising a question whether the original recognizance release order remained intact. The Court of Criminal Appeal referred to a "norm" for mandatory imprisonment under Commonwealth legislation of 60 to 66 per cent of the head sentence, affected by special circumstances, citing earlier authorities including Bernier, Viana, CAK & CAL and Ruha. It nevertheless fixed a 50 per cent pre-release period on the basis of special circumstances including first time in custody, good rehabilitation prospects and need for community reintegration.
Hili and Jones sought special leave to appeal to the High Court on three main questions: the existence of a norm, the sufficiency of the Court of Criminal Appeal's reasons on manifest inadequacy, and the status of the recognizance release order on the money-laundering count. The High Court (French CJ, Gummow, Hayne, Crennan, Kiefel and Bell JJ, Heydon J agreeing in the result but with additional observations) granted special leave on grounds one to six, refused it on ground seven, treated the appeals as instituted and heard instanter, and dismissed them.
Why the court decided this way
The joint judgment began from the fundamental proposition that sentencing federal offenders is governed by Pt IB of the Crimes Act 1914 (Cth), which makes exhaustive provision for non-parole periods and recognizance release orders. Section 68 of the Judiciary Act 1903 (Cth) picks up State laws only to the extent that Pt IB does not otherwise provide. Division 4 of Pt IB (ss 19AB-19AK) therefore displaces State or Territory provisions on fixing periods to be served before release. Section 16A(1) requires the court to "impose a sentence or make an order that is of a severity appropriate in all the circumstances of the offence". Section 16A(2) lists mandatory considerations if relevant. The provision accommodates common-law principles such as totality (Mill v The Queen) because they give content to the statutory phrase, but it forbids generalisations across federal offences that would depart from the requirement to sentence for the particular offence.
A recognizance release order under s 20(1)(b) may take effect at any point during the sentence. For sentences not exceeding three years, s 19AC(1) requires such an order unless it is inappropriate. No statutory ratio is prescribed between head sentence and pre-release period except for terrorism and like offences under s 19AG. The Court of Criminal Appeal's reasons made no reference to these provisions. Instead it endorsed the "norm" of 60-66 per cent drawn from State and Territory intermediate appellate decisions. The High Court held this approach apt to mislead for several reasons. First, "norm" is ambiguous: it might be prescriptive or descriptive, yet no statutory root for a prescriptive rule exists. Second, it distracts attention from the obligation to assess severity appropriate to all circumstances of the particular offence. Third, it suggests the same proportion should apply to fraud, narcotics importation or any other federal offence. Fourth, it implies that factors such as prior good character or cooperation can affect the proportion only within a narrow band or upon demonstration of "special circumstances", inverting the proper sequence.
The joint judgment approved the reasoning in Ruha that all matters in s 16A(2) must be considered in fixing both the head sentence and the pre-release period. The period that justice requires the offender to serve in custody must be determined by reference to the principles in Power v The Queen, Deakin v The Queen and Bugmy v The Queen. Deterrent and punitive effects for serious tax fraud must be reflected in both components. A mechanistic or formulaic approach that narrows the statutory range for recognizance release orders is therefore wrong. The Court cited Hunt J in Paull, the Court of Criminal Appeal in Bernier and Viana, and the Western Australian Court of Appeal in Bertilone to the same effect.
On manifest inadequacy the Court applied House v The King. The Director had alleged no specific error, only that the result was unreasonable or plainly unjust. The Court of Criminal Appeal was entitled to reach that conclusion if, having regard to all relevant matters, the sentences were outside the range. The offending was sustained, planned and deceitful, motivated by greed, involved substantial sums, and was difficult to detect. It affected the whole community. Comparable cases (Wheatley, where a public figure received two and a half years with release after 15 months for $318,000 evasion, and Ly, where a tax agent received six years with a four-year non-parole period for $329,000) demonstrated that the original sentences were markedly lower. The difference was so large that error could be inferred even though its precise location was not apparent. The re-sentenced terms of three years with release after 18 months were within the proper range. The reasons, though not ideal, sufficiently revealed why the first-instance sentences were inadequate by identifying the circumstances of the offending, the offenders' backgrounds and the yardstick provided by comparable cases.
Heydon J agreed in the result but emphasised that while legal principles must be applied consistently, the discretionary judgment in sentencing permits a range of legitimate outcomes. Two judges applying identical principles to identical facts may legitimately differ without either being wrong. Precedent does not convert the discretionary exercise into a search for a single correct answer.
Before and after state of the law
Before Hili intermediate appellate courts in New South Wales, Queensland and elsewhere had repeatedly referred to a "norm" or usual proportion of 60 to 66 per cent for non-parole periods or pre-release periods in federal sentencing. Bernier (1998) 102 A Crim R 44 stated that this Court had "usually" thought the proportion ought to be somewhere between 60 and 66 per cent. Viana [2001] NSWCCA 171 repeated that statement. CAK & CAL [2009] QCA 23 and Ruha [2010] QCA 10 treated the range as the "norm", departure from which required unusual factors or special circumstances. Sentencing judges and counsel routinely argued by reference to that range. The practical effect was that the statutory language of s 16A(1) and Div 4 of Pt IB was sometimes subordinated to a judicially created rule of thumb.
After Hili that practice could no longer be followed. The High Court made clear that any reference to a "norm" is at best ambiguous and at worst misleading. It has no statutory root. It distracts from the obligation to impose a sentence of appropriate severity in all the circumstances of the particular offence. The proportions observed in past cases are the result of the application of conventional principles to those cases; they cannot dictate the outcome in a new case. Sentencing judges must begin with the text of Pt IB, apply the mandatory considerations in s 16A(2), determine the head sentence, then fix the pre-release period by asking what period justice requires the offender to serve in custody having regard to deterrence, punishment, rehabilitation and all other relevant matters. Comparable sentences remain useful only as a yardstick when accompanied by an articulation of the unifying principles they reveal. The decision therefore restored primacy to the statutory scheme enacted by the Commonwealth Parliament and to the common-law principles that inform the statutory language.
The judgment also clarified the respective roles of first-instance judges and intermediate appellate courts in achieving consistency. Numerical tables or graphs are unhelpful. Consistency is "reasonable consistency" in the application of principle across Australia. Intermediate appellate courts should not depart from each other's interpretations of Commonwealth legislation unless convinced the earlier view is plainly wrong. Appellate review on manifest excess or inadequacy remains available but does not require the appellate court to identify specific error when the result itself bespeaks error.
Key passages with plain-English translation
"[3] The first question should be answered 'no'. There neither is, nor should be, a judicially determined norm or starting point (whether expressed as a percentage of the head sentence, or otherwise) for the period of imprisonment that a federal offender should actually serve in prison before release on a recognizance release order."
Plain English: Stop using any fixed percentage as a starting point for deciding how long someone must stay in gaol before early release on a Commonwealth sentence. The law does not allow it.
"[2] … the consistency that is sought is consistency in the application of the relevant legal principles, not some numerical or mathematical equivalence. Consistency in sentencing federal offenders is achieved by the proper application of the relevant statutory provisions, having proper regard not just to what has been done in other cases but why it was done, and by the work of the intermediate courts of appeal."
Plain English: Consistency does not mean every tax cheat gets exactly the same percentage of time behind bars. It means every judge must follow the same legal rules, look at why earlier judges sentenced as they did, and explain their own reasoning. Appeal courts keep the system honest.
"[27] … ss 16A(1) and (2) 'make it plain that all of the circumstances, including the matters in the non-inclusive list in s 16A(2), must be taken into account in making recognizance release orders just as they must be taken into account in imposing a sentence of imprisonment'." (quoting Ruha)
Plain English: The judge must look at every relevant fact twice—once for the total sentence length and again for the early-release date. There is no shortcut or formula.
"[47] … 'Sentencing patterns are, of course, of considerable significance in that they result from the application of the accumulated experience and wisdom of first instance judges and of appellate courts.' But the range of sentences that have been imposed in the past does not fix 'the boundaries within which future judges must, or even ought, to sentence'. Past sentences 'are no more than historical statements of what has happened in the past. They can, and should, provide guidance to sentencing judges, and to appellate courts, and stand as a yardstick against which to examine a proposed sentence'." (approving Simpson J in De La Rosa)
Plain English: Old cases are useful because experienced judges have already thought hard about similar crimes. But they do not set rigid minimum or maximum numbers that later judges must obey. Think of them as a ruler, not a cage.
