Ретроактивність змін в судовій практиці ЄСПЛ

Abstract

Jenny Sandvig LL.M. Program Retroactivity of Changes in the Case-law of the European Court of Human Rights April, 2018 Supervisor: Professor Vicky Jackson LL.M. 50-Page Paper

RETROACTIVITY OF CHANGES IN THE CASE- LAW OF THE EUROPEAN

COURT OF HUMAN RIGHTS

Jenny Sandvig1

The paper addresses a gap in the literature on what to do when the European Court of Human Rights (ECtHR), which supervises the implementation of the European Convention of Human Rights (ECHR), changes its understanding of the Convention. More precisely, it asks whether the ECtHR should apply changes in case-law retroactively to find violations in cases that were decided by domestic courts in accordance with ECtHR case -law as it stood bef ore the change . The Court has predominantly answered this question in the affirmative. Yet, its approach is far from consistent. Some of its judges have warned that the ECtHR , by defaulting domestic courts for applying the Convention correctly, undermines domestic implementation of the Convention and upsets legal certainty. The

paper makes two contributions. First, it systematizes inconsistencies in ECtHR case- law on retroactivity to tease out circumstances in which the Court has found reason to limit retroactivity. Second, it draws on the U.S. Supreme Court’s (SCOTUS) limited retroactivity doctrines and U.S. scholarly debate about retroactivity to argue that the ECtHR should limit the temporal effect of changes in case-law when reviewing domestic court decisions that were correctly decided under ECtHR case-law at the time.

1 LL.M. Candidate 2018, Harvard Law School. This LL.M. paper is written under the excellent supervision of Professor Vicky Jackson, Harvard Law School. I thank Professor Alex Whiting, Harvard Law School, and Assistant Clinical Professor of Law and Director of the International Human Rights Clinic, Chicago Law, Claudia Flores, for valuable and inspiring comments during the Salzburg Cutler Fellows Program in Washington, D.C., on February 23, 2018. I also thank S.J.D. Candidate, Kári Holmar Ragnarsson, Harvard Law School, and Special Assistant to the Graduate Program, Jane Fair Bestor, for their comments on earlier drafts. Faults are mine.

Table of Content 1 Introduction .............................................................................................................................. 1 2 The Problem of Adjudicative Retroactivity ............................................................................. 3 2.1 Overview .......................................................................................................................... 3 2.1.1 Description of the Problem ...................................................................................... 4 2.1.2 Some Terminology ................................................................................................... 4 2.1.3 Delineations ............................................................................................................. 4

2.2 Method ............................................................................................................................. 5 2.3 The Scholarly Debate on Retroactivity ............................................................................ 5 2.3.1 The Debate on ECtHR Retroactivity ........................................................................ 5 2.3.2 The Debate on Retroactivity in the United States .................................................... 8 2.3.3 My Contribution ..................................................................................................... 12 3 The ECtHR’s Approach to Retroactivity ............................................................................... 13

3.1 Overview ........................................................................................................................ 13 3.2 Cases of Absolute Retroactivity ..................................................................................... 13 3.2.1 Introduction ............................................................................................................ 13 3.2.2 Valenzuela Contreras v. Spain ............................................................................... 13 3.2.3 Aoulmi v. France .................................................................................................... 14 3.2.4 Nicola Silvestri v. Italy .......................................................................................... 15

3.3 Cases of Limited Retroactivity ...................................................................................... 16 3.3.1 Introduction ............................................................................................................ 17 3.3.2 Cases Balancing Third Party Interests ................................................................... 17 3.3.3 Denial of Compensation in Transsexual Legal Recognition Cases ....................... 20 3.4 Cases Where Judges Disagreed over Retroactivity ........................................................ 23 3.4.1 Introduction ............................................................................................................ 23

3.4.2 Borgers v. Belgium ................................................................................................ 23 3.4.3 Demir and Baykara v. Turkey ................................................................................ 24 3.4.4 Lucky Dev. v. Sweden ........................................................................................... 25 3.5 Preliminary Conclusions ................................................................................................ 27 4 SCOTUS’ Approach to Retroactivity .................................................................................... 28 4.1 Overview ........................................................................................................................ 28

4.2 Criminal Cases ............................................................................................................... 29 4.3 Civil Cases ..................................................................................................................... 32

4.4 Qualified Immunities Doctrine ...................................................................................... 34 4.5 Summary ........................................................................................................................ 35 5 A Non-Retroactivity Approach for the ECtHR? .................................................................... 35 5.1 Preview of the Argument ............................................................................................... 35 5.2 Would Non-Retroactivity Preclude Dynamic Interpretation? ........................................ 37 5.3 The Distinction Between Pending Cases ....................................................................... 38

5.4 Assessing the Underlying Principles .............................................................................. 39 5.4.1 Considerations Advanced for Retroactivity ........................................................... 39 5.4.2 Considerations Advanced in Favor of Non-Retroactivity ...................................... 43 5.5 Bright Line-Rule or Discretionary Standard? ................................................................ 48 5.6 Conclusion ..................................................................................................................... 49 6 Final Remarks ........................................................................................................................ 49

1/50 1 Introduction This LL.M. paper addresses whether the European Court of Human Rights (ECtHR) should apply changes in its case- law with absolute retroactive effect. This question frequently arises where the ECtHR departs from case- law in one case, and thereafter must decide whether the newly established interpretation should apply with retroactive effect to other complaints in which the facts predate the new rule, and where the domestic judicial decisions at issue have been decided in accordance with ECtHR case-law before it shifted.

Absolute retroactivity, therefore, entails finding Contracting States in breach of human rights because national courts have implemented ECtHR case- law as it sto od when the matter was decided domestically. Over the years, the ECtHR has approached the question of retroactivity in different ways. While often opting for absolute retroactivity over state protests, the Court has occasionally limited the retroactive eff ect of new case- law, or displayed an ongoing debate about retroactivity through dissents.

The question of absolute retroactivity is of practical and normative significance. It is of practical significance because the ECtHR often reverse its jurisprudence in order to interpret the Convention dynamically in light of “present-day conditions”,2 or to harmonize split approaches to a question taken by its different Chambers. 3 While decisions by the ECtHR are not formally binding on others than the parties to a given case, 4 the interpretations of the Convention are considered in practice to have erga omnes effects.5 Thus, from what point in time is the Contracting State, or other Contracting States, required to implement a new interpretation? From what point in time may other individuals claim

remedies for similar situations? Retroactivity is also of normative importance, since it asks in essence whether law is found or judge -created, and whether dynamic interpretat ion in light of the present ought also to change the past.

2 Tyrer v. The United Kingdom, App. No. 5856/72, Eur. Ct. H.R. (ser. A) 26 (1978). 3 See for instance Sergey Zolotukhin v. Russia [GC], App. No. 14939/03, Eur. Ct. H.R. 2009-I; A and B v. Norway [GC], App. Nos. 24130/11, 29758/11, Eur. Ct. H.R. 2016. The author of this paper was counsel for the Norwegian Government in the latter case.

4 Article 46 § 1 of the ECHR reads: “The High Contracting Parties undertake to abide by the final judgment of the Court in any case to which they are parties.” 5 Oddný Mjöll Arnardóttir, Res Interpretata, Erga Omnes Effect and the Role of the Margin of Appreciation in Giving Domestic Effect to the Judgments of the European Court of Human Rights, 28 EUR. J. INT. LAW 819–843 (2017).

