HKSAR v. HUNG CHAN WA

CACC 411/2003
CACC 61/2004

IN THE HIGH COURT OF THE

HONG KONG SPECIAL ADMINISTRATIVE REGION

COURT OF APPEAL

CRIMINAL APPEAL NOS. 411 OF 2003 AND 61 OF 2004

(ON APPEAL FROM HCCC NOS. 17 OF 2003 AND 211 OF 2003)

----------------------

CACC 411/2003

CACC 61/2004

BETWEEN

______________

----------------------

J U D G M E N T

----------------------

Hon Stock JA (giving the judgment of the Court):

Introduction

1. On 23 June 2005, this Court handed down judgment in these two appeals, but reserved for further argument a consequential issue raised by the respondent.  This judgment deals with that consequential issue.

2. Each of the appellants had been convicted of trafficking in dangerous drugs and sentenced to substantial terms of imprisonment.  The issue canvassed upon appeal was whether the presumptions raised by sections 47(1) and (2) of the Dangerous Drugs Ordinance, if read as imposing persuasive burdens of proof upon defendants, were compatible with the presumption of innocence to which constitutional protection is afforded by article 39 of the Basic Law and, if not compatible, whether it was possible, in accordance with applicable canons of construction, to read the provisions as imposing an evidential burden.  We decided that the subsections, if read as imposing persuasive burdens, were not compatible with the presumption of innocence thus protected, but that it was reasonably possible to construe the presumptions as ones that created an evidential onus only and that they should be so read.  Since at the trials of the appellants the juries were directed to apply persuasive burdens, we allowed the appeal of each appellant, quashed each conviction, and ordered that each appellant be retried for the offences alleged by the respective indictments.

3. In the course of argument, the court was invited to declare, in the event of a determination that a persuasive burden was in the context of these subsections unconstitutional, that the effect of the decision was prospective so that it would apply only to such cases as were determined thereafter; but not wholly prospective, meaning that we were asked both by the respondent as well as the appellants to enable the decision to enure to the benefit of these appellants as well as to other convicted persons still ‘in the system’, meaning those whose appeals were pending at the date of the decision as well as those whose time for filing an application for leave to appeal had not at the date of the decision expired.  Short notice had been given of this issue and written argument was rather scant, so we intimated that, should the appeal on the question of the persuasive burdens go against the respondent, we would hear full argument on another occasion with the benefit, if we saw fit, of submissions by an amicus curiae.

4. By reason of our determination on the constitutional issue, the respondent’s application to impose a temporal limit upon the effect of our decision falls now to be considered.  To this end, we have received full submissions from the respondent, and from Mr Yu SC, as amicus curiae.  Although this application does not affect the appellants, the Director of Legal Aid has nonetheless instructed counsel for the appellants in relation to this application and we have received submissions from counsel thus instructed.

Motive for this application

5. What motivates this application is clear enough.  The respondent fears that as a result of this Court's decision, thousands of previous convictions for trafficking in, and possession of, dangerous drugs will be the subject of challenge or attempted challenge.

6. The respondent has produced statistical tables and charts showing the number of convictions for such offences from 1969 to mid-1991, and from mid-1991 to 30 September 2005.  1969 is used as a starting date because that was the year in which the predecessor of section 47 was first enacted.  The relevance of 1991 is that it was in that year that the Hong Kong Bill of Rights Ordinance was enacted giving legislative effect to the International Covenant on Civil and Political Rights (“ICCPR”); and it is the protection of the presumption of innocence expressed by that Covenant that lies at the heart of our decision on the presumptions in section 47.  Figures have also been supplied for convictions since 1 July 1997, although we think that for present purposes it is the figure since 8 June 1991 that is relevant, since the constitutional protection of the right in question goes back to that date.  Over 12,000 persons have been convicted since then of trafficking in dangerous drugs, and over 86,000 of possession of dangerous drugs other than for the purpose of trafficking.  There are also figures for convictions for other offences for the proof of which statutory presumptions in the same terms as section 47 of the Dangerous Drugs Ordinance apply: section 24, Firearms and Ammunition Ordinance, Cap 234 and section 10, Control of Chemicals Ordinance, but the numbers are small.

7. The respondent contends that our decision as to the presumptions will be followed by applications for leave to appeal out of time against convictions entered prior to June 2005, the date of our judgment, by persons seeking to rely on the retrospective effect of a declaration of invalidity; in this case the declaration that application of a persuasive burden in such cases was unlawful.  The respondent proffers, as a suggested realistic scenario, the prospect of thousands of applications for leave to appeal out of time against conviction.  This, it is contended, will result in serious disruption to the administration of justice because the courts, prosecutors, defence counsel and the Legal Aid Department will be inundated with applications for leave to appeal out of time, so much so that the criminal justice system will grind to a halt.  Accordingly, this Court is invited to limit the temporal effect of its decision by making it prospective (save for those in time at the date of that decision) to ensure that those making such applications would have them dismissed in limine, utilising for this purpose the provisions of section 83S of the Criminal Procedure Ordinance, Cap 221.

8. The figures provided do not distinguish between those who pleaded guilty, on the one hand, and those, on the other, who were convicted after trial, although experience suggests, and Mr McCoy SC for the respondent agrees, that it is likely that the great proportion of convictions, certainly a majority, will have followed pleas of guilty.  Nor do the figures reveal in how many of the contested cases reliance was placed on the persuasive burden of proof, a burden which, according to our earlier judgment, was not properly applicable in such cases.  As we intimated in that judgment, there will have been many cases where such a presumption was not utilized.  Be that as it may, we are prepared, for the purpose of this application, to assume the possibility of a very significant number of applications for leave to appeal out of time. 

9. The respondent’s case goes further, however.  Not only does the respondent say that the device of prospective overruling should be used to avoid administrative chaos; but it goes on to contend that the court has no choice in the matter, for prospective overruling is mandated by article 160 of the Basic Law. 

Defining the application

10. The term ‘prospective overruling’ is imprecise if it is taken to mean, as it implies on its face, that the device to which it refers is utilized only where an earlier judicial decision has been overturned by a subsequent one; for the device is intended to cover also those situations in which a legislative provision is declared invalid or is interpreted by the courts, whether or not there has been a prior judicial interpretation or determination as to the validity of that provision, and where it is sought to limit the temporal effect of the decision.  The term is used ‘to cover any technique by which a court places an explicit limit on the retrospective effect of its decision’: see “Prospective Overruling” Dame Mary Arden (2004) 120 LQR 7.  The term is in a further sense not strictly accurate:

“The important point is not that such remedies, when granted, have prospective effect, as the vast majority of declarations are assumed to have this in any case.  The relevant point is that a ‘prospective declaration’ may be prospective only.  A more accurate term is therefore ‘non-retrospective declarations’.  But this description is too clumsy to adopt.”

Lord Woolf: “The Additional Responsibilities of the Judiciary in the New Millenium” The Clifford Chance Lectures (Hart Publishing, 2000) pp 142, 142-143.

11. In this context, it serves little purpose to engage upon an analysis of the degree to which the presumptions in sections 47(1) and (2) of the Dangerous Drugs Ordinance were previously studied by the courts for their constitutional validity.  We referred in our earlier judgment [paragraphs 92 et seq] to R v Sin Yau-ming [1992] 1 HKCLR 127 which suggested, without deciding, that the presumption now contained in section 47(2) of the Ordinance (then section 47(3)), if read as a persuasive burden, was acceptable.  Those comments were however obiter.  Section 47(3) (as it then was) was struck down by that decision on the footing that it was a presumption upon other presumptions that were themselves invalid.  We referred also to R v Lum Wai-ming (1992) 2 HKPLR 1782, and R v Chan Wai-ming (No 2) (1992) 2 HKPLR 231, both first instance decisions, saying that in so far as it was held that the presumptions (if persuasive) were valid, we disagreed.  The relevance for present purposes of those two decisions and our respectful disagreement with them, in so far as they touched upon the presumptions at issue in the cases before us, is that the respondent says that they are evidence of the previous understanding of the law upon which prosecutors have for many years relied.  To the question of reliance, we shall later return.  But for definitional purposes, it suffices to say that what we are asked to do is to limit the temporal effect of our decision; that is the sense intended in this judgment for the term “prospective overruling”.

