What happened
Matthew James Elliott and Bronson Matthew Blessington, aged 16 and 14 respectively at the time of the offence, were convicted alongside Stephen Jamieson of the murder of Janine Balding committed on 8 September 1988. Following a joint trial before Newman J in the Supreme Court of New South Wales, all three were convicted on 21 June 1990. On 18 September 1990 Newman J sentenced Jamieson to the mandatory life sentence then applicable under ss 19 and 442 of the Crimes Act 1900 (NSW). For the two juveniles Elliott and Blessington, Newman J exercised his discretion to impose life sentences rather than a lesser term and added the now-famous non-release recommendation: "So grave is the nature of this case that I recommend that none of the prisoners in this matter should ever be released."
Appeals against conviction by Jamieson and Elliott, and against sentence by Elliott and Blessington, were heard by the Court of Criminal Appeal (Gleeson CJ, Hope AJA, Lee AJ) and determined on 17 February 1992. The Court dismissed the conviction appeals and, while granting leave to appeal against sentence, dismissed those appeals as well, holding that the life sentences imposed on the juveniles fell within the range of Newman J's discretion. Gleeson CJ expressly noted the absence of any statutory basis for the making of the non-release recommendation or for any appeal against it.
More than a decade later, legislative amendments to the regime governing existing life sentences altered the significance of such recommendations. The Sentencing Act 1989 (NSW), as amended by the 1997 legislation, defined a "non-release recommendation" and used it as a trigger for a higher eligibility threshold (20 years served) before an offender could apply to the Supreme Court under s 13A for determination of a minimum term. The Crimes (Sentencing Procedure) Amendment (Existing Life Sentences) Act 2005 (NSW) further clarified that the definition captured recommendations even if they had been "quashed, set aside or called into question".
Armed with the discovery that the 1992 orders of the Court of Criminal Appeal had never been perfected by entry on the indictment, Elliott and Blessington sought leave in 2006 to reopen their sentence appeals and, if necessary, to appeal out of time against the non-release recommendation itself. The Court of Criminal Appeal (Spigelman CJ, Howie J; Kirby J dissenting) refused leave. Kirby J would have reopened the appeals, quashed the life sentences and substituted fixed terms of 28 years with non-parole periods expiring on 8 September 2009. The majority refused both the reopening application and any extension of time. Elliott and Blessington then appealed to the High Court, which heard the matters together and, on 8 November 2007, dismissed both appeals.
Why the court decided this way
The High Court began from the principle of finality articulated by Barwick CJ in Ratten v The Queen: once a jury returns verdicts of guilty, the judicial power in respect of the indictment is spent upon sentencing, subject only to the statutory appellate regime created by the Criminal Appeal Act 1912 (NSW). That statute, as explained in Grierson v The King, does not confer a general appellate power but instead defines the available grounds, the procedure and the duty of the Court of Criminal Appeal. Once an appeal has been heard and determined on the merits, the Court has no further authority to reopen it.
Two distinct routes to relief were advanced. The first was that Newman J's non-release recommendation itself constituted a "sentence" or "order made by the court of trial" within the statutory definition in s 2(1) of the Criminal Appeal Act, thereby opening an avenue of appeal under s 5(1)(c). The Court rejected this for two reasons. At the time Newman J spoke, the recommendation had no statutory foundation; it was simply an observation directed to the executive administration of the ticket-of-leave system. Analogy to parole or probation orders made under express statutory power was inapt. Nor did later legislation confer the necessary character upon the recommendation. In Baker v The Queen the Court had characterised the 1997 amendments as merely selecting the existence of a non-release recommendation as a "trigger" or criterion for a more onerous eligibility rule. That legislative choice did not mean the recommendation had been made "by" the court of trial so as to satisfy the statutory definition of "sentence". The 2005 Amendment Act's precautionary extension of the definition to recommendations that had been called into question did not alter this analysis. The recommendation therefore never became an appealable order, and the Court of Criminal Appeal had been correct to refuse leave to bring a fresh or out-of-time appeal against it.