2/50 This tension between evolving case-law and absolute retroactive application has been widely discussed in common law jurisdictions, and in particular in the United States (U.S.). This can be explained by the fact that a rule pronounced in a common law case will have general binding effects erga omnes. In civil law jurisdictions, this tension has traditionally been less pronounced. Ordinary civil law courts typically perform ex post reviews of concrete facts with effects limited to the parties ( inter partes )6 without formally establishing binding rules for others.7 Hence, retroactive application of new case-law from ordinary civil law courts has not raised the same issues in Continental Europe as those that

have fueled the debate in common law jurisdictions.8 Courts and scholars in the U .S., in particular, have attempted and evaluated various solutions to the retroactivity problem , ranging from non-retroactivity on a case -by case basis to absolute retroactivity.9 Today, changes in constitutional inte rpretation by the Supreme Court of the United States (SCOTUS) apply with retroactive effect to all pending cases, criminal and civil, but normally not to pending criminal cases on collateral review (habeas corpus) .10 In cases on collateral review, the Supreme Court reviews the constitutionality of the state court’s deci sion within the legal context when it was pronounced, reserving the possibility of case-by-case retroactive application to watershed

changes in consti tutional law, or to decriminalized conduct.11 Under the qualified immunities doctrine, moreover, SCOTUS shields officials from liability insofar their actions did not violate “clearly established law” at the time.12 As the ECtHR ha s adopted a dynamic or evolving approach to treaty interpretation, with de facto binding effects of its interpretations erga omnes , I argue that the considerations that underlie the doctrines on retroactivity in U.S. constitutional law, are of interest to the ECtHR. Notwithstanding the significant differences between SCOTUS and the ECtHR, the paper argues that the ECtHR should consider the doctrines developed by SCOTUS in

6 P. POPELIER ET AL., THE EFFECTS OF JUDICIAL DECISIONS IN TIME (2014); EVA STEINER, COMPARING THE

PROSPECTIVE EFFECT OF JUDICIAL RULINGS ACROSS JURISDICTIONS (2015).

7 STEINER, supra note 6 at 4. 8 Id. at 4. 9 To be discussed in Chapter 4. 10 Griffith v. Kentucky, 479 U.S. 314 (1987); Teague v. Lane, 489 U.S. 288 (1989). 11 Teague, 489 U.S. 12 Harlow v. Fitzgerald, 457 U.S. 800 (1982).

3/50 search of a coherent and better approach to retroactivity. As the ECtHR, unlike SCOTUS, is not bound by a principle of stare decisis, it is all the more important to balance the Court’s changing interpretations with considerations for legitimate expectations and legal certainty of final decisions taken by national courts. Moreover, the ECtHR’s approach to retroactivity should reflect the shared responsibility with domestic courts for safeguarding the rights and freedoms in the Convention, and the subsidiary character of its own supervision.13 The importance of deference to domestic implementation was emphasized most recently in the final Copenhagen Declaration.14 I argue that SCOTUS’s approach to

final state court decision s on collateral review could inform this exercise in judicial restraint and inter- court comity. Otherwise, the ECtHR risks undermining domestic implementation and weakening the legal certainty of its own decisions.15 To oper ationalize these concerns, the paper proposes that the ECtHR should adopt a doctrine on non-retroactivity. While new case-law should apply retroactively to cases that were still pending before domestic courts when the ECtHR changed its case-law, it should not apply retroactively to cases that were already finally decided by domestic courts in accordance with ECtHR case- law at the time, save in instances where the situation complained of is ongoing. Moreover, departures from domestically implemented case-law

should be made ex nunc, without retroactive remedies. The paper is divided into six chapters. The second chapter provides an overview of the retroactivity question, before placing the contribution of the paper within the scholarly debate. A third chapter discusses ECtHR’s conflicting approach to retroactivity. A fourth chapter outlines SCOTUS’ approach to retroactivity. A fifth chapter develops the proposal on non-retroactivity and discusses likely objections. A sixth chapter concludes.

2 The Problem of Adjudicative Retroactivity 2.1 Overview 13 Copenhagen Declaration, (2018). 14 Id. 15 See similar concerns voiced in the separate opinion of Lucky Dev v. Sweden, App. No. 7356/10, Eur. Ct. H.R., November 27, 2014.

4/50 2.1.1 Description of the Problem Any court de cision is in a sense retroactive, in that it adjudicates preceding facts. This inherent temporal aspect of adjudication is not at issue here. Insofar as a court decision constitutes a foreseeable interpretation of the law when the facts occurred, the decision will not upset the actors’ legitimate expectations at the material time. The problem of adjudicative retroactivity arises first where a court departs from previous understandings of the law, and decides predating facts according to a new rule which was not foreseeable at the material time. Such retroactivity taxes legal certainty and legitimate expectations.

These strains on legal certainty and legitimate expectations are further aggravated where a superior court (international or federal) applies new case-law to review a superior domestic court decision which has acquired a res judicata effect, and which was decided before the case-law changed, in accordance with previous case-law. This is the problem I address.

2.1.2 Some Terminology I distinguish between absolute retroactivity and limited retroactivity. Absolute retroactivity means applying newly announced case-law to any older case, even cases that were finally decided at the time case- law shifted.16 This has been described by one critic as inserting error in proceedings that were errorless at the time they were concluded. 17 Limited retroactivity means applying a change in case-law to the case-at-point and similarly situated parties in cases pending direct review, but not to cases that were finally decided at the time.

2.1.3 Delineations Article 7 of the ECHR prohibits retroactive criminal law, understood to include statutory and adjudicative criminalization. This prohibition has been widely discussed elsewhere, and raises particular problems which lie outside the confines of this paper.18 16 F. Tulkens & S. Van Drooghenbroeck, The Shadow of Marckx for a Renewed Debate on the Temporal Effects of Judgments of the European Court of Human Rights, in THE GLOBAL COMMUNITY. YEARBOOK OF INTERNATIONAL LAW AND JURISPRUDENCE I, 251 (Giuliana Ziccardi Capaldo ed., 2011).

17 Kermit Roosevelt III, A Little Theory is a Dangerous Thing: The Myth of Adjudicative Retroactivity, 31

CONN. LAW REV. 1075–1138 (1998).

18 See for instance Simon Atrill, Nulla poena sine lege in comparative perspective: retrospectivity under the ECHR and US constitution, PUBLIC LAW 107–131 (2005); Susana Sanz - Caballero, The Principle of Nulla Poena Sine Lege Revisited: The Retrospective Application of Criminal Law in the Eyes of the European Court of Human Rights, 28 EUR. J. INT. LAW 787–817 (2017).

5/50 2.2 Method My paper will address the question of retroactivity by analyzing the jurisprudence of the ECtHR and attempt ing a reor ganization. To this end, I will employ in part a doctrinal method and in part a recasting method. To shed light on the inter -court dimension of retroactivity and illustrate how the question can be approached differently in different kind of cases, I will a lso compare the ECtHR’s jurisprudence with the SCOTUS’ doctrines on non-retroactivity. This comparative analysis will flesh out the considerations in play in analogous situations in the U.S., and explore the transfer value of the U.S. debate about limited retroactivity.

2.3 The Scholarly Debate on Retroactivity 2.3.1 The Debate on ECtHR Retroactivity It is worthwhile addressing whether changes in ECtHR’s case -law should be nonretroactive, for several reasons. First, the issue pertains to the ongoing discussion on subsidiarity and the need for consistency in the Court’s case- law, in order to enable efficient domestic implementation of the Convention. A much debated Draft D eclaration circulated by Denmark in the lead -up to the High Level Conference on Reform of the Human Rights Convention System in Copenhagen, April, 2018, argued that the Court “should take positions which are stable and coherent”, so that “national authorities can apply and enforce the Convention principles at [the] domestic level and rule with certainty

[…] without running the risk of subsequent disavowal.” 19 Critics, including the Court, countered that Contracting States should not curtail the Court’s ability to develop the Convention.20 I will argue that a doctrine on non -retroactivity could strike a functional 19 Draft Copenhagen Declaration, (2018) § 57.