12. Prospective overruling may take several forms.  In its pure form, it operates in the future only and does not even enure to the benefit of the parties to the proceedings in respect of which the new rule or the interpretation is declared.  Or it may take a modified form, whereby it has no retrospective effect save for the benefit of the parties and, perhaps, those whose cases or appeals are pending at the date of the declaration in question.  In the event, what this court is asked to do by the respondent is to make a declaration in a modified form.  That modification is particularised at paragraph 3 above: the present appellants have benefited from our determination, and we have preserved the position for those whose applications for leave to appeal were pending, or who were within time for lodging applications for leave to appeal, at the date of our decision.  In the course of the hearing of the present application, Mr McCoy has proposed yet a further modification by reason of which, if adopted, our determination would also operate retrospectively for those who could show exceptional circumstances “justifying or excusing the failure to apply within time”.  We shall later explain the genesis of that proposed extra modification.

The starting point

13. The starting point is that judgments normally have retrospective as well as prospective effect.  They have retrospective effect because judges necessarily are engaged upon the adjudication of past events and because in determining what law is to be applied to the past facts under consideration, judges ascertain and declare what the law was when those events took place.  The rights and obligations of the parties are then decided according to the law as it was at the time of those events.  Judgments also have prospective effect because of the role that precedent plays in the common law.  Legislatures, on the other hand, give prospective and not retrospective effect to the laws they enact.

14. In the interpretation of a statute, judges decide what the meaning of an enactment was at the date of its enactment and, by reason of the declaratory theory of judicial decisions, this applies even where there has been an earlier judgment that has provided a contrary interpretation which the later decision overrules; for the effect of the overruling is that the earlier decision did not represent the law: see Blackstone Commentaries on the Laws of England (1st edn, vol 1, p 70); and Kleinwort Benson Ltd v Lincoln County Council [1999] 2 AC 349, referred to by the Court of Final Appeal in Lau Kong Yung v Director of Immigration (1999) 2 HKCFAR 300, 326.  It is recognized that in deciding cases, judges may “on occasion develop the common law though as a general rule he does this ‘only interstitially’” (per Lord Goff in Kleinwort Benson Ltd at p 378B); and, further, that since societal factors that affect assessments of proportionality when judging the permissibility of infringements of rights are factors that may change with prevailing circumstances, the interpretation of statutory provisions that impinge on fundamental rights is itself liable to change.  That said, the starting point is that:

“... the interpretation [of a statute] the court gives an Act of Parliament is the meaning which, in legal concept, the statute has borne from the very day it went on to the statute book.”

Re Spectrum Plus Ltd: National Westminster Bank plc v Spectrum Plus Ltd and others [2005] 4 All ER 209, at 225, per Lord Nicholls.

15. In the present cases, this court has been engaged upon an exercise in statutory interpretation.  What we have done is to interpret statutory provisions in the light of constitutional rights protected by the Basic Law.  Those rights were constitutionally preserved on and from 1 July 1997, because that is the date upon which the Basic Law came into effect; although the protection of the presumption of innocence has, of course, a longer history than that: it was given constitutional and statutory protection in Hong Kong from 8 June 1991 by virtue of the Letters Patent and the Bill of Rights Ordinance, respectively.

16. It must follow, as a starting point, article 160 aside, that the effect of our decision in June 2005 is a determination of the meaning of the relevant provisions at the date of their enactment, which in this case was 1992. 

Article 160 Basic Law

17. Article 160 of the Basic Law provides that:

“Upon the establishment of the Hong Kong Special Administrative Region, the laws previously in force in Hong Kong shall be adopted as laws of the Region except for those which the Standing Committee of the National People's Congress declares to be in contravention of this Law.  If any laws are later discovered to be in contravention of this Law, they shall be amended or cease to have force in accordance with the procedure as prescribed by this Law.

Documents, certificates, contracts, and rights and obligations valid under the laws previously in force in Hong Kong shall continue to be valid and be recognized and protected by the Hong Kong Special Administrative Region, provided that they do not contravene this Law.”

18. The respondent’s case is that by reason of this provision of the Basic Law, prospective overruling is the required norm in those cases where it is decided after 1 July 1997 that a law enacted before 1 July 1997 contravenes the Basic Law.  That is not to suggest that it has been the norm in practice.  Indeed, the point has not been taken before.  Nonetheless, what is said is that on 23 February 1997 there was adopted by the Standing Committee of the National People's Congress a Decision on Treatment of the Laws previously in force in Hong Kong, designed to give effect to articles 8 and 160 of the Basic Law, and that annexed to the Decision was a list of Ordinances as well as particular statutory provisions previously in force that were thereby declared to be in contravention of the Basic Law and therefore not adopted as the laws of the Hong Kong Special Administrative Region.  None of the provisions of the Dangerous Drugs Ordinance enacted in 1992, with which we have been concerned, appears in those Annexes.  Whilst it is not suggested that the list purported to be an exhaustive one of laws that were in fact in contravention of the Basic Law, it is suggested that since the statutory provisions in question in the present cases were not there included, they must succumb to the fate of all laws not there listed but subsequently discovered to be in contravention of the Basic Law, which fate is to be found in the phrase in article 160: “… shall be amended or cease to have force” (emphasis added).  “Cease to have force” means, according to this argument, “cease to have force immediately after the date of that discovery”.  This means that until such time as a particular provision is discovered and declared to be inconsistent with the Basic Law, it remains in force.  Not only does the Basic Law not prohibit prospective overruling but, so it is argued, the language of article 160 mandates it.  It requires prospective overruling as the norm.

19. We do not agree with this argument.  There are fundamental problems with it, not only in relation to this particular case, but generally.

20. In relation to this particular case, the fundamental problem with the argument is that we have not found either statutory provision, that is to say either of the presumptions created by sections 47 (1) and 47(2) of the Dangerous Drugs Ordinance, to be inconsistent with the Basic Law.  What we have determined (see paragraph [137] of the earlier judgment) was that “in the face of competing available constructions, the legislature should be taken to have intended a provision compatible with the requirements of the Covenant”, and that it was reasonably possible to construe the words “until the contrary is proved” in sections 47(1) and (2) of the Dangerous Drugs Ordinance as imposing only an evidential burden upon an accused.  We said, at paragraph [143]:

“We are, in the event, satisfied that on the principles of interpretation which we are enjoined to apply, it is reasonably possible, and therefore incumbent upon us, to construe the presumptions enacted by sections 47(1) and (2) of the Dangerous Drugs Ordinance as presumptions that each create an evidential onus, and that the provisions are not to be regarded as imposing persuasive burdens of proof.”

In other words, we did not hold the relevant statutory provisions to be unconstitutional.  What we did was to assume an intent on the part of the legislature to enact provisions that were compatible with the ICCPR, to which domestic effect had been given before the date of enactment, and to construe them accordingly.  It follows that the premise upon which the article 160 argument is based does not arise in the present case.

21. It is said by Mr McCoy that that is artificial, that the reality is that this court has declared that in so far as the presumptions were on their face persuasive burdens, they contravened the Basic Law and that that was to assert or declare the invalidity of a previous law.  Even if, contrary to the view we express in paragraph [20] above, that were so, there arises a further obstacle to the article 160 argument as it applies to this case.  It is that our determination about the burdens imposed by section 47 must apply equally to the provisions as they stood before 1 July 1997, so that the supposed previous law – the persuasive burden – was not one that can properly be said to have been ‘a law previously in force’ upon the coming into effect of the Basic Law.  The provisions were enacted in 1992 and the particular persuasive burdens, if that is what the provisions required, were not then permissible, and the Letters Patent then in force precluded the enactment of such provisions since they contravened the ICCPR.  That such invalidity runs back to the date of enactment is clear.  This approach accords with the judgment of the Court of Final Appeal in Solicitor v Law Society of Hong Kong & Secretary for Justice [2004] 1 HKLRD 216.