The second route relied upon the undisputed fact that the 1992 orders had never been perfected by entry on the indictment. The Court accepted that, until perfection, the Court of Criminal Appeal retained power to reopen. That power, however, is not at large. Drawing on the incidental powers doctrine confirmed in DJL v Central Authority and the criteria articulated by Mason CJ in Autodesk Inc v Dyason (No 2), reopening is available only where the court has proceeded on a misapprehension of fact or law that cannot be attributed solely to the default of the party seeking relief. The 1992 decision had applied the law as it then stood under the Crimes Act and had correctly concluded that life sentences were within discretion. Subsequent legislative changes to the redetermination regime could not retroactively create a misapprehension in the 1992 reasons. Section 6(3) of the Criminal Appeal Act requires the appellate court to ask whether some other sentence "is warranted in law and should have been passed". That inquiry looks to the law at the time of sentencing and appeal, not to later statutory amendments. To hold otherwise would undermine the statutory scheme for timely appeals.
The Court also rejected the proposition that refusing reopening would show disrespect to Parliament. It is the judicial branch's province to apply the law governing the case before it. Unless the later legislation expressly or impliedly amended the Criminal Appeal Act itself (a proposition disavowed by the Solicitor-General), the existence of a legislative policy favouring strict treatment of non-release recommendation cases could not preclude the proper exercise of appellate jurisdiction. On the facts, however, no miscarriage of justice had been shown in the 1992 decision, and the passage of time and volume of intervening legislation did not itself justify reopening. Accordingly both appeals were dismissed.
Before and after state of the law
Prior to the 1989–1990 reforms, sentencing judges in New South Wales commonly made non-release recommendations when imposing life sentences. These remarks were directed to the executive's administration of the ticket-of-leave system then operating under s 463 of the Crimes Act 1900 (NSW). There was no statutory basis for the recommendation itself, nor any recognised avenue of appeal against it, as Gleeson CJ observed in the 1992 decision.
The Prisons (Serious Offenders Review Board) Amendment Act 1989 (NSW) and the Sentencing (Life Sentences) Amendment Act 1989 (NSW) effected a fundamental change. Section 463 was repealed, the ticket-of-leave system was replaced, and s 13A of the Sentencing Act 1989 (NSW) created a new original jurisdiction in the Supreme Court to fix minimum and additional terms for existing life sentences after eight years had been served. The court was required to have regard to any relevant comments made by the original sentencing judge. The 1997 amendments then introduced an eligibility threshold of 20 years for offenders who were the subject of a non-release recommendation (defined broadly to include any expression of opinion that the person should never be released) and imposed a "special reasons" test before a minimum term could be set. Baker v The Queen characterised this as the selection of a legislative criterion rather than the creation of new legal effect "by" the original sentencing court.
The Crimes (Sentencing Procedure) Amendment (Existing Life Sentences) Act 2005 (NSW) added a further clarification that the definition of existing life sentence included any non-release recommendation even if it had been quashed, set aside or called into question. This was apparently inserted for more abundant caution after Gleeson CJ's 1992 remarks disapproving Newman J's recommendation.
After the High Court's 2007 decision the law remained that a non-release recommendation does not itself constitute an appealable order. The statutory redetermination pathway under what is now Part 7 of the Crimes (Sentencing Procedure) Act 1999 (NSW) continues to treat the existence of such a recommendation as a relevant but not determinative factor. The strict criteria for reopening unperfected orders of the Court of Criminal Appeal, anchored in the Autodesk principles, continue to govern applications brought long after the original decision.
Key passages with plain-English translation
Paragraph [17] quotes Gleeson CJ's 1992 observation: "There does not appear to have been any statutory basis for the making of the 'recommendation', nor, for that matter, does there seem to be any statutory basis for appealing against it." In plain English, the Chief Justice was saying that Newman J was simply expressing a view to the executive; he was not making a formal order that could be challenged in a court of appeal.
At [24] the Court endorses Spigelman CJ's reasoning that treating the 1997 Act's use of the recommendation as a criterion is "inconsistent with the proposition that the legal consequence of a non-release recommendation now could be said to arise from anything done 'by' the court of trial". Translation: Parliament's later decision to give the recommendation evidentiary weight in a parole-style review does not magically turn the judge's earlier comment into a court order that was appealable all along.