20 European Court of Human Rights, OPINION ON THE DRAFT COPENHAGEN DECLARATION (2018); Alice Donald & Philip Leach, A WOLF IN SHEEP’S CLOTHING: WHY THE DRAFT COPENHAGEN DECLARATION MUST BE REWRITTEN EJIL: TALK!, https://www.ejiltalk.org/a-wolf-in-sheeps-clothing-why-the-draftcopenhagen-declaration-must-be-rewritten/ (last visited Apr 12, 2018); Andreas Follesdal & Geir Ulfstein,

THE DRAFT COPENHAGEN DECLARATION: WHOSE RESPONSIBILITY AND DIALOGUE? EJIL: TALK!,

https://www.ejiltalk.org/the-draft-copenhagen-declaration-whose-responsibility-and-dialogue/ (last visited Apr 12, 2018); Leonie Huijbers, THE DRAFT COPENHAGEN DECLARATION – PROCESS-BASED REVIEW AND SUBSIDIARITY STRASBOURG OBSERVERS (2018), https://strasbourgobservers.com/2018/02/27/the-draftcopenhagen-declaration-process-based-review-and-subsidiarity/ (last visited Apr 12, 2018); Antoine Buyse,

THE DRAFT COPENHAGEN DECLARATION – WHAT ABOUT CIVIL SOCIETY? STRASBOURG OBSERVERS

(2018), https://strasbourgobservers.com/2018/03/01/the-draft-copenhagen-declaration-what-about-civil-

6/50 middle ground to resolve some of this controversy. It would avoid disruptive retroactive changes to settled implementations, and further the interaction between the ECtHR and domestic courts implementing the Convention, while still preserving t he Court’s prerogative to develop the Convention.

Second, there is a need for ECtHR’s case-law on retroactivity to be systematized, because it is inconsistent and largely overlooked. While the Court predominantly applies new changes with absolute retroactiv ity, it limits temporal effects in some type of cases, and splits on the issue in others. This inconsistency is evidence of concern on the part of some ECtHR judges.21 In Lucky Dev v. Sweden, 2014, three judges warned that it is “disruptive for national courts following the Court’s case -law faithfully to find themselves – without any warning – accused of a breach of the Convention”.22 They stressed the need for “careful consideration” of the retroactivity problem in order not to upset the interaction with

domestic courts, and called for “heightened attention” to this issue.23 So far, however, this call for action has escaped scholarly attention. I have not been able to find recent articles in French, English, Norwegian, Danish or Swedish dealing in particular with ECtHR retroactivity after Lucky Dev . The few scholars that discussed retroactivity before Lucky Dev are divided. Some have analyzed retroactivity on declaratory grounds.24 Popelier argues that since the ECtHR refers to its judgments as “essentially declaratory”, a new interpretation of the Convention must be considered part of the Convention from its entry into force.25 Christoffersen seems to share this view. He argues that because the Court

views its judgments as “declaratory in nature”, there can be “no temporal restrictions on society/ (last visited Apr 12, 2018); Roisin Pillay, THE EUROPEAN CONVENTION ON HUMAN RIGHTS: THE

DRAFT COPENHAGEN DECLARATION AND THE THREAT TO THE EUROPEAN COURT OPINIO JURIS,

http://opiniojuris.org/2018/03/02/33469/ (last visited Apr 12, 2018); Janneke Gerards & Sarah Lambrecht,

THE DRAFT COPENHAGEN DECLARATION - COMMENT SERIES I ECHR BLOG,

http://echrblog.blogspot.com/2018/02/the-draft-copenhagen-declaration_26.html (last visited Apr 12, 2018); Joint NGO Response to the Draft Copenhagen Declaration on the European Convention on Human Rights system, (2018), http://www.omct.org/statements/2018/02/d24721/ (last visited Apr 12, 2018).

21 Chapter 3 will discuss concerns raised by ECtHR judges further. 22 Lucky Dev v. Sweden, App. No. 7356/10, Eur. Ct. H.R., November 27, 2014. 23 Id. 24 Patricia Popelier, The European Court of Human Rights, in THE EFFECTS OF JUDICIAL DECISIONS IN TIME 265–284 (Patricia Popelier et al. eds., 2014); J. CHRISTOFFERSEN, FAIR BALANCE: PROPORTIONALITY,

SUBSIDIARITY AND PRIMARITY IN THE EUROPEAN CONVENTION ON HUMAN RIGHTS (2009).

25 Popelier, supra note 24 at 280; CHRISTOFFERSEN, supra note 24 at 435.

7/50 the effect of the judgments.” 26 Christoffersen finds support for this understanding in what he refers to as an “essential positivistic notion that the legal obligations expressed in the Court’s judgments do not flow from the judgment as such, but from the binding force of the ECHR; the judgments are intermediate declarations of the law as it would stand irrespective of the judgments.”27 This argument is misleading, however, for at least two reasons. 28 First, it is incompatible with the kind of dynamic interpretation which give rise to the retroactivity problem in the first place. If judges merely declare the law which has always been , Convention rights would not be subject to change. Yet, by virtue of the “living instrument” -doctrine,29

Convention rights unmistakably change, in particular when the Court overrules itself. Thus, the new law cannot be the law that was without rewriting the past. A more candid recognition that judges make law has prompted a nuanced discussion about retr oactivity regarding decisions by SCOTUS (to be discussed below). Second, it is debatable whether the ECtHR’s characterizations of its judgment s as “essentially declaratory” necessarily mean that changes in case- law have absolute retroactive effect. Rather, it could simply mean that ECtHR judgments are not prescriptive, and leave t o the Contracting State the choice of measures to implement the decision.30 Other commentators take a different view. Tulkens and S Van Drooghenbroeck call out the

Orwellian character of the absolute retroactivity of abrupt changes in ECtHR case-law, and point to inconsistencies in the treatment of cases. 31 While they indicate sympathy for limited retroactivity , or even pure prospectivity, they do not propose any conclusive solution.32 Besson goes further. She argues that because “human rights are indeterminate” 26 CHRISTOFFERSEN, supra note 24 at 435.

27 Id. at 435. 28 Richard H. Jr. Fallon & Daniel J. Meltzer, New Law, Non-Retroactivity, and Constitutional Remedies,

104 HARV. LAW REV. 1731–1833, 1758 (1990).

29 Tyrer v. The United Kingdom, App. No. 5856/72, Eur. Ct. H.R. (ser. A) 26 (1978). 30 Marckx v. Belgium (Plenary), App. No., 6833/74, Eur. Ct. H.R. (ser. A) 31 (1979), § 58: “the Court’s judgment is essentially declaratory and leaves to the State the choice of the means to be utilized in its domestic legal system to comply with the judgment"; Moreira Ferreira v. Portugal (No. 2) [GC], App. No.

19867/12, Eur. Ct. H.R. 2017, § 47: “[F]indings of a violation in its judgments are essentially declaratory and, by Article 46 of the Convention, the High Contracting Parties undertook to abide by the final judgments of the Court in any case to which they were parties”.

31 Tulkens and Van Drooghenbroeck, supra note 16. 32 Id.

8/50 and must be subject to “dynamic” interpretation, “qualification or overruling of the precedent is always possible” and “jurisprudential authority only temporary”. 33 On this basis, she posits that new interpretations of the Court can “never have retroactive effect”.34 One problem with this view, however, is that it is contradicted by the body of cases from the ECtHR that give new changes absolute retroactive effect. Besson does not mention these cases. Moreover, her claim appears to overstate the particularity of human rights, while overlooking that considerations underlying retroactivity – justice, equality, legal certainty, reliance – might call for different solutions in different types of cases.

2.3.2 The Debate on Retroactivity in the United States If scholarly debate about ECtHR retroactivity is underdeveloped, the opposite is true of the debate in the U.S.35 Traditionally, the debate about retroactivity in the U.S. also turned on whether judges declare or make the law. 36 The declaratory position is often attributed to Blackstone, who wrote that judges are “not delegated to pronounce a new law, but to maintain and expound the old one.” 37 On this view, departures from overruled cases wer e 33 Samantha Besson, The Erga Omnes Effect of Judgements of the European Court of Human Rights,

COUR EUR. DROITS HOMME APRÈS PROTOC. 14 - PREM. BILAN PERSPECT. 125–175, 160 (2011).