"[58] Intervention is warranted only where the difference is such that, in all the circumstances, the appellate court concludes that there must have been some misapplication of principle, even though where and how is not apparent from the statement of reasons." (quoting Wong)
Plain English: An appeal court can increase a sentence that looks far too light even if it cannot pinpoint the exact legal mistake the first judge made. The overall result can itself prove the mistake.
What fact patterns trigger this precedent
Hili is triggered whenever a federal offender is to be sentenced or re-sentenced and a party invites the court to adopt a percentage "norm" or starting point for the pre-release period under a recognizance release order or non-parole period. It applies to all federal offences, not merely tax fraud. The decision is engaged whenever counsel or a judge refers to the 60-66 per cent range derived from Bernier, Viana, CAK & CAL or Ruha. It is also engaged when manifest inadequacy is alleged on a Crown appeal and the appellate court must decide whether the first-instance sentence is unreasonable or plainly unjust having regard to the circumstances of the offence, the offender, and comparable cases.
Typical triggers include: sustained course of deliberate dishonesty against the Commonwealth (tax, customs, social security or grant fraud); large amounts of revenue lost; offending that is difficult to detect; offenders who have otherwise been of good character but acted from greed rather than need; early pleas coupled with cooperation and undertakings under s 21E; and any submission that seeks to narrow the statutory discretion by reference to historical sentencing patterns without articulating the underlying principles. The precedent also applies when an intermediate appellate court is asked to depart from another intermediate appellate court's interpretation of Pt IB of the Crimes Act. Because the reasoning rests on the exhaustive character of Div 4 and the text of s 16A, it governs every federal sentencing exercise where a recognizance release order or non-parole period must be fixed.
How later courts have treated it
The source judgment itself records that the Court of Criminal Appeal's reasons, although ultimately reaching the correct result on manifest inadequacy, stated applicable principles "in a way which, at the very least, is apt to mislead" by endorsing the 60-66 per cent norm. The High Court expressly disapproved that statement at [39] and [23]. It approved instead the fuller analysis in Ruha that rejected a mechanistic or formulaic approach and insisted on beginning with the statutory text. The joint judgment cited with approval the earlier observations of Hunt J in Paull, the Court of Criminal Appeal in Bernier and Viana, and the Western Australian Court of Appeal in Bertilone that set ratios mask consideration of individual facts.
The decision reinforced the binding status of Wong on the impermissibility of guidelines that distract from s 16A(2) and of House v The King on the two categories of appellate intervention. It confirmed the continuing authority of Power, Deakin and Bugmy as supplying the content of "severity appropriate in all the circumstances". The judgment also affirmed the principle in Australian Securities Commission v Marlborough Gold Mines and Farah Constructions that intermediate appellate courts should not depart from each other's interpretations of Commonwealth legislation unless convinced the earlier view is plainly wrong. Subsequent treatment within the source text therefore treats Hili as restating and clarifying existing principle rather than breaking new ground, while correcting a line of intermediate authority that had hardened a descriptive observation into a prescriptive norm.
Still-open questions
The joint judgment left open whether the Court of Criminal Appeal's view that charging money laundering on facts already necessary to prove the fraud and deception counts raised "serious issues relating to double jeopardy" was correct. That question was not argued and the Court expressed no view. It also left open the precise mechanism for correcting any error in the recognizance release order on Jones's money-laundering count, noting that s 19AH(1) permits the Attorney-General, Director or the offender to apply at any time to set aside an improperly made order and make a proper one. The Court declined to deal with the point because no such application had been made to the Court of Criminal Appeal.
Heydon J's separate observations leave open the precise boundary between legal principle (which must be applied uniformly unless plainly wrong) and discretionary judgment (which permits legitimate variation even on indistinguishable facts). The joint judgment did not engage with that distinction. The decision does not address how sentencing courts should quantify the value of future assistance undertakings under s 21E when the assistance has not yet been given, nor the interaction between Commonwealth and State sentencing regimes in cases where both jurisdictions are involved. Finally, while the judgment emphasises that past sentences are a yardstick only when accompanied by articulation of unifying principles, it leaves open what degree of detail in that articulation is required before a sentencing judge or appellate court may safely rely on a collection of prior cases. These matters remain for future decisions to resolve.
Catchwords
Hili v The Queen
Judgment (69 paragraphs)
[1]
Special leave to appeal granted on grounds one to six inclusive of the draft notice of appeal.
[2]
Appeal treated as instituted and heard instanter, and dismissed.
[3]
Special leave to appeal granted on grounds one to six inclusive of the draft notice of appeal, but refused on ground seven.
[4]
Appeal treated as instituted and heard instanter, and dismissed.
[5]
On appeal from the Supreme Court of New South Wales
[6]
J T Svehla with R J Webb for the applicants in both matters (instructed by Snelgroves)
[7]
P W Neil SC for the respondent in both matters (instructed by Commonwealth Director of Public Prosecutions)
[8]
Notice: This copy of the Court's Reasons for Judgment is subject to formal revision prior to publication in the Commonwealth Law Reports.
[9]
Criminal law - Sentence - Principles - Federal offences - Applicants pleaded guilty to federal offences - Prosecution successfully appealed on ground of manifest inadequacy against head sentences and recognizance release orders imposed by sentencing judge - Court of Criminal Appeal stated that "the 'norm' for a period of mandatory imprisonment under the Commonwealth legislation is between 60 and 66% [of head sentence]" - Whether any judicially determined "norm" for ratio between time to be served in custody by federal offender and length of head sentence imposed - How consistency in federal sentencing to be achieved - Whether sentences imposed by sentencing judge manifestly inadequate - Whether Court of Criminal Appeal's reasons sufficient.
[10]
FRENCH CJ, GUMMOW, HAYNE, CRENNAN, KIEFEL AND BELL JJ. The applicants seek special leave to appeal against sentences imposed by the Court of Criminal Appeal of the Supreme Court of New South Wales, following a successful prosecution appeal against sentences that had been imposed on them in the District Court of New South Wales. The applicants had pleaded guilty, in the District Court, to offences committed in evading income taxation.
[11]
Mr Hili pleaded guilty to one charge of obtaining a financial advantage by deception from a Commonwealth entity (the Commissioner of Taxation) contrary to s 134.2(1) of the Criminal Code Act 1995 (Cth) ("the Code"). Mr Jones pleaded guilty to three charges: one of defrauding the Commonwealth contrary to s 29D of the Crimes Act 1914 (Cth) ("the Crimes Act"), one of obtaining a financial advantage by deception from a Commonwealth entity (the Commissioner of Taxation) contrary to s 134.2(1) of the Code, and one of money laundering contrary to s 400.4(1) of the Code.
The applicants were friends. In 1997, they had formed a company to perform painting and carpentry contract work for the New South Wales Department of Housing. Mr Hili had then introduced Mr Jones to his accountants, a firm which carried on its practice in Burwood, New South Wales. Each thereafter used the firm to perform accounting work. It was this firm of accountants which later invited each applicant to participate in a scheme for evading taxation.
It was agreed, for the purposes of sentencing, that each applicant had evaded taxation that would otherwise have been payable in respect of the years ended 30 June 2001, 2002 and 2003 by a company or companies he controlled, and evaded taxation that would otherwise have been payable personally for the same years. Mr Jones was also alleged to have evaded taxation otherwise payable in respect of the years ended 30 June 2000 and 30 June 2004 by a company he controlled. Taxation was evaded by claiming deductible expenses for fees falsely said to have been paid by the relevant company. The amounts falsely claimed as deductions were passed through an international round robin of transactions. These transactions passed most of the money said to have been paid by the relevant company, through accounts in the names of companies owned and operated by an accounting and business advisory firm based in Vanuatu, into the hands of one or other of the applicants. The receipts were falsely described as loans by a foreign lender.