22. As for the proposition generally, that article 160 mandates, as a norm, declarations with purely prospective effect, the proposition sits ill with the scheme of the Basic Law as a whole, and with any purposive construction of it, and for this purpose we assume a ruling that does declare unconstitutional a provision enacted prior to 1 July 1997:

(1)   The result of the respondent’s proposition, if correct, would be contrary to the remedy proposed.  What is proposed is that the effect of our earlier judgment should enure to the benefit of these appellants, as well as to others in respect of those convictions where time for appealing had not then expired.  But if the respondent is correct as to the proper reading of article 160, then it would have to follow that only pure prospective overruling was permitted.  Where would be the room for any retrospective effect at all?  Certainly not from article 160 itself.  If the construction of article 160 for which the respondent contends were a proper construction, no defendant could benefit, either at first instance or subsequently, from any declaration at his trial or upon his appeal that a statutory provision creating the offence for which he was on trial or convicted was unconstitutional.  To answer this point, Mr McCoy says that the requirement of article 160 is a norm, not an absolute, and that the courts have concurrent jurisdiction to modify the austere effect of the pure prospectivity that article 160, if wrongly viewed in isolation, would dictate.  The common law powers of judges are preserved by the Basic Law and inherent in those powers, he says, is the right, indeed the requirement, to fashion such remedies as are just.  We are not attracted by this answer.  If those who framed the Basic Law had in mind by article 160 that the norm of a declaration of invalidity would be prospective with powers, however, in a court to depart from that norm, they would have said so.

(2)   The ramifications of the proposition, if correct, would lead to a bizarre result.  The Article applies only to laws in force prior to 1 July 1997 that are subsequently discovered to be inconsistent with the Basic Law.  It does not relate to laws enacted after 1 July 1997.  Article 11 of the Basic Law provides that: “No law enacted by the legislature of the Hong Kong Special Administrative Region shall contravene this Law”; and article 39 itself provides that: “The rights and freedoms enjoyed by Hong Kong residents shall not be restricted unless as prescribed by law.  Such restrictions shall not contravene the provisions of the preceding paragraph of this Article”; the ‘preceding paragraph’ being that which gives constitutional effect to the ICCPR.  There can be no question but that any incompatible law enacted after 1 July 1997 is intended by the Basic Law to be ultra vires the power of the legislature, never to have acquired the force of law.  The respondent’s argument must therefore presuppose that in criminal cases, quite apart from non-criminal cases, involving constitutional points, liability will depend upon the date of enactment of the offence-creating provision.  So a defendant on trial in, say, 2002 might successfully challenge the constitutionality of a provision enacted on 2 July 1997 but not the very same one if enacted on 30 June 1997 since, according to the effect of the respondent’s argument, Article 160 preserves willy-nilly as good law any pre-July 1997 enactment that was not declared unconstitutional by the Standing Committee in February 1997, and preserves it as good law until such date as it is declared unconstitutional; whereas the Basic Law negates ab initio any unconstitutional law enacted in and after July 1997. 

(3)   In paragraph [21] above, we referred to Solicitor v Law Society of Hong Kong and applied its logic to the present cases, concluding that if a persuasive burden had ever been the law in the context of section 47, it was not a law in force on 1 July 1997.  Mr McCoy serves notice that the respondent will seek to re-litigate the point in the Court of Final Appeal, on the basis that the Court of Final Appeal was not asked to consider the issue in the light of article 160.  As to this, we would make only the following observation, that if the respondent were correct – that an invalid law not declared invalid before 1 July 1997 was to be treated as good law as at and after 1 July 1997 – the effect of the proposition, if good, becomes more curious still when one considers the pre-July 1997 position of any legislative provision that contravened the protections afforded by the ICCPR.  Article VII(3) of the Letters Patent, which had effect in Hong Kong from June 1991 until 30 June 1997, prohibited the enactment of laws after 8 June 1991 that restricted the rights and freedoms enjoyed in Hong Kong if the restriction was inconsistent with the ICCPR as applied to Hong Kong.  Prior to 1 July 1997, the effect of the provision was that any law enacted after 8 June 1991, which was also the effective date of the Bill of Rights Ordinance, was, to the extent of an inconsistency with the ICCPR, rendered nugatory ab initio.  The respondent’s submissions, if correct, would mean that a provision that was not ICCPR-compliant had no lawful effect prior to 1 July 1997, but that lawful effect was conferred upon the provision on 1 July 1997, such lawful effect to continue until such time as it was declared unconstitutional.  To illustrate the practical result of this, let us transpose to 1997 the circumstances of Mr Lee Kwong-kut (AG of Hong Kong v Lee Kwong-kut [1993] AC 951) from 1990, which was the date of the alleged offence by him under section 30 of the Summary Offences Ordinance.  Suppose the acts upon which that prosecution was based had been committed by him on 30 June 1997, and suppose no prior challenge to section 30, but a challenge at his trial.  He would then be acquitted because we can legitimately suppose a determination that at the date of his offence, section 30 was not good law.  Suppose, however, that the acts in question had instead been committed on 1 July.  Well, in that case, according to the respondent’s theory, until such time as there was a challenge and a finding of unconstitutionality, the law would be treated after 1 July 1997 as valid and Mr Lee would be guilty.

(4)   The respondent’s contention is at odds with other provisions of the Basic Law.  Article 8 of the Basic Law provides that:

“The laws previously in force in Hong Kong, that is, the common law, rules of equity, ordinances, subordinate legislation and customary law shall be maintained, except for any that contravene this Law… .” (Emphasis added).

Article 18 provides that:

“The laws in force in the Hong Kong Special Administrative Region shall be this Law, the laws previously in force in Hong Kong as provided for in article 8 of this Law, and the laws enacted by the legislature of the region.” (Emphasis added).

These provisions mean what they say, namely, that legislation that does contravene the Basic Law shall not be maintained after 1 July 1997. 

Articles 8 and 18 were reflected in paragraph 1 of the Standing Committee’s Decision of 23 February 1997:

“The laws previously in force in Hong Kong, which include the common law, rules of equity, ordinances, subordinate legislation and customary law, except for those which are in contravention of the Basic Law, are adopted as the laws of the Hong Kong Special Administrative Region.” (Emphasis added).

The meaning of that is clear enough.  It covers all non-compliant laws, not just those separately specified in paragraphs 2 and 3 of, and Annexes 1 and 2 to, that Decision. Domestic effect was given to that Decision by section 7 Hong Kong Reunification Ordinance:

“The laws previously in force in Hong Kong, that is the common law, rules of equity, ordinances, subsidiary legislation and customary law, which have been adopted as the laws of the HKSAR, shall continue to apply.” (emphasis added).

It follows that non-compliant laws did not then become part of our laws.

(5)   The argument ignores the continuum that was self-evidently intended by the Basic Law for the preservation of fundamental rights, as well as the very purpose of Chapter III of the Basic Law, an intention to which the Chief Justice referred in Solicitor v Law Society of Hong Kong and Secretary for Justice [2004] 1 HKLRDR 214, 221 as “a most important theme of the Joint Declaration and the Basic Law”.  See also Secretary for Justice v Lau Kwok Fai [2005] 3 HKLRD 88, 101G-H per Sir Anthony Mason NPJ.  Moreover, the “…intention of the Basic Law was to entrench constitutionally the rights and freedoms in Chap III [of the Basic Law]”: Gurung Kesh Bahadur v Director of Immigration (2002) 5 HKCFAR 480, 491.  It would be an odd exercise in entrenchment that envisaged a suspension of those very rights until some unascertainable time when a particular statutory provision happened to fall for judicial scrutiny for its compatibility with the Basic Law.