34 Id. at 160. 35 Roosevelt, supra note 17; PETER W. LOW, FEDERAL COURTS AND THE LAW OF FEDERAL-STATE

RELATIONS (7th ed. ed. 2011); LOUISE WEINBERG, FEDERAL COURTS: CASES AND COMMENTS ON JUDICIAL

FEDERALISM AND JUDICIAL POWER (1994); Kit Kinports, Habeas Corpus, Qualified Immunity, and Crystal Balls: Predicting the Course of Constitutional Law, 33 ARIZ. LAW REV. 115–198 (1991); Lee-ford Tritt, Moving Forward by Looking Back: The Retoractive Application of Obergefell, 2016 WIS. LAW REV. 873– 946 (2016); Fallon and Meltzer, supra note 28; Ann Woolhandler, Public Rights, Private Rights, and Statutory Retroactivity, 94 GEORGETOWN LAW J. 1015–1064 (2005); Beryl Harold Levy, Realist Jurisprudence and Prospective Overruling, 109 UNIV. PA. LAW REV. 1–30 (1960); Richard S. Kay, Retroactivity and Prospectivity of Judgments in American Law, in COMPARING THE PROSPECTIVE EFFECT OF JUDICIAL RULINGS ACROSS JURISDICTIONS 209–246 (2015); Gil J. Ghatan, The Incentive Problem with

Prospective Overruling: A Critique of the Practice, 45 REAL PROP. TRUST ESTATE LAW J. 179–212 (2010); Pamela J. Stephens, The New Retroactivity Doctrine: Equality, Reliance and Stare Decisis, 48 SYRACUSE LAW REV. 1515–1576 (1998); Steven W. Allen, Toward a Unified Theory of Retroactivity, 54 N. Y. LAW SCH. LAW REV. 105–146 (2009); James B. Haddad, Finality Distinction in Supreme Court Retroactivity Analysis: An Inadequate Surrogate for Modification of the Scope of Federal Habeas Corpus, 79 NORTHWEST. UNIV. LAW REV. 1062–1079 (1984); Paul J. Mishkin, Foreword: The High Court, the Great Writ, and the Due Process of Time and Law Supreme Court 1964 Term, The, 79 HARV. LAW REV. 56–102 (1965); Orin S. Kerr, Good Faith, New Law, and the Scope of the Exclusionary Rule, 99 GEORGETOWN

LAW J. 1077–1118 (2010); Jill E. Fisch, Retroactivity and Legal Change: An Equilibrium Approach, 110

HARV. LAW REV. 1055–1123 (1996).

36 BEN JURATOWITCH, RETROACTIVITY AND THE COMMON LAW 41 (2008) with further references.

37 BLACKSTONE, COMMENTARIES 69 (15 ed. 1809).

9/50 only corrections of previous courts’ failure to discern and declare the “true law”. 38 Since the underlying “true law” was the same, even if judgments changed, the declaratory theory did not recognize a problem with retroactivity.39 At the turn of the ninet eenth century, however, critics would call this out as a fiction.40 Yet, while most U.S. scholars today reject the declaratory theory as “antiquated”41 and “irrelevant”,42 it has its occasional proponents.

Three related positions are of interest here. First, in what has been referred to as a “Neo -Blackstonian” position, Dworkin argues that “legal decisions enforce pre-existing rights, and right answers exists in almost all cases”.43 The best rule, on this view, should apply to any action regardless of time. The problem with this position, however, is that it neglects the fact that what is the “right answer” or the “best rule” in fact changes with time.44 Therefore, it does not solve the question of whether to apply old or new law to old facts.

A second position has been argued by Justice Scalia. In American Trucking, for instance, he objected against the very notion of retroactivity, because it “presupposes a view of our decisions as creating the law, as opposed to declaring what the law already is.”45 It did not make sense, he asserted, to limit the retroactive effect of the Court’s interpretations “since the [Constitution] does not conform to our decisions, but our decisions are supposed to conform to it”. 46 This position on retroactivity , however, is premised on a distinct originalist method of U.S. constitutional interpretation which differs from the ECtHR’s dynamic approach.

Third, it has been argued that the declaratory theory of adjudication still informs the retroactivity doctrines in the U.S. 47 Allen posits that “any doubt that the current Supreme 38 JURATOWITCH, supra note 36. 39 Id. 40 Levy, supra note 35.

41 Fisch, supra note 35 at 1082. 42 Fallon and Meltzer, supra note 28 at 1759.

43 RONALD DWORKIN, A MATTER OF PRINCIPLE 119 (1985).

44 Roosevelt, supra note 17. 45 American Trucking Ass’ns v. Smith, 496 U.S. 167 (1990) (Scalia, J., concurring). In United States v. Virginia, 496 U.S. 167 (1990) (Scalia, J., dissenting), he opposed finding male-only military education in violation of the equal protection clause under the 14th Amendment, because “the assertion that [the] tradition has been unconstitutional through the centuries is not law, but politics-smuggled-into-law”.

46 Id. 47 Allen, supra note 35 at 108.

10/50 Court accepts this jurisprudential principle [was] recently put to rest in Danforth v. Minnesota”.48 Justice Stevens, writing for the majority, stated that the “source of a ‘new rule’ is the Constitution itself, not any judicial power to create new rules of law. 49 Accordingly, the underlying right necessarily pre-exists our articulation of the new rule.”50 This seems, at face value, to be reminiscent of the declaratory accounts of ECtHR retroactivity discussed above.

The opposing position to the declaratory theory openly acknowledges that judges do make law, and hence, that the l aw changes when judges overrule previous cases. 51 This acknowledgment prompts the question of whether it is appropriate to give new judge-made law retroactive effect. It has produced proposals along a continuum from full retroactivity to non- retroactivity. Pamela Stephens, for instance, argues for a presumption for retroactivity out of concerns for equality. 52 This presu mption could be rebutted u pon showing of actual reliance, whereby reliance interests are weighed against the benefits of retroactivity in the specific context. 53 David Lehn, on the other hand, argues for a presumption for non-retroactivity. He views retroactivity as a preclusion pro blem, which

should be solved under a “reliance-based cost-benefit test”, with a rebuttable presumption for non-retroactivity, in both pending and final cases. 54 Kermit Roosevelt III proposes a “decision-time model” in which courts decide the case according to the best legal answer at the time of the decision. 55 The decision-time model attempts to solve the problem of applying new law in habeas corpus cases, by confining “decision- time” to the time of the final direct review . This notion is clever , but it does not solve the normative question of whether individuals should benefit from new interpretations on collateral review.

Retroactivity has also been analyzed by way of equilibrium theory or remedial law . Jill Fisch argues that retroactivity is justified in the context of an unstable regulatory 48 Id. at 108. 49 Danforth v. Minnesota, 552 U.S. 264 (2008).

50 Id. 51 Levy, supra note 35. 52 Stephens, supra note 35 at 1574. 53 Id. at 1574. 54 David Lehn, Adjudicative Retroactivity as a Preclusion Problem: Dow Chemical Co. v. Stephenson, 59

N. Y. UNIV. ANNU. SURV. AM. LAW 563–608, 565 (2003).

55 Roosevelt, supra note 17 at 1119.

11/50 equilibrium, and unjustified in the context of a stable regulatory equilibrium.56 Others have sought to distill the remedial aspect of retroactivity. In his minority opinion in American Trucking, Justice Stevens proposed a distinction between the application of a new rule (retroactivity) and the legal effects of that application (remedy), and argued that courts could deny remedies otherwise warranted by a retroactive rule if such relief would upset parties’ expectations. 57 He gathered a seven vote majority for this view in Danforth v.

Minnesota.58 Here, Justice Stevens argued that the issue of retroactivity relates “not to the temporal scope of a newly announced right, but [to] whether a violation of the right that occurred prior to the announcement of the new rule will entitle a criminal defendant to the relief sought”.59 In Footnote 5, he contended that it would make more sense to speak of “redressability” instead of “retroactivity” of newly announced rights.60 Fallon and Meltzer develop this distinction between interpretation and implementation further.61 They argue that the law of remedies provides the “appropriate doctrinal framework” for considering the problem of new law. 62 They identify two fundamental principles – interest in individual redress and corrective justice, and instrumental

deterrence.63 Balancing these two interests, they argue that a presumption for remedies must occasionally yield in instances of new law out of considerations for “reliance, disruption, or s ome similar reason”. 64 The threshold test for new law is to be made with reference to the decision’s “relative predictability”,65 counting “extreme unpredictability”66 and discounting “ordinary legal evolution”.67 Their proposed remedial calculus allows for withholding remedies in situations where awarding them retroactively would be 56 Fisch, supra note 35 at 1123.