The adoption of Mason CJ's statement in Autodesk at [34] is central: the jurisdiction to reopen arises only where the court has laboured under a misapprehension of fact or law not solely the fault of the applicant, and "is not to provide a backdoor method by which unsuccessful litigants can seek to reargue their cases". In plain language, you cannot ask a court to have another go simply because you have thought of better arguments or the law has since changed.
Paragraph [42] states: "The subsequent legislation affecting the position of the appellants did not create any miscarriage of justice by the 1992 decision which called for interception in the perfection of the orders which had then been pronounced." Translation: even though the law changed, the 1992 judges got the law right for their time; later parliamentary choices do not retroactively make the earlier decision wrong.
What fact patterns trigger this precedent
This decision is triggered whenever an offender seeks to characterise a judicial remark—particularly a non-release or "never to be released" observation—as an appealable component of sentence under the Criminal Appeal Act 1912 (NSW). It applies to any attempt to reopen a finally determined appeal against sentence on the basis that orders were not formally perfected where the only new matter is subsequent legislative change that alters the practical consequences of the original sentence.
The precedent is engaged where an existing life sentence was imposed before the 1990 reforms and the offender relies on the 1997 or 2005 amendments to argue that a non-release recommendation has acquired new legal character. It also governs any application to reopen an unperfected Court of Criminal Appeal order more than a decade after the event when the asserted error is not present in the original reasons but arises only from later statutory policy.
Fact patterns falling outside the ratio include cases where the original sentencing judge made an actual order under statutory power (for example, a parole order under Commonwealth legislation as in R v Carngham) or where the alleged misapprehension existed on the material available at the time of the original appeal and was not the product of later legislative intervention.
How later courts have treated it
The High Court itself treated Grierson v The King as definitively excluding any general power to reopen a finally determined criminal appeal, while acknowledging that the incidental powers of a statutory court extend to reopening before perfection. Baker v The Queen was applied for the limited proposition that the 1997 Act merely selected a legislative criterion rather than altering the character of the original judicial act. Autodesk Inc v Dyason (No 2) supplied the controlling test for the exercise of the reopening discretion and was treated as stating a principle of general application to both civil and criminal appellate courts.
The joint judgment carefully distinguished appeals by way of rehearing (where later law may be applied) from the error-based jurisdiction under s 6(3) of the Criminal Appeal Act. It cited DJL v Central Authority for the proposition that a statutory court possesses only those powers incidental to its jurisdiction, reinforcing the absence of any inherent power to reopen.
The decision also confirmed that s 12(1) of the Criminal Appeal Act, which permits the Court to exercise powers available to the Supreme Court in civil matters where necessary in the interests of justice, does not authorise reopening on the broad grounds that might be available in ordinary civil litigation once the criminal finality principles are engaged.
Still-open questions
The judgment leaves open the precise boundary between a mere recommendation and an order that has legal effect under later legislation. While the 2005 Amendment Act's precautionary language was held not to convert the recommendation into an order "made by" the trial court, the Court did not decide whether a different form of words by a sentencing judge might cross that line.
The interaction between the reopening principles and fresh evidence of post-sentencing rehabilitation or other developments remains expressly undecided; Kirby J's dissenting reasons canvassed authorities from several jurisdictions but the majority found it unnecessary to enter that field once reopening was refused on other grounds.
Whether an express parliamentary amendment to the Criminal Appeal Act itself that altered the finality rule would be constitutionally valid was not required to be decided, the Solicitor-General having disavowed any such argument. The constitutional questions raised in written submissions were held not to arise once the statutory construction issues were resolved against the appellants.
The Court did not determine the outer limits of the "misapprehension" test in Autodesk where the later legislative change is so fundamental that it might be said to have altered the very basis upon which the original sentence was understood to operate. Spigelman CJ's references to respect for parliamentary power were disapproved in form but the result was upheld on narrower grounds, leaving some uncertainty as to how future courts should weigh legislative policy when exercising the reopening discretion. These questions await further litigation in which the precise statutory context differs from the 1997–2005 life-sentence regime considered here.