The frauds were effected over a long time. Mr Hili engaged in 15 round robin transactions between January and November 2002. He lodged false income tax returns for two companies and for himself for the year ended 30 June 2002, for one of the companies and for himself for the year ended 30 June 2003, and for the same company for the year ended 30 June 2004. Mr Jones engaged in 11 round robin transactions between April 2001 and February 2004. He lodged false tax returns for his company for the year ended 30 June 2000, for himself and for his company for the years ended 30 June 2001, 30 June 2002 and 30 June 2003, and for the company for the year ended 30 June 2004.
[12]
The prosecution appealed against the sentences imposed, alleging that the sentences were manifestly inadequate. The Court of Criminal Appeal (McClellan CJ at CL, Howie and Rothman JJ) allowed[1] the appeal in respect of the single sentence imposed on Mr Hili, and allowed the appeal in respect of the sentences imposed on Mr Jones with respect to the first two counts, of defrauding the Commonwealth and obtaining a financial advantage by deception. On re‑sentencing, the Court of Criminal Appeal sentenced Mr Hili to a term of three years with a recognizance release order to take effect after 18 months. Mr Jones was re‑sentenced to a fixed term of 12 months on the first count (of contravening s 29D of the Crimes Act) and two years six months on the second count (of contravening s 134.2(1) of the Code) to commence six months after the commencement of the sentence for the first count. The effect of those sentences was that Mr Jones was to be imprisoned for three years. A recognizance release order was made to take effect after 18 months' imprisonment.
No order was made by the Court of Criminal Appeal with respect to the sentence of 18 months' imprisonment imposed on Mr Jones for money laundering. The Court of Criminal Appeal expressed[2] the view that the facts alleged to found the money laundering offence were all facts necessary to establish the other offences with which Mr Jones was charged, and that[3] to charge the money laundering offence raised "serious issues relating to double jeopardy". The correctness of that view was not in issue in this Court and had not been in issue in the Court of Criminal Appeal. It would be wrong to express any view about it in these reasons. What is presently relevant is that, on one view of the matter, the fact that the Court of Criminal Appeal made no order with respect to the sentence imposed for money laundering left intact the recognizance release order made at first instance in respect of that sentence.
[13]
The applicants seek special leave to appeal against the sentences imposed by the Court of Criminal Appeal. That application was referred for argument, as on appeal, before the whole Court.
The applicants sought to argue three questions. First, is there, or should there be, "a norm or starting point, expressed as a percentage" for the period of imprisonment that a federal offender should actually serve in prison before release on a recognizance release order? Second, did the Court of Criminal Appeal give adequate reasons for its conclusion that the sentences imposed at first instance were manifestly inadequate? Third, did the orders made by the Court of Criminal Appeal leave intact the recognizance release order made at first instance in respect of the sentence imposed on Mr Jones for money laundering?
The first question should be answered "no". There neither is, nor should be, a judicially determined norm or starting point (whether expressed as a percentage of the head sentence, or otherwise) for the period of imprisonment that a federal offender should actually serve in prison before release on a recognizance release order.
As to the second question, concerning adequacy of reasons, the Court of Criminal Appeal was right to conclude that the sentences imposed at first instance were manifestly inadequate. There is no reason to doubt that the sentences ordered by the Court of Criminal Appeal were within the range of sentences that could properly be imposed on the applicants following a successful prosecution appeal. The reasons given by the Court of Criminal Appeal for arriving at its orders did not state the relevant principles in a way that should be adopted, but sufficiently revealed why the sentences imposed were manifestly inadequate. The applicants did not demonstrate that, if proper principles had been applied, any lesser sentence should have been passed by the Court of Criminal Appeal on either applicant.
As to the third question, if, as the applicants submitted, the orders made by the Court of Criminal Appeal left intact the recognizance release order made at first instance in respect of the charge of money laundering preferred against Mr Jones, this Court should not deal with the issue. No party submitted that the alleged error (if it was made) could not be corrected by the Court of Criminal Appeal on application by the Attorney‑General, the Director of Public Prosecutions or Mr Jones[4]. No application having been made to the Court of Criminal Appeal to correct any alleged deficiency in the recognizance release order it made with respect to Mr Jones, this Court should not now deal with that aspect of the matter.
[14]
It will be convenient to deal with the issues that are raised in relation to the first question ("norm" or starting point) and the second question (sufficiency of reasons) by examining relevant provisions of the Crimes Act that deal with the sentencing, imprisonment and release of federal offenders, next noticing some aspects of the reasons of the Court of Criminal Appeal, and then dealing with the questions of "norm" or starting point and sufficiency of reasons in that order.
The question of "norm" or starting point raises questions about consistency in sentencing federal offenders. It will therefore be necessary to examine what is meant by "consistency", and to consider the means by which consistency is achieved. These reasons will show that the consistency that is sought is consistency in the application of the relevant legal principles, not some numerical or mathematical equivalence. Consistency in sentencing federal offenders is achieved by the proper application of the relevant statutory provisions, having proper regard not just to what has been done in other cases but why it was done, and by the work of the intermediate courts of appeal.
Reference to the proper application of the relevant statutory provisions directs attention to a fundamental point that is the starting place for any consideration of the issues in these applications.
[15]
For the most part, the relevant statutory provisions that were to be applied in sentencing the applicants are to be found in Pt IB of the Crimes Act.
Of their own force the laws of the States with respect to the sentencing of offenders could have no operation with respect to the sentencing of offenders against laws of the Commonwealth[5]. Any relevant operation is by reason of a federal law which "picks up" State law. By operation of s 68 of the Judiciary Act 1903 (Cth) ("the Judiciary Act"), some State and Territory laws in relation to the sentencing of offenders are picked up and applied when a court, exercising federal jurisdiction conferred by s 68, sentences a federal offender[6]. But, to the extent to which Pt IB of the Crimes Act otherwise provides, State and Territory laws in relation to the sentencing of offenders are not picked up. As explained in Putland v The Queen[7], s 68(1) of the Judiciary Act is "to be read in the sense it would have if, as a matter of express statement rather than implication, there was a qualification for provision otherwise made from time to time by the laws of the Commonwealth".
Of most immediate relevance to the first of the questions that is to be considered in these applications (the question of "norm" or starting point) it is to be observed that State and Territory provisions relating to the fixing of the period an offender must serve in prison, before being released, or eligible for release, are not picked up by s 68(1). Division 4 of Pt IB of the Crimes Act, concerning the fixing of non‑parole periods and the making of recognizance release orders with respect to federal offenders, makes exhaustive provision for the subject with which it deals. Because it makes exhaustive provision for that subject, State or Territory laws relating to the fixing of non‑parole periods are not picked up by, and therefore are not applied by, s 68(1) of the Judiciary Act.
[16]
Division 2 of Pt IB of the Crimes Act sets out general sentencing principles that are to be applied in sentencing federal offenders. Division 1 provides definitions of terms used in the Part.
Chief among the principles stated in Div 2 is that provided by s 16A(1): that "[i]n determining the sentence to be passed, or the order to be made, in respect of any person for a federal offence, a court must impose a sentence or make an order that is of a severity appropriate in all the circumstances of the offence". Section 16A(2) prescribes a list of matters that a sentencing court must take into account in determining the sentence to be passed or the order to be made, if those matters are relevant and known to the court. Section 16A(2) recognises that there may be other matters which may or must be taken into account.
As noted in Johnson v The Queen[8], s 16A of the Crimes Act, on its proper construction, accommodates the application of common law principles of sentencing, such as the principle of "totality" discussed in Mill v The Queen[9]. Section 16A accommodates the application of that and some other judicially developed general sentencing principles because those principles give relevant content to the statutory expression "of a severity appropriate in all the circumstances of the offence" used in s 16A(1), as well as some of the expressions used in s 16A(2), such as "the need to ensure that the person is adequately punished for the offence"[10]. But s 16A does not permit the making of generalisations across all forms of federal offence about how individual sentences are to be fixed. To attempt such a generalisation would depart from the injunction that the sentencing court "must impose a sentence or make an order that is of a severity appropriate in all the circumstances of the offence" [scil. the particular offence for which a sentence is to be imposed].
Division 4 of Pt IB (ss 19AB‑19AK) of the Crimes Act governs the fixing of non‑parole periods and the making of recognizance release orders.