57 American Trucking Ass’ns v. Smith, 496 U.S. 167 (1990). 58 Danforth v. Minnesota, 552 U.S. 264 (2008). 59 Id. 60 Id. 61 Fallon and Meltzer, supra note 28 at 1832; See also Roosevelt, supra note 17 at 1119 arguing that concerns about notice and reliance are relevant to the “remedial calculus”, not the question of what law applies.

62 Fallon and Meltzer, supra note 28 at 1883. 63 Fallon and Meltzer, supra note 28. 64 Id. at 1797. 65 Id. at 1794. 66 Id. at 1796. 67 Id. at 1796.

12/50 “potentially disruptive”. 68 The relevant values, they argue, have been “c aptured” by the Supreme Court Stovall v. Denno and Chevron Oil Co. v. Hudson. 69 Here, the Supreme Court formulated three-part standards for when newly announced rights should be given retroactive effect. These cases, and the doctrines that replaced them, will be discussed further in Chapter 4.

2.3.3 My Contribution This review has shown that whereas scholarly discussi on of ECtHR retroactivity is underdeveloped, there is a vigorous U.S. debate that is bursting with various approaches to what is, at least in some respects, the same issue. Given the frequency of the ECtHR’s departures from precedents, the occurrence of sp lit approaches in different ECtHR chambers,70 and the ongoing debate about reforming the Convention system, 71 there is an urgent need to assess and evaluate how the ECtHR can apply new changes to review domestic implementation in a coherent and foreseeable manner. This paper addresses this problem by systematizing the EC tHR’s approach, and exploring the transfer value of the

US debate and doctrines about limited retroactivity. It contributes to scholarly debate by proposing that the ECtHR ought to adopt a doc trine on non- retroactivity in cases where domestic courts have decided the case in accordance with the existing ECtHR case-law at the time. This proposal would in some ways resemble the non- retroactivity doctrine of federal habeas corpus cases in the U.S., with its emphasis on incentivizing state courts to apply the Constitution as it was interpreted at the time a case was finally decided. It has been argued elsewhere that the time is ripe for European common and civil law jurisdictions to contemplate intro ducing the U.S. practice of non-retroactive overruling.72

68 Id. at 1793. 69 Id. at 1797.

70 KATIA LUCAS-ALBERNI, LE REVIREMENT DE JURISPRUDENCE DE LA COUR EUROPÉENNE DES DROITS DE

L’HOMME (2008).

71 High-Level Conference on reform of Human Rights Convention system, HUMAN RIGHTS AND RULE OF LAW, https://www.coe.int/en/web/human-rights-rule-of-law/news/- /asset_publisher/qI77sHb3q28L/content/high-level-conference-on-reform-of-human-rights-conventionsystem (last visited Apr 11, 2018).

72 Eva Steiner, Judicial Rulings with Prospective Effect-from Comparison to Systematisation, in COMPARING THE PROSPECTIVE EFFECT OF JUDICIAL RULINGS ACROSS JURISDICTIONS 1–23 (Eva Steiner ed., 2015).

13/50 Yet, to the best of my knowledge, no one has so far conducted an integrated analysis of the approach to retroactivity taken by the ECtHR and SCOTUS.73 3 The ECtHR’s Approach to Retroactivity 3.1 Overview As I spell out above, ECtHR’s case-law on retroactivity is wanting in consistency . In general, changes in case -law are given abso lute retroactive effect, but occasionally the Court limits the temporal effect of new interpretations. Only rarely has the Court engaged in reflections over the temporal effects that changes in its case-law ought to have. 74 This chapter is structured as follows. Section 3.2 gives some examples of the predominant rule of absolute retroactive effect. Section 3.3 discuss es two types of cases where the Court

gives its judgments limited retroactive effect . Section 3.4 discuss es cases where judges have expressed different views on the temporal effects of new case-law. Section 3.5 draws some preliminary conclusions. 3.2 Cases of Absolute Retroactivity 3.2.1 Introduction In cases where the Court applies new case- law to review olde r facts and decisions, one sometimes find submissions by governments that t he Court should decide the case with respect to the applicable legal context at the time of the final national court decisions. The Court is typically dismissive of such arguments. To flesh out the dialectic, which illuminates the tension between absolute retroactivity and evolutive interpretation, I will

describe three cases in more detail. 3.2.2 Valenzuela Contreras v. Spain 73 Popelier, supra note 24; Steiner, supra note 72 A discussion of the ECtHR’s approach briefly figures in Popelier’s volume, and a discussion of the US approach figures in Steiner’s volume. .

74 Lucky Dev v. Sweden, App. No. 7356/10, Eur. Ct. H.R., November 27, 2014; Demir and Baykara v. Turkey [GC], App. No. 34503/97, Eur. Ct. H.R. 2008-V.

14/50 A first example of absolute retroactivity is Valenzuela Contreras v. Spain, 1998.75 At issue was whether a Spanish court ruling , which in 1990 authorized the monitoring of the applicant’s telephone , was lawful under Article 8 . The domes tic ruling applied the ECtHR’s case-law as it stood at the time of decision.76 Five years later, however, in Kruslin and Huvig v. France, the ECtHR formulated detailed requirements for monitoring to be lawful. These requirements were that “the guarantees stating the extent of the authorities’ discretion and the manner in which it is to be exercised must be set out in detail in domestic law so that it has a binding force which circumscribes the judges’ discretion in the

application of such measures”. 77 The Spanish Government argued that the national court “could not have been expected to know the conditions laid down in the Kruslin and Huvig judgments five years before those judgements were delivered in 1990”. 78 The Court rejected this argument, claiming that “the conditions” formulated in Kruslin and Huvig “stem from the Convention itself .”79 Spain was thus held to have Article 8 and its requirement of foreseeable laws, in a judgement which was hardly foreseeable at the material time.

3.2.3 Aoulmi v. France Another t elling example is Aoul mi v. France, 2006. At issue in this case was whether France violated ECHR Article 34 by carrying out the applicant’s expulsion for drug violations in August 1999, when the Court had requested an interim measure to stay the expulsion under Rule 39 of the Rules of the Court.80 In two cases from 1991 and 2001, the Court had concluded that interim measures under Rule 39 were not binding. 81 In 2005, however, the Grand Chamber overturned these cases. In Mamatkulo v and Askarov v.

France, the Grand Chamber changed its case-law to hold that failure to comply with a Rule 75 Valenzuela Contreras v. Spain, App. No. 27671/95, Eur. Ct. H.R., Judgment (1998). 76 Id. 77 Id. at § 60. 78 Id. at § 60. 79 Id. at § 60.

80 Article 36 reads: The Court may receive applications from any person, nongovernmental organisation or group of individuals claiming to be the victim of a violation by one of the High Contracting Parties of the rights set forth in the Convention or the Protocols thereto. The High Contracting Parties undertake not to hinder in any way the effective exercise of this right.

81 Cruz Varas and Others v. Sweden (Plenary), App. No. 15576/89, Eur. Ct. H.R. March 20, 1991; Čonka v. Belgium, App. No. 51564/99, Eur. Ct. H.R., Judgment 2002-I.