(6)   The argument ignores as well the context in which article 160 was promulgated, as well as the phrase ‘in accordance with the procedure as prescribed by this Law’, which appears immediately after the words ‘shall be amended or cease to have force’.  In order to give effect to those articles of the Basic Law, such as article 8, which mandated the continuation of laws across the reunification dateline, a mechanism for adoption of laws consistent with the Basic Law was necessary.  The mechanism utilized was the NPC Decision of 23 February 1997 and, domestically, section 7 of the Reunification Ordinance.  Article 17 of the Basic Law vested the Region with legislative power and a procedure was there laid down for scrutiny of laws enacted by the Region’s legislature.  Article 73 of the Basic Law conferred upon the legislature power to enact, amend or repeal laws, with further stipulations under other articles for the work of the legislature, such as the constitution of a quorum and rules of procedure (article 75) and the method by which a bill passed was to take effect (article 76).  It was, in our judgment, envisaged by article 160 that since the pre-1997 scrutiny of laws was not intended to be the last word on compatibility, future legislative action might be necessary to amend or to repeal laws subsequently discovered to be in violation of the Basic Law.  If that were to happen, action would then be taken in accordance with the procedures defined by the Basic Law.  It is to such procedures that article 160 refers.  It could hardly have been intended that the discovery to which reference is made in article 160 was discovery by the courts, or that the ‘procedures’ envisaged were judgments of the courts.  If judicial ‘discovery’ was intended, at what stage was such discovery to take effect and result in the death of the infringing law?  Would a law cease to have effect when a magistrate declared it to be inconsistent with the Basic Law only to revive on appeal, if that finding were held to be incorrect?  Sections 23 and 27 of the Interpretation and General Clauses Ordinance were drawn to our attention to highlight the suggested prospective effect of an article 160 discovery.  Those sections do not affect the matter since in our judgment article 160 has nothing to do with judgments of the courts. Whatever the temporal effect of article 160 and of sections 23 and 27 of the Interpretation and General Clauses Ordinance, it makes no sense, in our opinion, to conclude that they apply to judicial decisions.

Article 160 apart 

23. We are thus brought back to the starting point, which is that an interpretation of a statute constitutes a determination of its meaning at the date of its enactment.  Similarly, a determination that a statutory provision is unconstitutional means that the law was ultra vires the legislature and never acquired legal force and effect.  It is therefore incumbent upon the respondent, if it is to succeed in this application, to demonstrate a constitutional provision, or a statutory provision or other rule of law, that would permit this court to limit the temporal effect of its decision; and if there is a such a power, the question would be whether the power should in this case be exercised.

24. Some constitutions, such as the South African Constitution, expressly permit temporal limitation by the courts upon the effect of a declaration of constitutional invalidity.  The Basic Law does not express such a power.  Nor does the respondent point to any statutory instrument that empowers this court to do as it asks.  The respondent’s case is that the power to order that the effect of its determination be limited to the future is part of the inherent jurisdiction of the court, a jurisdiction that includes an ability to supply such judicial remedies as it finds are in the interests of justice.  Mr McCoy points to jurisprudence in some jurisdictions that evidences a willingness to use the technique of prospective overruling.  To that jurisprudence we shall shortly turn, but it suffices at this point to say that the examples are few, the approach decidedly cautious, the circumstances in which such jurisdictions are willing to countenance use of the technique are exceptional and, even then, the courts are entirely reluctant to apply it where to do so would affect the liberty of the subject.  Against that backdrop, the respondent’s stance is bold, for the suggestion is made in terms that in Hong Kong temporal limitation by the courts upon the effect of declarations of invalidity is appropriate in every case, whether criminal or not.  This line does not depend on article 160, for it is said that this is the approach that the courts here should adopt in any event.

25. We are satisfied that in this case, it is not necessary to decide whether temporal limitation upon declarations of invalidity of laws, or upon the effect of judicial interpretation of a Statute, is or is not part of the remedial armoury available to the courts in this jurisdiction.  That is because we are satisfied that, even if it were available, there is no conceivable justification for extending the parameters of its use beyond the exceptional circumstances that have been delineated by jurisprudence elsewhere, and that this case falls well outside those parameters.

Other common law jurisdictions

26. The courts of the United Kingdom have, until very recently, set their face against the technique of judicial prospective overruling.  Particularly where the issue is one of statutory interpretation, the device has been seen as inconsonant with the judicial function which is to interpret and apply, not to legislate.  Thus it was that as recently as 1998, Lord Goff said in Kleinwort Benson at 379:

“... such a system [prospective overruling], although it has occasionally been adopted elsewhere with, I understand, somewhat controversial results, has no place in our legal system”.

27. Although in that jurisdiction a change of attitude is heralded by the decision of the House of Lords in National Westminster Bank plc v Spectrum Plus Ltd and others [2005] 2 AC 680, it remains the fact, so far as we have been informed, that the device has not there been used.  It is unnecessary here to rehearse the facts of the case, save to record that it was a civil action.  The effect of the judgment of the majority was, as the headnote relates, that there could be cases:

“… where a decision on an issue of law was unavoidable but the decision would have such gravely unfair and disruptive consequences for past transactions and happenings that the House of Lords would be compelled to depart from the normal principles relating to the retrospective and prospective effect of court decisions”;

although indicating clearly that were that country’s supreme court to follow such a course, it would be a course adopted “altogether exceptionally”.  Even in relation to statute law, the majority did not exclude the possibility of a justifiable use of the technique although Lord Scott and Lord Steyn disagreed, Lord Scott commenting, p 726, para [125] that:

“Where an interpretation and application of a statute is the issue, a prospective ruling would, absent legislative authority… appear to constitute an improper usurpation by the judiciary of the role of the legislature.”

Dame Mary Arden, at p 11 of the Law Quarterly Review article which we cite at paragraph [10] above, has also suggested that the technique should not be used ‘where the decision turns purely on the construction of a statute.  The question whether there should be relief in respect of past breaches of an Act of Parliament must surely be a matter for Parliament itself.  Likewise the concept probably cannot in general be invoked so as to prevent a person from claiming a violation of human rights.’

28. In Spectrum, Lord Nicholls summarized overseas experience of the device.  He noted, at p 693 paras [18-19], that in other common law countries the technique had taken root only in the United States of America and India and that in the United States its fortunes had “waxed and waned”.  He noted, too, that the United States Supreme Court has retreated from its previous limited adoption of the rule, and for our purposes it is particularly noteworthy that in Griffith v Kentucky (1987) 479 US 314 ‘the court abandoned prospective overruling when directly reviewing criminal cases’.  The position in India assists the respondent’s stance little, since India finds its jurisdiction in this regard in express terms in the Indian Constitution.  Canada has rejected the technique: see Re Edward and Edward (1987) 39 DLR (4th) 654, although it has invoked the doctrine of state necessity to prevent a legal vacuum: Reference re Manitoba Language Rights under the Manitoba Act 1870 (1985) 19 DLR (4th) 1.  The European Court of Justice has utilized the device on occasions, but it will do so only exceptionally: R (on the application of Bidar) v Ealing LBC Case C-209/ 03 [2005] QB 812, whereas the European Court of Human Rights has in particular circumstances relieved States parties from ‘re-opening legal acts antedating the delivery of its judgment.’.  The High Court of Australia has firmly rejected the concept as a judicial device, holding that the Court has no power to overrule cases prospectively: see Ha v State of New South Wales, Walter Hammond & Associates v State of New South Wales (1997) 146 ALR 355. 

29. To this list we would add South Africa, but there the power is exercised pursuant to a specific provision of the Constitution.  There are other instances, for example, the Scotland Act 1998, section 102, by which power is conferred by legislation to limit the retrospective effect of a declaration of invalidity of a statute.  The very fact of these few instances of specific legislative empowerment to use the technique evidences the caution with which the notion of its use, absent such provisions, is approached.  Yet we are urged to say that in Hong Kong it should be used almost as a matter of course. 

The circumstances of its use, where used

30. A study of the circumstances in which those jurisdictions that have used the technique have been prepared to do so, reveals that it has been used exceptionally and that when used, its objective has been to protect those who have entered upon transactions giving rise to rights and obligations and who have, in doing so, genuinely relied upon that which was at the time fairly taken to be settled law, and where to reverse the consequences of that reliance would cause them undue hardship.  What is further demonstrated is that courts recoil from its use where liberty is at stake.