A recognizance release order is defined in s 16 of the Crimes Act as an order made under s 20(1)(b). Section 20(1)(b) enables a court, before which a person is convicted of one or more federal offences, to sentence the person to imprisonment in respect of the offence, or each offence, "but direct, by order, that the person be released, upon giving security of the kind referred to in [s 20(1)(a)] either forthwith or after he or she has served a specified period of imprisonment in respect of that offence or those offences that is calculated in accordance with [s 19AF(1)]". As is apparent from the phrase "either forthwith or after he or she has served a specified period of imprisonment", s 20(1)(b) permits a sentencing court to make a recognizance release order that will take effect at any time within the period of the sentence imposed: from the time at which the sentence is imposed, to a very short time before it expires. Whatever time the recognizance release order takes effect, s 19AF requires that the sentencing court must fix the pre‑release period (the period to be served before being eligible for release on giving security) so that it ends not later than the end of the sentence, as reduced by any remissions or reductions under s 19AA. (The remissions and reductions picked up by s 19AA are remissions or reductions provided for by a law of a State or Territory in respect of State or Territory sentences.)
[17]
The reasons of the Court of Criminal Appeal were given by Rothman J. The other members of the Court agreed with those reasons. The reasons made no reference to the provisions of Pt IB of the Crimes Act. The failure to do so led the Court to state applicable principles in a way which, at the very least, is apt to mislead.
The Court of Criminal Appeal noted[14] the submission put by the prosecution that there was an "appropriate ratio between a non‑parole period and the head sentence", and that "the non‑parole period should be between 60% and 66% of the total sentence". In written submissions to the Court of Criminal Appeal, the prosecution had put forward those propositions in support of the general proposition that "an important principle to be observed is consistency in sentencing of Commonwealth offenders". A large part of the reasons given by the Court of Criminal Appeal was directed to responding to the submission that there is, or should be, some normal ratio between the time to be served in custody by a federal offender and the length of the head sentence imposed.
The Court of Criminal Appeal noted that in R v Viana[15], Meagher JA had said:
[18]
"The principles of law applicable in this area have been laid down by this Court in Bernier v The Queen[16]. There is in fact no statute which requires the non‑parole period to bear any particular proportion in relation to the head sentence, nor is there any mandatory precedent in this Court which requires a fixed sentence. The most that can be said is that this Court has usually in cases of this sort, thought the proportion ought to be somewhere between 60 and 66 per cent. That is not to say that higher percentages cannot stand."
[19]
The Court of Criminal Appeal also noted[17] what had been said by Atkinson J, for the Court of Appeal of the Supreme Court of Queensland, in R v CAK & CAL; Ex parte Commonwealth DPP[18]:
[20]
"The norm for non‑parole periods and periods required to be served before a recognizance release order for Commonwealth offences is generally considered to be after the offender has served 60 to 66 per cent of the head sentence. The precise figure may be outside this range as it is a matter of judicial discretion and is not necessarily capable of precise mathematical calculation, but that is the usual percentage of the sentence. A sentence that was well outside that range would have to have most unusual factors to justify it. In this case taking into account the offenders' early pleas of guilty, by way of ex officio indictment, the past co‑operation by the respondents, the payment of the loss sustained to the Commonwealth by the respondents and their apparent rehabilitation, the appropriate period before a recognizance release order was appropriate would have been towards the lower end of that range in the region of 60 per cent or after serving 21 and a half months imprisonment." (emphasis added, footnotes omitted)
[21]
The Court of Criminal Appeal quoted extensively from the decision of the Queensland Court of Appeal in R v Ruha, Ruha & Harris; Ex parte Commonwealth DPP[19], a decision given after CAK & CAL. The conclusion that the Court of Criminal Appeal drew from Ruha was expressed[20] in the following terms:
[22]
"In my view, the Queensland Court of Appeal have, with great respect, accurately recounted the principles applied by this Court, namely, that the 'norm' for a period of mandatory imprisonment under the Commonwealth legislation is between 60 and 66%, which figure will be affected by special circumstances applicable to a particular offender." (emphasis added)
[23]
Despite the repeated citation of references to there being a "norm" for the relationship between the term to be served and the head sentence imposed, of between 60 and 66 per cent, the Court of Criminal Appeal, when re‑sentencing the applicants, fixed a recognizance release order in respect of each of the applicants which was 50 per cent of the head sentence imposed. The Court of Criminal Appeal did that on the footing[21] that there were grounds upon which the sentencing judge "could find that there are special circumstances that warrant fixing [the period of time to be served in custody] as low as 50% (but no lower)", and that the Court of Criminal Appeal should continue the sentencing judge's approach of fixing a "lower than usual" proportion of the head sentences as the term to be served in custody. The special circumstances were said[22] to include "that this is the first time these offenders will be in prison; the good prospects of rehabilitation; and the necessity to ensure assistance in assimilating back into the community and dealing with their past alcohol issues".
How do the orders made and reasons given by the Court of Criminal Appeal accord with Pt IB of the Crimes Act? It is best to approach that issue by first considering the utility of references to a "norm".
[24]
The proposition stated[23] by the Court of Criminal Appeal in this matter: "that the 'norm' for a period of mandatory imprisonment under the Commonwealth legislation is between 60 and 66%, which figure will be affected by special circumstances applicable to a particular offender" should not be accepted. Its error is revealed by consideration of two points: the first point is a general proposition about references to a "norm"; the second comprises a set of more specific propositions drawn from what was said by the Queensland Court of Appeal in Ruha.
First, any reference to a "norm" for non‑parole periods for federal offences is, at the very least, apt to mislead. Reference to a "norm" is ambiguous. It does not reveal whether the proposition is prescriptive or descriptive. That is, is the "norm" that is identified a statement of what ought to be, or is it an observation of what has been done in past cases? If it is the former, what is its statutory root? As the earlier description of the applicable statutory provisions shows, there is none. Is it a proposition of universal application, or are there exceptions? Apparently there are exceptions: in "special circumstances". What are "special circumstances"? What is the source of these exceptions? None was identified. If reference to a "norm" is intended as a compendious description of what has been done in other cases, what are those other cases? Why are they useful comparators? Is the historical description of what has been done intended to guide what should be done thereafter? What is the principle that will tell a sentencing judge when or how the "norm" should be applied?
Even if the ambiguities inherent in references to a "norm" were to be resolved, references to a "norm" will necessarily mislead if they distract attention from the applicable statutory provisions: Pt IB of the Crimes Act. They will mislead if they suggest that the same kind of sentencing outcome will generally be expected in the sentencing of any federal offender. That is, they will mislead if they are read as saying that the same proportionate relationship should (or should normally) exist, between the time that is to be served in prison and the length of the head sentence imposed, in relation to all federal offences, no matter whether the offender has defrauded the Commonwealth, has been knowingly concerned in the importation of prohibited imports, or has committed some other federal offence. They will mislead if they suggest that matters such as the absence of prior convictions, or the willingness to co‑operate with authorities, can have no effect on fixing a non‑parole period, or time to be served before a recognizance release order takes effect, greater than a stated small percentage of the head sentence. They will mislead if they suggest that the offender must demonstrate some special circumstance to warrant departure from a set, mathematically calculated, relationship between the time to be served in custody and the head sentence.
[25]
"because the relevant factors and the relative differences in the weight to be afforded to each factor in the different aspects of the overall sentencing process may differ according to infinitely variable circumstances, there can be no 'mechanistic or formulaic'[33] approach which requires sentencing judges to ensure that the proportion which the pre‑release period bears to the sentence of imprisonment must or must usually fall within a range which is substantially narrower than the whole period of the imprisonment, which is the range the statute expressly contemplates for recognizance release orders. The proportions commonly encountered in the decided cases should themselves be the results of application of conventional sentencing principles to the particular circumstances of each case: the appellant's argument inverts that proper approach by requiring that the sentence in a particular case be substantially dictated by a pre‑determined range unless there are unusual factors." (emphasis added)
[26]
The same point had been made 20 years earlier by Hunt J, at first instance in the Supreme Court of New South Wales in R v Paull[34]. His Honour said that "the application of set ratios in fixing non‑parole periods necessarily masks the consideration which must be given to the individual facts of a particular case". And the point was repeated by the Court of Criminal Appeal in Bernier[35], and in Viana[36], and was later made by the Court of Appeal of the Supreme Court of Western Australia in Bertilone[37].