15/50 39 request for interim measures amounted to a violation of Article 34 .82 The French Government argued, in vain, that the applicant’s expulsion had taken place five years before the departure in Mamatkulov and Askarov, and submitted that the Court was required to reach its findings with reference to the “ applicable legal context at the time of the impugned measure.”83 The Court rejected this argument. It stressed that “even though the binding nature of measures adopted under Rule 39 had not yet been expressly asserted at the time of the applicant’s expulsion, Contracting States were nevertheless already required to comply with Article 34 and fulfil their ensuing obligations.” 84 The Cour t

reiterated its jurisprudential principle on precedents, namely that “in the interests of legal certainty and foreseeability it should not depart, without good reason, from its o wn precedents”.85 It rejected, however, to regard these considerations as dispositive. Instead, the Court held that “it was of crucial importance that the Convention be interpreted and applied in a manner which rendered its rights practical and effective, not theoretical and illusory”.86 France was therefore held to have violated Art icle 34 in 1999 because t he Court’s interpretation had changed in 2005, even though the Court as late as 2001 had found that Article 34 could not be violated under similar circumstances. The Court’s

rationale is to render the Convention right, which it adm its did not exist at the material time, “practical and effective, not theoretical and illusory.” 3.2.4 Nicola Silvestri v. Italy Lastly, in Nicola Silvestri c. Italy, 2009, the Court applied a previous departure in its caselaw to hold that Italy violated the right to a fair trial under Article 6 before the change in case-law took place. 87 The crucial issue in Nicola Silvestri was whether an employment dispute about the position as prison director fell within the ambit of Article 6. The civil limb of this article pertains only to proceedings that are determinative of “civil rights”. The Court’s Grand Chamber had previous ly, in Pellegrin v. France, 1999, held that disputes

“raised by public servants” were not determinative of “civil rights”, and thus outside Article 82 Mamatkulov and Askarov v. Turkey [GC], App. Nos. 46827/99, 46951/99, Eur. Ct. H.R. 2005-I. 83 Borgers v. Belgium (Plenary), App. No. 12005/86, Eur. Ct. H.R. (ser. A) 214-B (1991).

84 Aoulmi v. France, App. No. 50278/99, Eur. Ct. H.R. 2006-I, , § 111. 85 Mamatkulov and Askarov v. Turkey [GC], App. Nos. 46827/99, 46951/99, Eur. Ct. H.R. 2005-I. 86 Aoulmi v. France, App. No. 50278/99, Eur. Ct. H.R. 2006-I, § 109.

87 Nicola Silvestri c. Italie, App. No. 16861/02, Eur. Ct. H.R., Judgment (2009).

16/50

6. In Eskelinen and others v. Finland, 2007, the Grand Chamber overruled Pellegrin, and

held that there was a “presumption that Article 6 applies” to labour disputes between civil servants and the State, as there was “no justification for the exclusion” of such disputes from Article 6.88 Nicola Silvestri lodged his complaint with the Court in 2002, five years prior to its departure in Eskelinen . His case had been determined by Italian courts in accordance with Pellegrin. The Italian Government asked the Court to hold the complaint inadmissible since the Court’s case-law did not recognize a right for the complainant under Article 6 at the time the complaint was communicated to the Government .89 The Court rejected this argument, stating that it had “since had the occasion to evolve its case- law

with regard to the applicability of Article 6 § 1 to disputes between the State and State officials”.90 Having found the complaint thus admissible, the Court went on to hold that Italy had violated Article 6. These examples provide three take-aways. First, they show that the Court lends a deaf ear to objections by Contracting States that the departure from previous case -law was not foreseeable at the time of the domestic court decision , which complied with ECtHR caselaw at the time. Second, they showcase how the Court’s backlog produces a considerable lag in time between the final domestic court decision , the departure in case -law, and the Court’s review decision. In all three cases, the departure in case-law took place five years

after the final domestic court’s decision. In Contreras and Silvestri, it took the ECtHR in total eight years to issue its judgments. In the interim measures case in Aoulmni, the Court delayed seven years. This means that while all cases were form ally pending before the ECtHR when the change in case -law occurred, they were far from being new complaints.

The third take-away is that the Court appears to contrast non-retroactivity against dynamic interpretation and efficiency. These aspects will be further discussed in Chapter 5. 3.3 Cases of Limited Retroactivity 88 Vilho Eskelinen and Others v. Finland [GC], App. No. 63235/00, Eur. Ct. H.R. 2007-II, , § 62.

89 Nicola Silvestri c. Italie, App. No. 16861/02, Eur. Ct. H.R., June 9, 2009, § 48. My translation. The original reads: “[...] eu l’occasion de faire évoluer sa jurisprudence relative à l’applicabilité de l’article 6 § 1 aux litiges entre l’Etat et ses agents”.

90 Id. at § 54.

17/50 3.3.1 Introduction There are at least two types of cases where the Court has limited the retroactive effect of new case-law. The first group concerns cases where changes in case-law would affect the outcome of settled inheritance rights, thus affecting third-party reliance interests. The second group regards the legal recognition of transsexuals . I will expound on these two groups in the following.

3.3.2 Cases Balancing Third Party Interests The first case in which the ECtHR limited the retroactive effect of new case-law is Marckx v. Belgium in 1979.91 At issue was whether the unfavorable legal and economic treatment of children born out of wedlock, as compared to children born in wedlock, infringed Article

14. In a previous case in 1967, the Commission had rejected a s imilar complaint as

manifestly ill-founded.92 In Marckx, the Court departed from this decision and held that the unequal treatment violated the Convention. The majority reasoned that there had be en an evolution towards equality between children born in and out of wedlock in Europe , and that the Convention, further to the “living document”-doctrine adopted the year before in Tyrer v. UK ,93 “must be interpreted in the light of present -day conditions”.94 Belgium, assuming such a finding would require the reopening of estates going back to the entry in force of the Convention, requested that the Court limit the retroactive effect.95 The Court responded that it did not examine the contested law “in abstra cto”, and that the

legal effects of its decision were limited to the parties.96 Yet, the Court acknowledged that it was “inevitable that the Court’s decision will have effects extending beyond the confines of this particular case, especially since the violations found stem directly from the contested provisions and not from individual measures of implementation”. 97 Cognizant that the 91 Marckx v. Belgium (Plenary), App. No., 6833/74, Eur. Ct. H.R. (ser. A) 31 (1979).

92 Id. at § 58. 93 Tyrer v. The United Kingdom, App. No. 5856/72, Eur. Ct. H.R. (ser. A) 26 (1978). 94 Marckx v. Belgium (Plenary), App. No., 6833/74, Eur. Ct. H.R. (ser. A) 31 (1979) § 41. 95 Id. at § 58. 96 Marckx v. Belgium (Plenary), App. No., 6833/74, Eur. Ct. H.R. (ser. A) 31 (1979).

97 Id.

18/50 Belgium Government had an “evident interest in knowing the temporal effects of the present judgment”, the Court reasoned as follows: 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" (8 April 1976, Defrenne v. Sabena, Reports 1976, p.

480). The reference to Defrenne v. Sabena is of interest. Here, the European Court of Justice (ECJ) established a doctrine of limited retroactivity for changes in its case-law, which is still good law, provided there is a risk of “serious economic repercussions” and the state has been encouraged in its understanding of EU-law by the Commission or the ECJ.98 The ECtHR went on to state that unequal treatment between “illegitimate” and “legitimate” children had been “permissible and normal” for many years in Contracting States, and that the Commission as recently as 1967 had rejected a similar complaint against Belgium. It then said:

Having regard to all these 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 acts or situations that antedate the delivery of the present judgment.

For many years, this limitation of retroactive effects in Marckx would be written off as an isolated incident in the Court’s otherwise consistent approach to absolute retroactivity. 99 The Court itself would on one occasion dismiss these lines as an obiter dictum.100 A closer reading of the Court’s case-law, however, shows that this understanding is inaccurate. Over the years, other complaints have been lodged with the Court by children born out of wedlock. Here, t he Court has in principle accepted limitations on retroactive effects to 98 Christian Waldhoff, Recent Developments Relating to the Retroactive Effect of Decisions of the ECJ, 46

COMMON MARK. LAW REV. 173–190 (2009).

99 Tulkens and Van Drooghenbroeck, supra note 16 at 259. 100 Unedic c. France (dec.), App. No. 20153/04, Eur. Ct. H.R., December 18, 2008.