31. The care to be taken not to create injustice by overruling previous decisions reasonably perceived to be settled law was alluded to by Lord Reid, albeit in the context of an analysis that pointed to the dangers of judge-made law:

“And there is another sphere where we have got to be very careful.  People rely on the certainty of the law in settling their affairs, in particular in making contracts or settlements.  It would be very wrong if judges were to disregard or innovate on what can fairly be regarded as settled law in matters of this kind.  When Parliament passes an act there is always objection to any proposal to make it retrospective.  But judge-made law is always retrospective.  We cannot say that the law until yesterday was one thing, from tomorrow it will be something different.  That would indeed be legislating.”

“The Judge as Law Maker” (1972-1973) 12 JSPTL 22, 23.

32. Although the United States has retreated from the technique, that is the jurisdiction in which its use was, albeit limited, more prevalent than elsewhere.  In a review of that earlier jurisprudence, Professor Traynor, formerly Chief Justice of California and Sir Arthur Goodhart Professor of Legal Science, University of Cambridge, summarized the objective of the decisions he reviewed:

“They have a common objective, to rid the law of an unsound rule and at the same time preclude undue hardship to a party that has justifiably relied on it.  Reliance plays its heaviest role in such areas as property, contracts, and taxation, where lawyers advise clients extensively and in their planning on the basis of existing precedents.  In most of the decisions I shall now review briefly, overruling was the rational judgment and prospective overruling was essential to preclude the injustice that retroactive application of the new rule would having entailed.”

“Quo Vadis, Prospective Overruling: A Question of Judicial Responsibility” (1977) 28 Hastings Law Journal 533, 543.

33. That review included reference to judgments that approved “the prospective application for decisions overruling settled rules of property or contract rights, on the ground that the ‘highest principles of justice’ militate against the investment of rights acquired in reliance on prior law” (p 543, in reference to Hill v Atlantic & N.C.R.R., 143 N.C. 539, 55 S.R. 854 (1906)); a decision of the Supreme Court of Ohio in Bingham v Miller 17 Ohio 445, 49 Am Dec 471 (1848) in which the normal retroactive effect of the decision would have invalidated legislative divorces that had been granted in the course of the previous forty years with the consequence, had there not been a temporal limitation on the effect of the decision, of rendering illegitimate those born of marriages in respect of which the decrees had been granted; and Hare v General Contract Purchasing Corp. 220 Ark 601 (1952) in which a decision of the Supreme Court of Arkansas reversed the court’s previous decision that had held a form of instalment contract not to be usurious but, because of widespread reliance on the previous decision, restricted the effect of the new decision prospectively.

34. In so far as Spectrum has hinted at the possibility of utilizing the technique in exceptional circumstances, it is, again, the injustice that would flow from undoing rights acquired in reliance on settled law that is stressed as the basis for its exceptional use:

“There could be cases where a decision on an issue of law, whether common law or statute law, was unavoidable but the decision would have such gravely unfair and disruptive consequences for past transactions or happenings that this House would be compelled to depart from the normal principles relating to the retrospective and prospective effect of court decisions.”

per Lord Nicholls at p 699, para. [40].

And this suggested premise for its use is an echo of the position of the European Court of Justice:

“The court has taken that step [of limiting the temporal effect of a ruling] only in quite specific circumstances, where there was a risk of serious economic repercussions owing in particular to the large number of legal relationships entered into in good faith on the basis of rules considered to be validly in force and where it appeared that both individuals and national authorities had been led into adopting practices which did not comply with Community legislation. …”

R v (Bidar) v Ealing London Borough Council at p 846H-847A, para 69.  See also Zamir & Woolf “The Declaratory Judgment’ (3rd ed. 2001 at 3.047, footnote 48); and Defrenne v Sabena [1981] 1 All ER 122, 137.

35. Even then, there is a caveat to be sounded on the issue of reliance, relevant because part of the respondent’s case in this application rests on suggested reliance on the section 47 presumptions as their effect had been understood.  The dependency, the expectation, that the law has truly been settled must be a genuine and justifiable dependence and expectation for, if that were not so, what was supposedly relied upon was not certain and if it was not final, a complaint that there has been an inroad upon the principle of certainty and finality is not well founded.

36. The European Court of Human Rights has recognized that in certain human rights cases, relief from the normal retrospective effect of an interpretation may be a necessary corollary to the fact that certain societal standards and needs move with the times, that such cases require an evolutive approach, and that the constitutional instrument against which the issue at hand is tested is a living instrument.  In ‘giving life to the abstract statements of fundamental rights… judges … will naturally be guided by what are thought to be the requirements of a just society in their own time.  In so doing they are not performing a legislative function. … On the contrary, they are applying the language of these provisions of the Constitution according to their true meaning’: Boyce v The Queen [2004] 3 WLR 786, 795.  The thinking is that within such a framework, an assessment that gives proper recognition to current societal attitudes and needs should not in its effect relate back to a time when standards and needs were materially different and when an assessment would, for that reason, then have reached a different conclusion.

37. In Marckx v Belgium (1979) 2 EHRR 330, the Court was engaged upon a determination of the permissibility of Belgian laws that drew a distinction between the rights of ‘illegitimate’ children and their mothers, and of ‘legitimate’ children and their mothers.  The Court said at p 346 that:

“It is true that, at the time when the convention of 4 November 1950 was drafted, it was regarded as permissible and normal in many European countries to draw a distinction in this area between “illegitimate” and the “legitimate” family.  However, the court recalls that this Convention must be interpreted in the light of present-day conditions.  In the instant case, the court cannot but be struck by the fact that the domestic law of the great majority of the member states of the Council of Europe has evolved and is continuing to evolve, in company with the relevant international instruments, towards full juridical recognition of the maxim mater simper certa est.”

And, at p 353, in relation to the temporal effect of the judgment:

“The Court’s judgment is essentially declaratory and leaves to the State the choice of the means to be utilised in its domestic legal system for performance of its obligation under Article 53.  Nonetheless, it remains true that the Government has an evident interest in knowing the temporal effect of the present Judgment.  On this question, reliance has to be placed on two general principles of law which were recently recalled by the Court of Justice of the European Communities: “the practical consequences of any judicial decision must be carefully taken into account”, but “it would be impossible to go so far as to diminish the objectivity of the law and compromise its future application on the ground of the possible repercussions which might result, as regards the past, from such a judicial decision”.  The European Court of Human Rights interprets the Convention in the light of present-day conditions but it is not unaware of the differences of treatment between “illegitimate” and “legitimate” children, for example in the matter of patrimonial rights, were for many years regarded as permissible and normal in a large number of Contracting States. … Evolution towards equality has been slow and reliance on the Convention to accelerate this evolution was apparently contemplated at a rather late stage.  ...   Having regard to all the circumstances, the principle of legal certainty, which is necessarily inherent in the law of the Convention as in Community Law, dispenses the Belgian State from re-opening legal practice or situations that antedate the delivery of the present Judgment.  Moreover, a similar solution is found in certain contracting states having a Constitutional Court: their public law limits the retrospective effect of those decisions of that court that annul legislation.”

38. A further example of the approach taken by the European Court of Human Rights which does not confine itself to identifying the meaning properly to be given to the Convention at the time it came into force, is referred to in Spectrum at P 695, paragraph [25]:

‘A notable instance of this “dynamic and evolutive” approach to interpretation can be found in the succession of cases relating to recognition of the rights of transsexual persons, culminating in the decision in Goodwin v UK (2002) 13 BHRC 120.  In Goodwin v UK the court held that the United Kingdom could “no longer claim” that the matter fell within its margin of appreciation and that the fair balance inherent in the Convention “now” tilted in favour of the applicant (see (2002) 13 BHRC 120 at 144 (para 93).).  Running through the court’s reasoning is an acceptance that the earlier, contrary decisions of the court remained correct statements of the interpretation and application of the convention when they were given.  Consistently with this the court held (at 148 (para 120)) that the finding of violation “with the consequences which will ensue for the future” was just satisfaction.’   