These are reasons enough to conclude that there neither is, nor should be, a judicially determined norm or starting point (whether expressed as a percentage of the head sentence, or otherwise) for the period of imprisonment that a federal offender should actually serve in prison before release on a recognizance release order. More particularly, these are reasons enough to conclude that it is wrong to say, as the Court of Criminal Appeal did[38], "that the 'norm' for a period of mandatory imprisonment under the Commonwealth legislation is between 60 and 66%, which figure will be affected by special circumstances applicable to a particular offender". It is wrong to begin from some assumed starting point and then seek to identify "special circumstances". Rather, a sentencing judge should determine the length of sentence to be served before a recognizance release order takes effect by reference to, and application of, the principles identified by this Court in Power, Deakin and Bugmy.
While this deals with the first question the applicants sought to argue in this Court, it is important to go on to consider the more general question of consistency in sentencing, upon which the prosecution based its arguments in favour of a mathematical approach to fixing the period to be served in custody for a federal offence.
[27]
The mathematical approach to fixing the period to be served in custody that was urged in the Court of Criminal Appeal (and maintained on the applications to this Court) was advanced as a means of ensuring consistency in the sentencing of federal offenders. How is consistency in federal sentencing to be achieved?
As Gleeson CJ pointed out, in Wong v The Queen[39]:
[28]
"All discretionary decision‑making carries with it the probability of some degree of inconsistency. But there are limits beyond which such inconsistency itself constitutes a form of injustice. The outcome of discretionary decision‑making can never be uniform, but it ought to depend as little as possible upon the identity of the judge who happens to hear the case. Like cases should be treated in like manner. The administration of criminal justice works as a system; not merely as a multiplicity of unconnected single instances. It should be systematically fair, and that involves, amongst other things, reasonable consistency." (emphasis added)
[29]
Consistency is not demonstrated by, and does not require, numerical equivalence. Presentation of the sentences that have been passed on federal offenders in numerical tables, bar charts or graphs is not useful to a sentencing judge. It is not useful because referring only to the lengths of sentences passed says nothing about why sentences were fixed as they were. Presentation in any of these forms suggests, wrongly, that the task of a sentencing judge is to interpolate the result of the instant case on a graph that depicts the available outcomes. But not only is the number of federal offenders sentenced each year very small, the offences for which they are sentenced, the circumstances attending their offending, and their personal circumstances are so varied that it is not possible to make any useful statistical analysis or graphical depiction of the results.
The consistency that is sought is consistency in the application of the relevant legal principles. And that requires consistency in the application of Pt IB of the Crimes Act. When it is said that the search is for "reasonable consistency", what is sought is the treatment of like cases alike, and different cases differently. Consistency of that kind is not capable of mathematical expression. It is not capable of expression in tabular form. That is why this Court held[40] in Wong that guidelines that the New South Wales Court of Criminal Appeal had determined should be used in sentencing those knowingly concerned in the importation of narcotics were inconsistent with s 16A of the Crimes Act. Those guidelines had made the weight of the narcotic the chief factor determining the sentence to be imposed, thus distracting attention from the several considerations set out in the non‑exhaustive list of matters prescribed by s 16A(2) as matters "the court must take into account" in fixing a sentence, if those matters are relevant and known to the Court.
The first and paramount means of achieving consistency in federal sentencing is to apply the relevant statutory provisions. And that requires the application of those provisions without being distracted or influenced by other and different provisions that would be engaged if the offender concerned were not a federal offender.
As was explained in Putland[41], Pt IB of the Crimes Act is not "an exhaustive statement of the will of the Parliament with respect to sentencing for federal offences". As noted earlier, there are some powers given by State or Territory law in relation to the sentencing of offenders that are picked up and applied by s 68(1) of the Judiciary Act when a court, exercising federal jurisdiction, sentences a federal offender. So, for example, in Putland, the Court held that s 68(1) picked up a provision of Northern Territory legislation relating to the imposition of an aggregate term of imprisonment.
[30]
"[R]ecording what sentences have been imposed in other cases is useful if, but only if, it is accompanied by an articulation of what are to be seen as the unifying principles which those disparate sentences may reveal. The production of bare statistics about sentences that have been passed tells the judge who is about to pass sentence on an offender very little that is useful if the sentencing judge is not also told why those sentences were fixed as they were."
[31]
Consistency in federal sentencing is to be achieved through the work of the intermediate courts of appeal. As was explained in Wong[51], the Court of Criminal Appeal was exercising federal jurisdiction in the present matters. That jurisdiction was invested in the Court by s 68 of the Judiciary Act. The laws of the State respecting the procedure for the hearing and determination of appeals (here an appeal by the Director of Public Prosecutions of the Commonwealth) were to apply and be applied, subject to s 68 of the Judiciary Act, so far as they were applicable. The relevant State provisions engaged by s 68 of the Judiciary Act were those of s 5D of the Criminal Appeal Act 1912 (NSW). Section 5D provides that the Attorney‑General or the Director of Public Prosecutions (in each case of the State) may appeal to the Court of Criminal Appeal against any sentence pronounced by the court of trial in any proceedings to which the Crown was a party, and that the Court of Criminal Appeal "may in its discretion vary the sentence and impose such sentence as to the said court may seem proper". And, as explained in Wong[52], the Attorney‑General of the Commonwealth (and by operation of s 9(1) of the Director of Public Prosecutions Act 1983 (Cth), the Director) may also appeal against a sentence imposed for a federal offence.
In dealing with appeals against sentences passed on federal offenders, whether the appeal is brought by the offender or by the prosecution, the need for consistency of decision throughout Australia is self‑evident. It is plain, of course, that intermediate courts of appeal should not depart from an interpretation placed on Commonwealth legislation by another Australian intermediate appellate court, unless convinced that that interpretation is plainly wrong[53]. So, too, in considering the sufficiency of sentences passed on federal offenders at first instance, intermediate appellate courts should not depart from what is decided by other Australian intermediate appellate courts, unless convinced that the decision is plainly wrong.
[32]
The single ground of appeal advanced by the Director in each appeal to the Court of Criminal Appeal was that the sentences imposed at first instance were manifestly inadequate. That is, the error which the Director asserted that the sentencing judge had made was of the last kind mentioned in House v The King[54]. By asserting manifest inadequacy, the Director alleged that the result embodied in the sentencing judge's orders was "unreasonable or plainly unjust". The Director did not allege that any specific error could be identified (as would be the case if the sentencing judge were said to have acted upon wrong principle, allowed extraneous or irrelevant matters to guide or affect her, mistaken the facts or not taken into account some material considerations). Rather, the Director asserted that it was to be inferred from the result that there was "a failure properly to exercise the discretion which the law reposes in the court of first instance"[55].
As was said in Dinsdale v The Queen[56], "[m]anifest inadequacy of sentence, like manifest excess, is a conclusion". And, as the plurality pointed out[57] in Wong, appellate intervention on the ground that a sentence is manifestly excessive or manifestly inadequate "is not justified simply because the result arrived at below is markedly different from other sentences that have been imposed in other cases". Rather, as the plurality went on to say[58] in Wong, "[i]ntervention is warranted only where the difference is such that, in all the circumstances, the appellate court concludes that there must have been some misapplication of principle, even though where and how is not apparent from the statement of reasons". But, by its very nature, that is a conclusion that does not admit of lengthy exposition. And, in the present matters, the Court of Criminal Appeal, having described the circumstances of the offending and the personal circumstances of the offenders, said[59] that "the sentence imposed in these matters is so far outside the range of sentences available that there must have been error".
The Court of Criminal Appeal also said[60] that "manifest error is fundamentally intuitive". That is not right. No doubt, as the Court went on to say[61], manifest error "arises because the sentence imposed is out of the range of sentences that could have been imposed and therefore there must have been error, even though it is impossible to identify it". But what reveals manifest excess, or inadequacy, of sentence is consideration of all of the matters that are relevant to fixing the sentence. The references made by the Court of Criminal Appeal to the circumstances of the offending and the personal circumstances of each offender were, therefore, important elements in the reasons of the Court of Criminal Appeal.