19/50 ensure legal certainty for third parties, but it has subjected these limitations to a rigorous proportionality review. In Fabris v. France, 2013, the Grand Chamber considered a cut-off date in a new law which amended previous legislation to grant children “born of adultery” identica l inheritance rights to legitimate children.101 This new law was enacted in 2001, further to the Court’s judgment in Mazurek v. France, 2000,102 where France was held to have violated Article 14 taken in conjunction with Article 1 of Protocol No., 1 because children “born of adultery” were treated different ly than legitimate children for inheritance purposes. The new law, while remedying this discrimination for the future, contained transitional

provisions which restricted application t o successions opened prior to December 4, 2001, the date on which the new law was pronounced.103 The purpose of this cut-off date was to “ensure the stability of completed inheritance arrangements”, which the Court recognized as a “legitimate aim”. 104 While recognizing, in principle, the need for retroac tive limitations, the Court disagreed on how France had balanced the competing interests in reaching the particular cut-off date. The Court concluded that the Convention was violated because the cut-off date went too far in favoring stability over the rights of children “born of adultery”. The Court took pains, however, to emphasize that it did not “call into question

the right of States to enact transitional provisions where they adopt a legislative reform with a view to complying with their obligations under Article 46 § 1 105 of the Convention”.106 However, the State was required “to prevent, with diligence, further violations similar to those found in the Court’s judgments”. 107 The Court stated that domestic courts must “ensure, in conformity with their constitutional order and having 101 Fabris v. France [GC], App. No. 16574/08, Eur. Ct. H.R., 2013-I § 16.

102 Mazurek v. France, App. No. 34406/97, Eur. Ct. H.R. 23 2000-II. 103 Fabris v. France [GC], App. No. 16574/08, Eur. Ct. H.R., 2013-I § 16. 104 Id. at § 66. 105 Article 46 § 1 of the ECHR reads: “The High Contracting Parties undertake to abide by the final judgment of the Court in any case to which they are parties.”

106 Fabris v. France [GC], App. No. 16574/08, Eur. Ct. H.R., 2013-I § 74. The Court referred, as examples, to Antoni v. the Czech Republic, App. No. 18010/06, Eur. Ct. H.R., November 25, 2010; Compagnie des gaz de pétrole Primagaz v. France, App. No. 29613/08, § 18, Eur. Ct. H.R., December 21, 2010; Mork v.

Germany, App. Nos. 31047/04 and 43386/08, §§ 28 to 30 and 54, Eur. Ct. H.R., June 9, 2011; and Taron v. Germany, (dec.), App. No. 53126/07, Eur. Ct. H.R., May 29, 2012. 107 Id. at § 75.

20/50 regard to the principle of legal certainty, the full effect of the Convention standards, as interpreted by the Court.”108 France had failed to do so in Fabris. Most recently, in the joint cases of Wolter and Sarfert v. Germany, 2017, the Court considered the validity of a cut -off date in transitional rules in Germany. 109 At issue was whether the applicants had suffered discrimination under Article 14 taken in conjunction with Article 1 of Protocol No. 1 when denied inheritance because they were born out of wedlock, and their fathers died prior to the cut-off date of a reform law which el iminated such unequal treatment for the future. The reform law had been enacted in 2011, further to

the Court’s judgment in Brauer v. Germany, 28 May 2009, in which the ECtHR held that Germany violated Article 14 taken in conjunction with Article 8 because the legal status of children born out of wedlock excluded them from inheritance on the same footing as children born in wedlock. The reform law was given retroactive effect to include cases where the parent had died after 28 May 2009, the date of the Brauer judgment , but not before. Since the applicants’ fathers passed away before 28 May 2009, the applicants were not recognized as statutory heirs by national courts. The ECtHR accepted, in principle, the need for a cut-off date to protect third-party interests, but drew the line at cases that were

final on the day that Brauer was decided. With reference to Marckx, the Court reasoned that it in such situations, “the inheritance rights of the deceased’s family have acquired legal force and can no longer be changed under national law”. 110 On the other hand, the Court held that where inheritance rights “are still open to be challenged under national law and thus the legal positions to be protected are only ‘relative’, the rights of children born outside marriage should be enforceable”.111 3.3.3 Denial of Compensation in Transsexual Legal Recognition Cases The second group of cases concerns the rights of transsexuals t o legal recognition. In this area, the Court seems to have gone so far as to limit retroactivity altogether.

108 Id. at § 75. 109 Wolter and Sarfert v. Germany, App. Nos. 59752/13, 66277/13, Eur. Ct. H.R., March 23, 2017. 110 Id. at § 64. 111 Id. at § 64.

21/50 At issue in Christine Goodwin v. the United Kingdom, 2002, and I v. United Kingdom , 2002, was whether the United Kingdom violated Article 8 and its protection of “private life” by refusing to legally recognize the new sex of transgendered applicants. In a string of earlier cases, most recently in 1998, the Court had held that the United Kingdom acted within its ma rgin of appreciation when it refused to alter the birth register or issue new birth certificates to transgendered applicants. 112 In Christine Goodwin and I., the Grand Chamber changed its course. 113 It referred to an evolution in European consensus on the matter, and concluded that the “fair bal ance that is inherent in the Convention now tilts

decisively in favour of the applicant”. 114 While the Grand Chamber found that Article 8 was violated, it refused to accommodate the applicants’ request for just satisfaction under Article 41. It reasoned that its holding was a product of dynamic interpretation which did not have retroactive effect, and that “until 1998 similar issues were found to fall within the United Kingdom’s margin of appreciation and that no breach arose.”115 The Court stated:

While there is no doubt that the applicant has suffered distress and anxiety in the past, it is the lack of legal recognition of the gender re assignment of post-operative transsexuals which lies at the heart of the complaints in this application, the latest in a succ ession of cases by other applicants raising the same issues. The Court does not find it appropriate therefore to make an award to this particular applicant.

The finding of violation, with the consequences which will ensue for the future, may in these circumstances be regarded as constituting just satisfaction.116 Four years later, the same cut-off date was applied to reject claims for damages prior to the date of Goodwin and I, July 7, 2002.117 In Grant v. United Kingdom, 2006, the applicant, a “post -operative male-to-female transsexual”, in an identical situation to Goodwin, 112 See the Rees v. the United Kingdom judgment of 17 October 1986, Series A no. 106, the Cossey v. the United Kingdom judgment, cited above; the X., Y. and Z. v. the United Kingdom judgment of 22 April 1997, Reports of Judgments and Decisions 1997-II, and the Sheffield and Horsham v. the United Kingdom

judgment of 30 July 1998, Reports 1998-V, p. 2011 113 Christine Goodwin v. The United Kingdom [GC], App. No. 28957/95, Eur. Ct. H.R. 2002-VI, , § 74. 114 Id. at § 93. 115 Id. at § 119. Both applicants claimed compensation under Article 41, which authorizes the Court to afford, “if necessary, […] just satisfaction to the injured party.” Christine Goodwin claimed pecuniary damage of GBP 38,200 and non-pecuniary damage of GBP 40,000. The other applicant, I., claimed pecuniary damage of GDP 88,181 and non-pecuniary damage of GDP 100,000.

116 Id. at 120. 117 Grant v. The United Kingdom, App. No. 32570/03, Eur. Ct. H.R., Judgment (2006).

22/50 claimed that her right to privacy under Article 8 had been infringed. 118 More specifically, she claimed that the law relating to transsexual persons and an agency’s refusal to give her a retirement pension afforded to women at the age of 60, violated Article 8. The Court rejected any claim arising before Goodwin, and explicitly tied its eventual finding of a violation to the period after Goodwin had been handed down. The Court reasoned that Grant “may claim to be a victim of this aspect of the lack of legal recognition from the moment, after the judgment in Christine Goodwin, when the authorities refused to give effect to her claim, namely, from 5 September 2002.” 119 The Court refused , however,

compensation prior to the date on which Goodwin was pronounced.120 It could be observed that this distinction between a violation and denied compensation ( ex tunc) for damages before case-law shifted has some likeness to the remedial approach to retroactivity in the U.S.

As would often seem to be the case with the ECtHR, however, this distinction has not been applied consistently across the board. In van Kück v. Germany, of June 12, 2003, the Court awarded damages for a similar complaint pertaining to a period before the pronouncement of Christine Goodwin and I on July 11, 2002.121 At issue in Kück was whether proceedings launched by the applicant against her insurance company for reimbursement of medical expenses of a transgender operation, where German courts held that she had failed to prove her surgery was necessary, complied with Articles 6 and 8 of the Convention. Although these proceedings took place between 1992 and 1995, the Court, nonetheless, held that

Germany had “overstepped the margin of appreciati on afforded to them” under Article 8 with reference to “recent developments” in I and Christine Goodwin, pronounced in 2002.122 The applicant was afforded non- pecuniary compensation under Article 41 as a result.