39. There is a further category of case, wholly exceptional, where a limit or stay has been placed upon the effect of a declaration of invalidity by the device of suspending its effect, where the objective of this variant is to avoid chaos or extraordinary administrative dislocation that would occur but for temporal limitation.  Such a case was Reference re Manitoba Language Rights under the Manitoba Act 1870 (1985) 19 DLR (4th) 1, where the Supreme Court of Canada declared legislation that had not been published and printed bilingually to be unconstitutional and where the Court invoked the doctrine of state necessity to enable it to deem the laws temporarily valid pending corrective legislative action, so as to avoid the legal vacuum and chaos that otherwise would ensue.  See also Brown v Board of Education 347 U.S. 483 (1954); Schacter et al. v Attorney General for Ontario 93 DLR (4th) 1; Nielsen v Canada (Employment and Immigration Commission) [1997] 3 FC 920. 

40. It is pertinent to note the range of remedies available to the Canadian court, according to Schacter, upon a finding of violation of the Canadian Charter: striking down the provision; striking down but temporarily suspending the declaration of invalidity; reading down; or reading in. It is pertinent, we say, because none of those remedies constitutes the stark remedy of a declaration of non-retroactivity which the respondent suggests is within our power to grant, despite the wide terms of section 24 of the Canadian Charter of Rights and Freedoms which affords to those whose rights and freedoms have been infringed ‘such remedy as the court considers appropriate and just in the circumstances’, and despite the fact that such a provision is not reflected by any express power conferred on our courts.  In this regard, the respondent has resorted, as it has in the past, to section 6 of the Bill of Rights Ordinance which, it contends, promotes flexibility in the development of judicial remedies.  We do not agree with that interpretation of Section 6.  Section 6 states that the court in a successful Bill of Rights challenge ‘may grant such remedy or relief … as it has power to grant or make in those proceedings’ (emphasis added).  Section 6 does not confer new powers.  It does not confer powers at large.  One has only to look at section 24 of the Canadian Charter to see what device could have been crafted had that been intended.  

41. There are also cases where the courts seek to design rules of practice for future operation by judges or by law enforcement authorities.  A question sometimes arises as to the effect of such decisions on past cases: see, for example, George Savvas (1991) 55 A Crim R 241.  It is unnecessary to delve into those, for the present cases are not examples in that category.

An exceptional technique

42. Even in those jurisdictions that have used, or have been prepared to countenance use of, the technique so as to preserve rights accrued in genuine and justifiable reliance upon that which has been viewed as settled law, the emphasis has been upon the exceptional nature of the remedy:

(1)   “Only occasionally will there be cases that clearly demand this technique”: Traynor, Hastings Law Journal, Vol 28 January 1977, P 542.

(2)   Lord Nicholls in Spectrum, at P 226 para [41], envisaged that if the technique were to be used, it is one that would be used ‘altogether exceptionally’ and, in relation to the device used in the Manitoba language case, he referred to it as one used in ‘extreme circumstances’. 

(3)   “It is only exceptionally that the court may, in application of the general principle of legal certainty inherent in the Community legal order, be moved to restrict the possibility for any person concerned to rely on a provision which has been interpreted with a view to calling into question legal relationships established in good faith …”: Grzelczyk v Centre public d’aide sociale Ottignies- Louvain-la-Neuve [2002] ICR 566, 600-601, referred to in Bidar, above, at p 1102. 

Criminal Cases

43. Not only does the respondent seek to persuade this court to draw our recent interpretation of the provisions in section 47 (or as the respondent would have it, our recent declaration of invalidity) within the net of temporal limitation when, where used at all, it is used exceptionally, but seeks to do so where, by its use, the liberty of those affected is at stake.  Going this far goes entirely against the grain:

(1)   “If ever there was a case where the declaratory theory should be applied it must surely be one where the liberty of subject is in issue – as it plainly is where the point relates to the entitlement of the subject to be released from custody”: R v Governor of Brockhill Prison Ex p Evans (No 2) [2001] 2 AC 19, 37.

(2)  In rejecting the technique generally, the High Court of Australia in Ha, p 504, remarked that it would be a perversion of judicial power to maintain in force that which was acknowledged not to be the law and added: “This would be especially so where … non-compliance with a properly impugned statute exposes a person to criminal prosecution”. 

(3)  “Whatever may be the function of prospect overruling, or the prospective announcement of new rules of law made by judges, it is not acceptable that they should operate in a field of law which affects liberty and where a risk of serious injustice exists”: per Kirby P in Savvas, above, p 292.

(4)   “Born out of a laudable wish to mitigate the seeming unfairness of a retrospective change in the law, prospective overruling can beget unfairness of its own. … This is most marked in criminal cases, where ‘pure’ prospective overruling would leave a successful defendant languishing in prison.  ‘Selective’ prospective overruling avoids this consequence but it could see a successful defendant freed while others in like case stayed in prison.” per Lord Nicholls in Spectrum, p 696, paras [26-27]. 

(5)   Even the rule in Linkletter v Walker (1965) 381 US 618, the United States decision that gave the green light to limits on retrospectivity in certain criminal cases – the decision since altered by Griffith v Kentucky (1987) 107 S Ct 708 – did so in relation to new constitutional rules of criminal procedure, including court procedure as well as police practice.  It was in those circumstances that the Supreme Court “invoked as a test ‘no likelihood of unreliability ’in the fact-finding process.  Given reliability, the Court was free to weigh official reliance and the advantages of orderly transition [to the new rule] against the usual factors in favor of retroactive application of judicial rules.”: Traynor, above, p 557.  In the same article, an article that preceded Griffith, he added, at p 559: “There is wisdom in the Linkletter … [test] that preclude[s] retroactivity to the extent of upsetting final judgments, when there is no appreciable risk that innocent defendants have been convicted.  Without such risk, the doctrine of res judicata, which precludes collateral attack on final judgments in civil cases, should apply with equal force in criminal cases.  Precluding complete retroactivity by adherence to conventional doctrine does not present any question of prospective overruling.  The new rule is applied retroactively within the conventional limits that have traditionally governed the application of new rules.”

44. In this context, it is timely to recall from para [104] of our earlier judgment the ramifications of the persuasive burdens with which that judgment was concerned:

“The consequence of a reverse onus provision such as that provided by section 47(1) of the Dangerous Drugs Ordinance, if read as a persuasive burden, is that an accused may be convicted by a jury that concludes that his account as to a central element of the offence… is as likely as not to be true.  That is a sobering thought, not least when one recognizes that the prescribed punishments are in the most severe category.”

Applying the principles to this case

45. Judicial power to engage the technique of prospective overruling is not, for the reasons we have provided, mandated by the Basic Law.  Nor is there any statutory provision that empowers the courts of Hong Kong to travel that route.  It is unnecessary for the purpose of the reverse onus provisions that have been under examination to decide whether we are empowered at common law to embrace the technique, either at all or in cases of an interpretation of the kind upon which we found it necessary to embark, or in cases where a statutory provision has been ruled inconsistent with the Basic Law.  As we have earlier intimated, such a determination is unnecessary since, even if the power existed and even if our determination constituted a declaration of incompatibility, we are satisfied that such a power would be one exercisable only in exceptional circumstances and not in a case of this kind.

46. Proceeding on the assumption that the power to invoke the technique exists – and we emphasize the hypothetical basis upon which we proceed – we fail to see whence comes the proposition that in Hong Kong the technique should be employed whenever a court in this jurisdiction determines a statutory provision to be unconstitutional, and that it should be employed even in criminal cases.  It is, with respect, a remarkable proposition.  The proposition would take us from the established starting point of retroactivity, well embedded in our law, to a point altogether at the other end of the spectrum, and entirely beyond anything heretofore embraced, indeed even hinted at, in other common law jurisdictions.