[33]
"the sentence imposed upon an accused person for an offence is a matter peculiarly within the province of the judge who hears the charge: he has a discretion to exercise which is very wide, but it must be exercised judicially, according to rules of reason and justice, and not arbitrarily or capriciously or according to private opinion."
[34]
This is not, however, inconsistent with the rule of law. Hayek saw one "essential point" of the rule of law as being "that the discretion left to the executive organs wielding coercive power should be reduced as much as possible."[72] That does not deny the legitimacy of judicial discretion as described by Starke J. The existence of judicial discretion in sentencing has been criticised[73]. But the starkest alternatives - fixed sentences or fixed minimum sentences - have been in general decline for the last two centuries, and, at least in Australia, a recent revival is very minor in scope and of unproven utility. Thus although, pending resolution of the problem in this Court, a court in one jurisdiction is at liberty to reach a point of view on a question of legal principle which contradicts an earlier court in another jurisdiction if the later court thinks that the earlier is plainly wrong, and although, subject to that, it is not possible for two courts to reach contradictory views on a point of legal principle which are both correct, it is possible for two courts, each acting on an identical legal principle, making no error of fact, omitting no relevant consideration and taking into account no irrelevant consideration, to arrive at different sentences without either of them being "wrong". As McHugh, Hayne and Callinan JJ said in Pearce v The Queen[74]: "Sentencing is not a process that leads to a single correct answer arrived at by some process admitting of mathematical precision". The circumstances of particular crimes and the "character, antecedents and conditions"[75] of particular offenders are so various, the combinations in which they can occur are so numerous, and the relationship between these factors and the purposes which criminal sentences are to serve can be so impalpable, that the application to them of discretionary judgment permitting a range of legitimate outcomes is inevitable. Section 16A(1) of the Crimes Act 1914 (Cth) requires the court to impose a sentence of "a" severity appropriate in all the circumstances: it does not follow that only one correct sentence is possible.
It is not surprising that the leading case in this Court on the power of appellate courts to intervene in discretionary judgments, House v The King, is a sentencing case. Nor is it surprising that in House v The King the Court divided appellable errors in discretionary judgments like sentencing judgments into two categories. The first was stated thus[76]:
[35]
"If the judge acts upon a wrong principle, if he allows extraneous or irrelevant matters to guide or affect him, if he mistakes the facts, if he does not take into account some material consideration, then his determination should be reviewed and the appellate court may exercise its own discretion in substitution for his if it has the materials for doing so."
[36]
The second existed where the order was unreasonable or plainly unjust[77].
[37]
"It may not appear how the primary judge has reached the result embodied in his order, but, if upon the facts it is unreasonable or plainly unjust, the appellate court may infer that in some way there has been a failure properly to exercise the discretion which the law reposes in the court of first instance. In such a case, although the nature of the error may not be discoverable, the exercise of the discretion is reviewed on the ground that a substantial wrong has in fact occurred."
[38]
In the second category, appellate intervention takes place because the character of the order indicates that some underlying error within the first category has taken place, even though it is not possible to identify it.
[39]
When the second category is relied on, the usual complaint is that the sentence is "manifestly excessive" or "manifestly inadequate". Mere excessiveness or inadequacy will not reveal that there is an error of either an identifiable or an unidentifiable nature. The difficulty which the principles in House v The King create for appellants in sentencing appeals - whether defendants complaining of "manifest excessiveness" or the prosecution complaining of "manifest inadequacy" - is that they give sentencing judges "a wide measure of latitude which will be respected by appellate courts."[78] But it does not follow that, when a sentencing judge in one case looks back on the reasons for judgment of an earlier sentencing judge in a similar case, the later judge should give the earlier one "a wide measure of latitude" in the sense of not departing from the outcome unless it is plainly wrong.
Sentences must be reasonably consistent. But it does not follow that disparities between them may not exist. Within the boundaries of reason, and leaving aside the special instance of co-offenders, where marked disparity renders sentences vulnerable on appeal[79], it cannot be said that any particular disparate sentence is necessarily wrong merely because it is disparate. Indeed, even within a single jurisdiction, one court, while bound by whatever this Court or the intermediate appellate court for that jurisdiction has held to be the correct legal principles (statutory or common law), may arrive at sentencing results in particular cases which are different from those reached by earlier courts in that jurisdiction without being open to appellate reversal or criticism for "error" merely because of those differences.
Thus two courts may arrive at different sentences because the later court considers the first to have erred, not in relation to the identification of legal principle, but in relation to factual reasoning or in relation to the exercise of discretionary judgment. It is open to a later court (whether an intermediate appellate court or a trial court) to depart from the sentencing conclusion of an earlier intermediate appellate court or trial court even though the circumstances seem indistinguishable. It is open for the later court to do this simply because the later court thinks that the earlier court erred in fact: in that event the circumstances become distinguishable. It is also open for the later court to do this merely because it thinks the earlier court erred in the exercise of discretionary judgment - that is, arrived at a sentence which the later court, accepting the correctness of the legal principles stated, the facts found and the considerations taken or not taken into account by the earlier court, considers nonetheless to be too high or too low. The later court's liberty to differ from the sentencing conclusion reached by the earlier court does not exist only where it thinks the earlier court to be plainly wrong. It exists where the later court thinks the earlier court's conclusion to be merely wrong. Indeed it exists even where the later court does not think the earlier court's conclusion to be "wrong", but just disagrees with it. The liberty of the later court continues even if more than one earlier court has reached a conclusion with which the later court disagrees. Even after a court carrying out the difficult obligation of sentencing has identified the correct legal principles, found the facts correctly, taken into account all relevant considerations and excluded all irrelevant considerations, the court is left with a field in which to exercise a discretionary judgment. It is no doubt right for a sentencing court to examine what has happened in cases similar to the one under consideration. And it is no doubt reasonable for a sentencing court to behave with humility in reading the opinions of other judges in earlier cases who may be abler, better qualified, more learned, or more experienced. But in exercising its discretionary judgment, the primary duty of a sentencing court is to be true to its own perception of what degree of severity or leniency is appropriate.
[40]
R v Jones; R v Hili (2010) 76 ATR 249. On 3 September 2010, after the applications for special leave had been filed, the Court of Criminal Appeal published supplementary reasons for judgment correcting two slips in its original reasons: R v Jones; R v Hili (No 2) (2010) 242 FLR 64. The supplementary reasons need not be examined in any detail. ↑
[41]
Section 19AH(1) of the Crimes Act 1914 (Cth) ("the Crimes Act") provides, so far as presently relevant:
[42]
"Where a court fails ... to make, or properly to make, a recognizance release order, under this Act:
[43]
(a) that failure does not affect the validity of any sentence imposed on a person; and
[44]
(b) the court must, at any time, on application by the Attorney‑General, the Director of Public Prosecutions or the person, by order, set aside any ... recognizance release order that was not properly ... made and ... make a recognizance release order under this Act." ↑
[45]
Bass v Permanent Trustee Co Ltd (1999) 198 CLR 334 at 352 [35]; [1999] HCA 9; Solomons v District Court (NSW) (2002) 211 CLR 119 at 134 [21]; [2002] HCA 47. ↑
[46]
Putland v The Queen (2004) 218 CLR 174; [2004] HCA 8. ↑
[47]
(2004) 78 ALJR 616 at 622 [15] per Gummow, Callinan and Heydon JJ; 205 ALR 346 at 353; [2004] HCA 15. Cf the express provisions considered in Chief Executive Officer of Customs v Labrador Liquor Wholesale Pty Ltd (2003) 216 CLR 161 at 168‑169 [15]‑[16], 178 [54], 205 [134]; [2003] HCA 49. ↑
[48]
(2010) 76 ATR 249 at [44]; see also (2010) 242 FLR 64 at [56]‑[64]. ↑
[49]
(2010) 76 ATR 249 at [44]; see also (2010) 242 FLR 64 at [60]. ↑
[50]
(1984) 58 ALJR 367; 54 ALR 765; [1984] HCA 31. ↑
[51]
(1990) 169 CLR 525; [1990] HCA 18. See also Inge v The Queen (1999) 199 CLR 295; [1999] HCA 55. ↑
[52]
See R v Harkness [2001] VSCA 87 per Callaway JA, quoting from his Honour's judgment in R v Pope (2000) 112 A Crim R 588 at 597 [28]. ↑
[53]
(2001) 207 CLR 584 at 591 [6]; [2001] HCA 64. ↑
[54]
(2001) 207 CLR 584 at 608 [65], 612‑613 [78], 616 [87] per Gaudron, Gummow and Hayne JJ, 632 [131] per Kirby J. ↑
[55]
Australia, House of Representatives, Crimes Legislation Amendment Bill (No 2) 1989, Explanatory Memorandum at 1. ↑
[56]
[2010] NSWCCA 194 at [304], citing Wong v The Queen (2001) 207 CLR 584 at 606 [59]. ↑
[57]
See Wong (2001) 207 CLR 584 at 602 [47], and the cases there cited. ↑
[58]
Australian Securities Commission v Marlborough Gold Mines Ltd (1993) 177 CLR 485 at 492; [1993] HCA 15; Farah Constructions Pty Ltd v Say‑Dee Pty Ltd (2007) 230 CLR 89 at 151‑152 [135]; [2007] HCA 22. See also CAL No 14 Pty Ltd v Motor Accidents Insurance Board (2009) 239 CLR 390 at 411‑412 [48]‑[50], 417 [63]; [2009] HCA 47. ↑
[59]
(2000) 202 CLR 321 at 325 [6]; [2000] HCA 54. ↑
[60]
"The Rule of Law", (2007) 66 Cambridge Law Journal 67 at 71. ↑
[61]
The Road to Serfdom, (1962) at 54 (emphasis added). ↑
[62]
Various arguments may be seen in Ashworth, Sentencing and Criminal Justice, 5th ed (2010) at 23-24, 41-42, 49, 52-53, 76-77, 276-277 and 417-418. ↑
[63]
(1998) 194 CLR 610 at 624 [46]; [1998] HCA 57. ↑
[64]
Lowe v The Queen (1984) 154 CLR 606 at 612 per Mason J; [1984] HCA 46. ↑
[65]
House v The King (1936) 55 CLR 499 at 505 per Dixon, Evatt and McTiernan JJ. ↑
[66]
House v The King (1936) 55 CLR 499 at 505 per Dixon, Evatt and McTiernan JJ. ↑
[67]
Postiglione v The Queen (1997) 189 CLR 295 at 336 per Kirby J; [1997] HCA 26. ↑
[68]
Lowe v The Queen (1984) 154 CLR 606; Postiglione v The Queen (1997) 189 CLR 295. ↑
[69]
Wong v The Queen (2001) 207 CLR 584 at 605 [57] per Gaudron, Gummow and Hayne JJ; [2001] HCA 64. ↑
Special leave to appeal granted on grounds one to six inclusive in each matter (refused on ground seven in Jones); appeals treated as instituted and heard instanter and dismissed in each matter.
Mr Hili and his companies evaded taxation totalling $398,537.82. Mr Jones and his companies evaded taxation totalling $362,925.24. Having regard to penalties and interest, Mr Hili became liable to pay over $1 million to the Australian Taxation Office as a result of his conduct. Mr Jones was liable to pay an amount of between $900,000 and $1 million. At the time of sentence, Mr Hili had paid some of what was owing, and was realising assets to pay the rest. Mr Jones was also taking steps to pay what he owed.
In the District Court, Morgan DCJ sentenced Mr Hili to imprisonment for a total of 18 months, with a recognizance release order to take effect after seven months. Mr Jones was sentenced on each count to 18 months. Each sentence was made concurrent with the others. Again, a recognizance release order was made to take effect after seven months.
The sentencing judge noted that each applicant was previously of good character. Each had pleaded guilty at the earliest opportunity. Each had admitted his guilt in the course of interviews with police, and provided the authorities with what police believed to be "all information and assistance available to him". Each undertook to co‑operate with law enforcement agencies, including by giving evidence in any subsequent proceedings brought in respect of four named persons who were alleged to have been instrumental in establishing and operating the evasion scheme the applicants had used. The undertakings are enforceable through the mechanisms prescribed by s 21E of the Crimes Act (by which the Director of Public Prosecutions may appeal against the reduction of a sentence where promised co‑operation is not provided). The sentencing judge took the view "that in each case the appropriate discount [on account of pleas of guilty and past and future assistance] is one of fifty per cent for the plea and the assistance of which twelve and a half per cent is referable to future assistance in accordance with the undertaking signed by each offender".
Special leave to appeal to raise the third question should be refused. Special leave to appeal should be granted in respect of the remaining grounds of appeal, but each appeal dismissed.
Subject to s 19AB(3), if a court imposes a federal sentence or federal sentences that in aggregate exceed three years (and the person is not already serving or subject to a federal sentence), the court must[11] either fix a single non‑parole period in respect of that sentence or those sentences, or make a recognizance release order. Section 19AB(3) permits a court to decline to fix a non‑parole period or make a recognizance release order "if, having regard to the nature and circumstances of the offence or offences concerned and to the antecedents of the person, the court is satisfied that neither is appropriate". Subject to s 19AC(3) and (4), if a court imposes a federal sentence or federal sentences that in aggregate do not exceed three years (and the person is not already serving or subject to a federal sentence) the court must[12] make a recognizance release order and must not fix a non‑parole period. The qualification provided by s 19AC(3) relates to sentences that, in aggregate, do not exceed 6 months. The qualification provided by s 19AC(4) is of the same kind as is made by s 19AB(3): the sentencing court need not make a recognizance release order if satisfied it is not appropriate to do so.
The general provisions made by ss 19AB and 19AC (as well as certain other provisions of the Crimes Act) have effect subject to s 19AG[13]. That section provides, in effect, that for certain specified offences (treachery, terrorism offences, and offences against Divs 80 or 91 of the Code, which deal with treason, sedition and offences relating to espionage and similar activities) a court must fix a single non‑parole period of at least three‑quarters of the sentence for that offence. But, as noted earlier, the combined effect of ss 20(1)(b) and 19AC was that, in the present cases, the statute made no provision fixing any relationship between the head sentence and a recognizance release order. On the contrary, the sentencing court had power to fix the recognizance release orders to take effect at any time during the period of the head sentences.
In Ruha, the Queensland Court of Appeal went to some lengths[24] to emphasise the cardinal importance of beginning consideration of the sentencing of any federal offender by examining the applicable statutory provisions, particularly Pt IB of the Crimes Act. The Court of Appeal summarised[25] the effect of the relevant provisions of the Crimes Act in three propositions, but neither the summary, nor any of the individual propositions is, or was intended to serve as, a substitute for the statutory language.
The Court of Appeal in Ruha examined[26] what considerations bear upon fixing the length of a pre‑release period under a recognizance release order. As the Court of Appeal rightly said[27], ss 16A(1) and (2) "make it plain that all of the circumstances, including the matters in the non‑inclusive list in s 16A(2), must be taken into account in making recognizance release orders just as they must be taken into account in imposing a sentence of imprisonment". In determining what recognizance release order is to be made, s 16A(1) requires the sentencing court to "make an order that is of a severity appropriate in all the circumstances of the offence". What is the "severity appropriate" is determined having regard to the general principles identified by this Court in Power v The Queen[28], Deakin v The Queen[29] and Bugmy v The Queen[30].
In the present cases, one consideration critical to the making of recognizance release orders was the determination of what was the period of imprisonment that justice required that each offender must serve in custody. And as the Queensland Court of Appeal pointed out in Ruha[31], again correctly, "the necessary deterrent and punitive effects of sentences for serious tax fraud must be reflected both in the head sentence and also in any provision for earlier release from custody".
The Queensland Court of Appeal was right to conclude, as it did[32], that:
In addition, there are respects in which Pt IB of the Crimes Act itself refers to and picks up State and Territory legislation affecting service of a sentence of imprisonment. Those provisions of the Crimes Act include: s 16E concerning the commencement of sentences; s 18(2) concerning imprisonment in a particular kind or class of prison; s 19A concerning detention of federal offenders in State or Territory prisons; s 19AA concerning remissions and reductions of sentences; and s 19AZD concerning leave of absence for and pre‑release of prisoners. But Div 4 of Pt IB (which deals with the fixing of non‑parole periods and the making of recognizance release orders) does not expressly engage any State or Territory law which prescribes how non‑parole periods are to be fixed in sentencing under State or Territory law. And, as was pointed out in Putland[42], the provisions of Div 4 of Pt IB are cast in terms that not only provide "a separate regime for fixing federal non‑parole periods rather than relying on applied State or Territory legislation"[43], those provisions deal exhaustively with that subject. State and Territory legislation concerning the fixing of non‑parole periods has no application to the sentencing of federal offenders.