47. In so far as the respondent contends that there has been reliance upon the law as previously understood, we see, for present purposes, no weight in the contention; even ignoring the question whether reliance by prosecutorial authorities and the courts could evercarry weight in relation to issues going directly to the question of guilt or innocence.  It is, in our judgment, a suggested reliance entirely outwith the category and type of reliance to which we have earlier referred as might constitute grounds for limiting the temporal effect of a judicial decision.  The principle of reliance is intended to benefit those who rely on the certainty of the law in settling their affairs, and temporal limitation in such cases is intended to avoid the grave hardship to them that would otherwise result.  It is entirely incongruous to apply this concept to the safety that prosecutors and courts may have felt in applying a reverse onus provision in criminal cases.  In any event, the reliance principle, where it does apply, must be genuine and must arise in respect of that which can reasonably be viewed as settled law.  What we have in the present case is a reverse onus provision, and it must be well known to anyone versed in human rights litigation that, absent onus-specific determinations by the highest court, reverse onus provisions are likely, at some time or another, to be the subject of challenge.  It is, in our judgment, idle to suggest – at any time, let alone in the early years of the development of human rights jurisprudence in a particular jurisdiction – that the law has been viewed as settled by reason of two first instance judgments not taken on appeal, and because of comments made obiter in one judgment of the Court of Appeal.  In this regard, it is relevant to note the reminder by Lord Scott in Spectrum, at p. 725, para [122] that “… the rulings of lower courts on points of law do not settle the law with finality.  They never have done.”  To the same effect, see Lord Hope at p 700, para [43].

48. We have referred in this judgment to a category of case recognised by the European Court of Human Rights and referred to in Spectrum, where, by virtue of the subject matter, courts are called upon to accord human rights instruments a “dynamic” interpretation.  The domestic courts of the European Community may accordingly be required to give legislation a changed meaning: see Spectrum, pp 698-699, para [37].  Even were we to assume that decisions of the courts of Hong Kong, through the application of a necessarily evolutive approach to a particular societal problem, may require, exceptionally, a temporal limitation upon the effect of a “new” decision, the examples revealed by the European cases are far removed from this case.  Ours was not an evolutive interpretation of the kind engaged by the European cases which we have mentioned.  What we are talking about in the two cases before us is a time span between June 1991 and 2005 in the context of the presumption of innocence.  We do not accept that in this context, standards, or the factors at play in the application of the proportionality test, have changed.

49. In so far as the respondent suggests that a failure to limit the temporal effect of our interpretation or, as the respondent would have it, our striking down, will halt the administration of justice by reason of a consequential flood of applications for leave to appeal out of time, this too is a line of argument that, in the present context, we find unpersuasive.  First, the evidence before us does not support the conclusion we are invited to draw.  No one can say how many applications are likely to be made, and the figures we have been given as to the number of cases since 1991 take no account of the fact that the majority will have been cases in which the defendant pleaded guilty.  Furthermore, the figures take no account of the fact that in many, perhaps most, of the remaining cases no use will have been made of the reverse onuses in question.  Nor does the doomsday scenario sketched by the respondent take account of the fact that all such applications as are made for leave to appeal against conviction, apart from those already ‘in the system’ at the date of our judgment or who have filed one within 28 days thereafter, will be applications out of time.

Applications out of time

50. What the respondent seeks by the course urged upon us is, at one swoop, to exclude from future judicial consideration – save for a limited exception to which we shall shortly turn – each and every out of time application in so far as it prays in aid the decision of 23 June 2005, regardless of the circumstances of the case, regardless of the fact that our law entitles applications to be made for leave to appeal against conviction to be made out of time, and regardless of the fact that there is a body of law that addresses the principles to be applied when such leave is sought because of a ‘new’ interpretation of the law. 

51. The essence of Mr. McCoy’s submission in this regard was this: that the effect of the jurisprudence of common law jurisdictions was that a change in the law, or in the settled understanding or interpretation of the law, will never in itself constitute a ground for granting an applicant leave to appeal out of time against a conviction entered before the ‘change’ was declared.  That being so, the argument goes, the temporal limitation that we are asked to provide will merely give effect to the practical consequence of that principle, yet will preclude a flood of hopeless applications that will otherwise be made by those hoping to find some loophole or fanciful exception to the limitation he invites us to impose.  However, in order to cater for those who have failed to lodge their applications in time because of some understandable mishap, a further category of those who should benefit from our judgment of June 2005 should, he says, be included in the declaration that the respondent asks this court to make.  The declaration that we are asked to make is that, subject to certain stated exceptions, the ‘new’ interpretation of section 47 of the Dangerous Drugs Ordinance ‘applies only from the next day subsequent to the delivery of the judgment of 23 June 2005’.  The proposed declaration goes on to include as beneficiaries of the interpretation the present appellants, those within time for applying for leave, and ‘any applicant or appellant who after 23 June 2005 applies for leave to appeal out of time, where there are exceptional circumstances justifying or excusing the failure to apply within time, on a prospective ground of appeal that depends on the new interpretation’; adding that ‘it is further declared that [for the purpose of this last category], the fact of the new interpretation is not in itself an exceptional circumstance’.

52. We were treated to a detailed analysis of decisions in other jurisdictions, as well as one in this jurisdiction, that seek to delineate the circumstances in which, if at all, a previous misconception about the law might afford a proper ground for extension of time in which to appeal in criminal cases.  For reasons upon which we shall shortly touch, this is not the appropriate place or time for a detailed rehearsal of principle, or for a determination by us of the principles this court will apply in such out of time applications as might follow our 2005 judgment.  It will suffice to give some idea of the approach taken elsewhere and in the Hong Kong case which is said by the respondent to have been wrongly decided.

53. The Canadian position is somewhat rigid.  The test that there prevails is:

“… whether or not the accused is still in the judicial system. … this test affords a means of striking a balance between the ‘wholly impractical dream of providing perfect justice to all those convicted under the overruled authority and the practical necessity of having some finality in the criminal process’”.

R v Wigman [1987] 1 S.C.R. 246, 257.

The application of the test was explained in R v Thomas [1990] 1 S.C.R. 713, 716:

“To be in the judicial system one of the following must apply:

1. an appeal has been launched to this Court;

2. an application for leave has been made within the time;

3. an application for an extension of time is granted based on the criteria that normally apply in such cases.

In a case in which the applicant alleges that he or she was convicted under a provision which has since been declared invalid, he or she should not be placed in a worse position than any other applicant.  On the other hand, since we cannot do perfect justice, the applicant should not artificially be brought into the system.

In the circumstances of this case, applying conventional principles, an extension of time should not be granted because:

(a)   there was no intention established to appeal within the time; and

(b)   the delay has not been adequately explained.

Accordingly, the application for an extension of time is dismissed.”

It is noteworthy that the respondent’s position is that the Canadian approach is hard-edged and in so far as it does not appear to make allowance for exceptional circumstances, it is not an approach that the respondent invites our courts to adopt.  That in itself is a significant concession, for it sits uncomfortably with the temporal declaration that we are asked to make.

54. The Australian approach is to be found in R v Unger [1977] 2 NSWLR 990 (following Victorian Stevedoring and General Contracting Co Pty Ltd v Dignan (1936) 46 CLR 73):

“There is no difference in principle between a subsequent judicial decision which has the effect of exposing a prior misconception in relation to the principle of law which was wrongly regarded as well founded at the time of the trial, and a subsequent judicial decision exposing the invalidity of regulations that were wrongly treated as valid at the time of the trial.  The trial having been concluded and the time for appeal having gone by, the general principle is that the matter is regarded as at an end.  It is to be borne in mind that the effect of a conviction in a criminal court, no less than a verdict and judgment in a civil court, is to merge in that conviction or judgment, as the case may be, all of the material upon which it proceeded. Dixon J, as the Chief Justice then was, said in Victorian Stevedoring: ‘if he has already been convicted, then because his liability has merged in the conviction, it no longer depends upon the law under which it arose, and it does not lapse with the revocation of the law.  The conviction has become the source of his liability for his offence, and the conviction continues in force because its operation does not depend upon the law creating the offence, but upon the authority belonging to a judgment or sentence of a competent Court.’”

55. Reviewing previous authorities in England, Scott Baker LJ in R v Ballinger [2005] Cr App R 433 said, at p 437-438:

“The court had [according to Lord Bingham CJ in R v Hawkins [1997] 1 Cr App R 234] traditionally been reluctant to grant an extension except where the extension was relatively short and a good reason was shown for the failure to apply in time.  He observed that indulgence had traditionally not been shown where the defendant acting on advice had pleaded guilty or where he had taken a conscious decision not to appeal.  He went on to say that the court had to be satisfied that good reason existed for giving leave to appeal out of time pointing out that a change in the law had not usually been regarded as a good reason. (see R v Mitchell (1977) 65 Cr. App. R 185 at 189 and 190).  Following the general practice the question was whether the defendant had suffered any substantial injustice or injury.

Lord Bingham CJ., at p. 240B described the rationale for the practice:

‘The practice may on its face seem harsh.  On the other hand, the consequences of any other rule are equally unattractive.  It would mean that a defendant who had roundly and on advice accepted that he had acted dishonestly and fraudulently, and pleaded guilty, or who had been found guilty and chosen not to appeal could after the event seek to reopen the convictions.  If such convictions were to be readily reopened it would be difficult to know where to draw the line or how far to go back.’”

56. The approach in New Zealand, which Mr Yu commends as striking the correct balance, is that stated in R v Knight [1998] 1 NZLR 583, 588-589:

“On its face that ‘still in the system’ approach would seem to preclude granting an extension of time thereafter even when the applicant was serving a long term of imprisonment.  Finality and certainty are very important values but in this society there must in our view be room for recognising that circumstances particular to the case of an applicant no longer in the conviction appeal system may lead to the conclusion that justice requires that leave be given.  It may be difficult to decide where to draw the line and expressions such as ‘the very rare case’ and ‘exceptional case’ are no doubt intended to counsel against ready resort to the discretion.  In Alofa v Department of Labour [1980] 1 NZLR 139 at p 146, Cooke J observed that the expression ‘the very rare case’ might be unduly restrictive if applied literally, but went on immediately to emphasise that ‘at least special circumstances must be shown to justify a departure from the principle of the finality of the results of court proceedings’. …

“… the starting point must be the principle that a conviction obtained according to law as it was then understood and applied should stand.  Leave to appeal out of time on the ground that there has been a restatement of the applicable law should be granted only where special circumstances can be shown to justify a departure from the principle of finality.  The applicant must demonstrate some special feature or features particular to the case that lead to the conclusion that in all the circumstances justice requires that leave be given.  Amongst the considerations which will also be relevant in that overall assessment are the strength of the proposed appeal and the practical utility of the remedy sought, the length of the delay and the reasons for delay, the extent of the impact on others similarly affected and on the administration of justice, that is floodgates considerations, and the absence of prejudice to the Crown.”

57. The approach in Hong Kong, as signalled by R v Kwok Hing Man [1994] 2 HKCLR 160, appears on its face to be altogether less circumscribed and we are invited by the respondent to say that it was wrongly decided.  That case followed the repeal of s 30 of the Summary Offences Ordinance by the decision in Attorney General of Hong Kong v Lee Kwong Kut [1993] AC 951.  Since the repeal took effect from the commencement of the Bill of Rights Ordinance, those who had since the date of commencement been convicted of a s 30 offence sought leave to appeal out of time against those convictions.  The application was opposed, counsel for the Attorney General relying upon Mitchell and Thomas, above.  The court took the view that there was a material difference between, on the one hand, the removal of a misconception of law that had been relied upon in order to secure a conviction and, on the other, the repeal of an offence creating statutory provision, effective before commission of the act charged.  It was in this context that the court said, at p. 164, that:

“We think that the relevant principle to be derived from [R v Wigman] is best expressed in the last words of the report of R v Ramsden [ [1972] Crim L R 547] , i.e., in the last analysis, it is in every case a matter of discretion for the court.  … We recognise the practical necessity of having some finality in the criminal process.  However it seems to us that the matter here is not so much one of the criminal process itself, but rather the direct result of the legislative scheme adopted which relates back the repeal to the commencement of the Bill of Rights Ordinance.  The practical problems that result seem to be much more amenable to legislative or executive, rather than judicial solution.  So far as the courts are concerned, we think the right approach in the present circumstances is that adopted… in R v Mitchell at p. 757G-H:

       ‘If we were to refuse him the extension of time in which to appeal against conviction, we should be keeping him in prison, so to speak, when we as a court were convinced that he had not committed an offence.  That again is not an attractive proposition, and it is one from which the court resiles.  This seems to us therefore to be the very rare case where the court should exercise its undoubted discretion to allow the extension of time and grant leave to appeal against conviction.’”

Noting that in the case before it ‘the resulting situation is that a person has been convicted of what can be plainly seen to have been unquestionably a nonexistent defence,’ the court concluded that it was ‘difficult to see how his application could be refused in the absence of some special circumstance.’

58. We are here invited to say that Kwok Hing Man was wrongly decided in so far as the court seemed, it is contended, to embrace an open-ended discretion; alternatively, to say that the decision stands in a peculiar position because of the unusual circumstances of the case whereby the applicants had been convicted of that which transpired to be a non-existent offence.  It is unnecessary, indeed unwise, for us to decide the point, or to determine which of the various approaches to which we have referred is to be preferred, assuming that they differ in essence from Kwok Hing Man, and assuming that it could be said that the test in Kwok Hing Man was obviously too wide.  We decline to take the course suggested because it is plain that, with the possible exception of the Canadian approach, none of the approaches excludes special cases, and that the declaration that we are asked to make does exclude them, save for an exception that is likely to arise most rarely.

59. Whilst the thread of authority is to exclude appeals out of time where the appeal is based upon a changed understanding of the law – and we need not for present purposes concern ourselves with non-existent offences – that thread constitutes the usual approach, the starting point, the general principle, but with allowance nonetheless for special circumstances.  Assuming for the moment that we were to lay down a general principle, a usual rule, that applicants out of time will never be granted leave to appeal on the mere basis of some previous assumption that the reverse onuses imposed by section 47 of the Dangerous Drugs Ordinance were persuasive onuses but assuming, again for the moment, that we were upon a full consideration of the matter to exclude the Canadian approach – and we are not invited by the respondent to adopt it – it would follow that the door would be open to the exceptional or special case.  If that were so, one could not reasonably predict, let alone delineate in some declaration, what those exceptional or special cases might be.  Yet the effect of the declaration that we are asked to make is not even to predict or delineate such exceptional or special cases.  Rather, it is to exclude them altogether.  That we are not now prepared to do.  

60. What the respondent has in fact been seeking in this application is a definition of the circumstances in which this court will permit applications to be made for leave to appeal out of time and so narrowly to define those circumstances as to turn the definition into a wholesale temporal limitation upon the effect of our decision of June 2005.  We are satisfied that that is an unacceptable approach, and we decline to adopt it.  If at a future date an application is, or applications are, made for leave to appeal out of time which are based in whole or in part upon a suggested changed interpretation or perception of the nature of the reverse onuses in section 47 of the Dangerous Drugs Ordinance, we will deal with them then for what they are – applications for leave to appeal out of time.  It may then be necessary in the light of the case before us – or, perhaps, a representative selection of several cases – to examine the applicable principles.  But to do so now would be to do so in the absence of any such application, and that we will not do.

Conclusion

61. In summary:

(1)   Article 160 of the Basic Law does not mandate temporal limitation upon the effect of a judicial interpretation of a statutory provision or upon a judicial declaration of incompatibility of a statutory provision with the Basic Law.

(2)   It is not necessary for the purpose of the issues decided by this court’s judgment of 23 June 2005 to determine whether, otherwise, temporal limitation upon the effect of judicial decisions – whether interpretations of statutes or declarations of incompatibility – is part of the remedial armoury available to the courts of this jurisdiction, for we are satisfied that even were that armoury available, there can be no justification for extending the limited boundaries of its use as exemplified in those jurisdictions that have utilized or expressed a willingness to utilize the technique and, for the reasons we have provided, we are further satisfied that the decisions with which we are here concerned fall well outside those boundaries.

62. Accordingly, we reject the respondent’s application for a declaration that our interpretation of the effect of the presumptions in section 47(1) and (2) of the Dangerous Drugs Ordinance is, subject to specified exceptions, to apply only from 24 June 2005.

63. There will be a costs order nisi that the costs of the amicus curiae and of the appellants be paid by the respondent, to be taxed if not agreed, the appellants’ costs to be taxed in accordance with the Legal Aid Regulations.