Next, in seeking consistency, sentencing judges must have regard to what has been done in other cases. In the present matter, the prosecution produced detailed information, for the sentencing judge and for the Court of Criminal Appeal, about sentences that had been passed in other cases arising out of tax evasion as well as cases of customs and excise fraud and social security fraud. Care must be taken, however, in using what has been done in other cases.
In Director of Public Prosecutions (Cth) v De La Rosa[44], Simpson J accurately identified the proper use of information about sentences that have been passed in other cases. As her Honour pointed out[45], a history of sentencing can establish a range of sentences that have in fact been imposed. That history does not establish that the range is the correct range, or that the upper or lower limits to the range are the correct upper and lower limits. As her Honour said[46]: "Sentencing patterns are, of course, of considerable significance in that they result from the application of the accumulated experience and wisdom of first instance judges and of appellate courts." But the range of sentences that have been imposed in the past does not fix "the boundaries within which future judges must, or even ought, to sentence"[47]. Past sentences "are no more than historical statements of what has happened in the past. They can, and should, provide guidance to sentencing judges, and to appellate courts, and stand as a yardstick against which to examine a proposed sentence"[48] (emphasis added). When considering past sentences, "it is only by examination of the whole of the circumstances that have given rise to the sentence that 'unifying principles' may be discerned"[49].
As the plurality said in Wong[50]:
The applicants' submissions criticising the sufficiency of the reasons given by the Court of Criminal Appeal pointed out that the Court of Criminal Appeal identified no specific error in the sentencing judge's findings of fact or reasons. That is right, but because the only ground advanced by the Director was the ground of manifest inadequacy, it had to be assumed that the Director alleged no specific error. That the Court of Criminal Appeal identified no specific error is, therefore, unsurprising. The absence of identification of such an error does not bespeak error on the part of the Court of Criminal Appeal. The reasons given by the Court of Criminal Appeal for concluding that the sentences passed were manifestly inadequate sufficiently revealed the bases for that conclusion.
In the present matters, the inadequacy of the sentences imposed at first instance was evident from consideration of all of the matters that were relevant to fixing a sentence (and making a recognizance release order) "of a severity appropriate in all the circumstances of the offence"[62]. The chief considerations which pointed to inadequacy in these cases were the nature of the offending, and the sentences that had been imposed in cases most closely comparable with the present.
The applicants' offending was sustained over a long time. It was planned, deliberate and deceitful, requiring for its implementation the telling of many lies. The applicants acted out of personal greed. The amount of tax evaded was not small. Detection of offending of this kind is not easy. Serious tax fraud, which this was, is offending that affects the whole community. As was pointed out in Ruha[63], the sentences imposed had to have both a deterrent and a punitive effect, and those effects had to be reflected in the head sentences and the recognizance release orders that were made.
Of the many other cases of fraud against the Commonwealth, to which the prosecution drew attention, at first instance and on appeal to the Court of Criminal Appeal, only one or two were closely comparable with the offences committed by the applicants. Rightly, the prosecution gave prominence to those cases in its written submissions to the Court of Criminal Appeal.
In the County Court of Victoria, Judge Wood had sentenced[64] to two and a half years' imprisonment an offender who had evaded taxation of about $318,000 by what were described[65] as "complex arrangements to funnel Australian earned funds offshore, under the guise of authenticity". His Honour had made a recognizance release order to take effect after 15 months' imprisonment. The offender, in that case, was a public figure. The sentencing judge found that the offender had been the subject of widespread public opprobrium for two years before he was sentenced. The offending was found to have been borne out of need, not greed. An appeal to the Court of Appeal of the Supreme Court of Victoria against the sentence was abandoned when the Court of Appeal informed counsel that consideration may be given to increasing the sentence that had been passed[66].
The second case to which the prosecution gave special emphasis[67] concerned multiple charges of tax fraud involving just under $329,000. The offender, a tax agent, was sentenced to six years' imprisonment with a non‑parole period of four and a half years. A non‑parole period of four years was fixed on appeal (the fixing of four and a half years being treated as a slip on the part of the sentencing judge) but the appeal against sentence otherwise dismissed.
The sentences passed on the applicants at first instance were very much lower than those passed in either of those cases. Allowing, as one must, for the different circumstances of each of the cases under consideration, the difference in sentences passed on the applicants at first instance, and those that were passed on the other offenders, is so large that the Court of Criminal Appeal was right to conclude that "there must have been some misapplication of principle [by the sentencing judge], even though where and how is not apparent from the statement of reasons"[68].
No separate argument was advanced by the applicants to the effect that the sentences that were passed on each applicant by the Court of Criminal Appeal were more severe than should have been passed. Having regard to the matters that have been mentioned, there is no reason to conclude that, applying proper principles, the Court of Criminal Appeal should have imposed any lesser sentence on either applicant than it did.
Each applicant should have special leave to appeal, limited in each case to grounds raising the first two questions identified at the start of these reasons (grounds one to six inclusive). Each appeal should be treated as instituted and heard instanter but dismissed.
HEYDON J. I agree with the orders proposed by French CJ, Gummow, Hayne, Crennan, Kiefel and Bell JJ, and the answers given to the three questions argued, but would adopt different reasoning in relation to the "norm" question in one respect.
I agree that no relevant statute requires the non-parole period of a sentence for a federal offender to represent any particular proportion of the head sentence. Nor does any common law rule. I agree that there must be consistency of decision among sentencing judges, and among intermediate appellate courts hearing sentencing appeals, in the sense that there must be consistency in the identification of the relevant legal principles; and that to that end an intermediate appellate court should not depart from a decision of another Australian intermediate appellate court about what those legal principles are unless convinced that that decision is plainly wrong. But the same does not necessarily apply in relation to aspects of sentencing other than legal principles.
A proposition asserted to be a proposition of legal principle is either right or wrong: if it is wrong, it will sometimes fall into the category of being "plainly wrong". To that statement qualifications must be made. Some time may pass before the moment arrives when the authorities have moved into a position permitting lawyers, and in particular courts, to say confidently that a proposition of legal principle - whether it be the construction of a statute or a common law rule - is right. The proposition may have developed as the result of particular choices being made as new problems to which the proposition must be applied come to attention. At least in the case of a common law principle, there will often be inherent in it seeds of future development, and some of those seeds may germinate in the medium term, the long term, or the very long term. Some authorities holding a principle to be right may later be overruled. But, subject to those qualifications, our law, including the law relating to sentencing, is characterised for the most part by principles which, at least in the short term, have a measure of certainty and predictability in being fixed and clear. That is because most judges avoid what Lord Bingham of Cornhill called "excessive innovation and adventurism"[69]. The less fixed and clear the principles are, the less is the system governed by the rule of law, and the closer it would go towards lacking a feature which in Hayek's opinion has been the greatest contribution to the prosperity of the West[70].
But a sentencing judgment does not rest only on identifying the correct legal principles. Those legal principles render some matters relevant and others irrelevant. They sometimes require an inquiry into particular factual circumstances. If they affect how the sentencing discretion should be exercised, being mandatory, they restrict the scope of the sentencing discretion. Thus so far as sentencing principles allocate the weight to be given to particular factors or mandate how they should be employed in relation to each other, those principles diminish the extent to which the court exercises a discretionary judgment, for those allocations and mandates would create duties, not discretions. The legal principles applicable to sentencing in our law do, however, ordinarily leave room for discretionary judgment. As Starke J said in House v The King[71]:
If the position were otherwise, a later court would be compelled to impose sentences on offenders which it thought to be too harsh or too lenient merely because earlier courts had followed that path, even though the question whether a sentence should be heavy or light is not a question of law. This would be a novel application of the doctrine of precedent. For a "sentence itself gives rise to no binding precedent. What may give rise to precedent is a statement of principles which affect how the sentencing discretion should be exercised"[80].