118 Id. 119 Id. at § 43. 120 Id. at § 56. 121 CHRISTOFFERSEN, supra note 24 at 436. 122 Van Kuck v. Germany, App. No. 35968/97, Eur. Ct. H.R., Judgment (2003).

23/50 I turn next to the last category of ECtHR c ases, those in which judges have voiced disagreement about the retroactive effect of new case-law. 3.4 Cases Where Judges Disagreed over Retroactivity 3.4.1 Introduction There are at least three cases where the ECtHR judges have openly disagreed over retroactivity. In all three cases, the domestic court’s decision complied with ECtHR case law at the time. The two first, Borgers v. Belgium, 1991, and Demir and Baykara v. Turkey, 2008, regard retroactivity of a departure from previous case-law in the case at issue. The third, Lucky Dev v. Sweden, 2014, mentioned in Section 2.3.1, regards retroactive application of a shift in case law to a complaint which was lodged after the shift occurred.

3.4.2 Borgers v. Belgium At issue in Borgers v. Belgium was whether the applicant’s right to a fair trial under Article 6 was violated because he could not contradict an opinion by the procureur general that had argued for the dismissal of the applicant’s appeal before the Court of Cassation. 123 A previous case, Delcourt v. Belgium, 1970, had endorsed the same practice, considering the procureur général sufficiently “impartial and independent”.124 The majority in the Borgers Court departed from Delcour. It asserted that the concept of “fair trial” had “undergone a considerable evolution in t he Court’s case -law, notably in respect of the importance attached to appearances and to the increased sensitivity of the public to the fair

administration of justice ”.125 Judge Storme ( ad hoc judge from Belgium) dissented. He lamented that Belgium had relied on the Delcourt judgment as being good law, and that the “fundamental right of a St ate to […] respect for legitimate expectations generated by the Court itself would be seriously infringed by a finding of a violation ex tunc.”126 If the Court should overrule Delcourt, Judge Storme reasoned, it should do so only with prospective 123 Borgers v. Belgium (Plenary), App. No. 12005/86, Eur. Ct. H.R. (ser. A) 214-B (1991).

124 Delcourt v. Belgium, App. No. 2689/65, Eur. Ct. H.R., Judgment (1970). 125 Borgers v. Belgium (Plenary), App. No. 12005/86, Eur. Ct. H.R. (ser. A) 214-B (1991). 126 Id.

24/50 effect. For support, he pointed to case-law from courts in the U.S., the ECJ, and the ECtHR (Marckx).127 3.4.3 Demir and Baykara v. Turkey Judge Storme’s critique would be repeated by Judge Zagrebelsky in Demir and Baykara v.

Turkey, 2008.128 Here, the Grand Chamber held that the Turkish Court of Cassation violated Article 11 when, in 1996, it failed to recognize a right to c ollective bargaining between a trade union and an employer.129 The decision of the Turkish court was consistent with a substantial body of ECtHR case -law, which said that Article 11 did not confer a right to collective bargaining. 130 As late as 2004, eight years after the Turkish judgment, the ECtHR found a similar, and indeed more recent denial of collective bargaining in Sweden, to be in compliance with Article 11.131 The Court departed from this case -law in Demir and Baykara . As grounds for this departure, the majority asserted that international law and domes tic legal systems had

evolved with respect to collective bargaining rights. It held that the right to bargain collectively with the employer “has become one of the essential elements of the right to form and to join trade unions” under Article 11. 132 This new right was then applied retroactively to fault the Turkish court’s decision issued twelve years before. The Court held that the Turkish trade union, “already at the material time [in 1996] enjoyed the right to engage in collective bargaining with the employing authority”.133 Judge Zagrebelsky dissented. He argued that if “new case- law extends the scope of a Convention provision and thus imposes a new obligation [on] States, a retrospective effect 127 Id. Judge Storme wrote: "Often a radical modification of the case-law interpreting a legal rule has been

accompanied by a “prospective overruling,” which was excellently explained in a judgment of the Illinois Supreme Court: “We feel justice will be served by holding that ... the rule herein established shall apply only to cases arising out of future occurrences” (Molitor case, cited by R. Joliet, Le droit institutionnnel des communautés européennes, Liège, 1981, 214; see also similarly: the judgments of Defrenne II (8 April 1976), Gravier (13 February 1985) and Blairot (2 April 1988) of the Court of Justice of the European Communities; see also and in particular the Marckx judgment of the European Court of Human Rights).".

128 Demir and Baykara v. Turkey [GC], App. No. 34503/97, Eur. Ct. H.R. 2008-V. 129 Id. 130 Swedish Transport Workers Union v. Sweden, App. No. 53507/99, Eur. Ct. H.R., Judgment (2006). 131 Id. 132 Demir and Baykara v. Turkey [GC], App. No. 34503/97, Eur. Ct. H.R. 2008-V.

133 Id.

25/50 that is automatic and not subject to directions by the Court would […] be diff icult to reconcile with the requirements of foreseeability and legal certainty, which are essential pillars of the Convention system.” 134 He also warned that retroactive application of expanded rights would make implementation of the Convention by domestic c ourts “difficult, if not impossible.” 135 Zagrebelsky thus argued that “provision be made for the period that precedes the departures from precedent.” In conclusion, he expressed regret that “the Court has once again allowed the “natural” retrospectiveness of judicial interpretation to impugn an approach that, at the material time, was (probably) not in breach of the

Convention.”136 This critique has been worded more strongly by others. Tulkens and Van Drooghenbroeck refer to Demir and Baykara as a “truly Orwellian temporality”, and write that the “retrospective dimension of the departure from precedent is reflected here in all its purity, and if we may dare to say so, in all its violence.”137 3.4.4 Lucky Dev. v. Sweden As I spell out above, the Court most recently split over the issue of retroactivity in Lucky Dev v. Sweden, 2014.138 In this case, the Court’s majority reasoned that by interpreting the prohibition against double jeopardy in Article 4 of Protocol 7 to the Convention in accordance with existing ECtHR case- law at the time , Swedish courts violated the

Convention.139 Three judges disagreed. The background to the case is as follows: In 2004, tax surcharges were imposed by administrative authorities on Lucky Dev for failing to declare business income from 2002. In c riminal proceedings initiated in 2005, Lucky Dev was convicted for tax fraud. Case- law from the ECtHR at the time stated that tax surcharges imposed in conjunction with tax fraud proceedings did not raise an is sue under the double jeopardy rule, since tax surcharges, which do not require guilt, was not the “same offence” as tax fraud, which requires guilt. In Rosenquist v. Sweden, 2004, the ECtHR rejected a similar complaint to Lucky Dev ’s over tax surcharges and double 134 Id.

135 Id. 136 Id. 137 Tulkens and Van Drooghenbroeck, supra note 16 at 254. 138 Lucky Dev v. Sweden, App. No. 7356/10, Eur. Ct. H.R., November 27, 2014. 139 Id.

26/50 jeopardy as “manifestly ill-founded”.140 The tax fraud proceedings against Lucky Dev were finally decided by Swedish courts in accordance with this case-law. Surprisingly, however, the Grand Chamber overturned Rosenquist and similar case-law in Zolotukhin v. Russia, 2009.141 Zolotukhin concerned administrative and criminal sanctions for disobedience in a military camp. Here, the Grand Chamber decided that the notion “same offence” in Article 4, Protocol 7, should be interpreted as comprising different offences based on essentially the same facts, regardless of whether the offences differed in their legal requirements. 142 This meant, by implication, that tax surcharges would be assessed as sanctions for the same of fence as convictions for tax fraud, as these two

offences under Swedish law were based on essentially the same facts. 143 Sweden implemented changes to its systems to comply with Zolotukhin from the date of its pronouncement, February 10, 2009. Since Lucky Dev’s case was finally decided prior to Zolotukhin, she fell outside the change. On January 21, 2010, Lucky Dev lodged a complaint before the ECtHR.144 The Swedish Government responded that the departure in Zolotukhin could not have inserted error in proceedings, such as Luck Dev’s case, which had already been finally decided in accordance with ECtHR case -law when case- law abruptly shifted.145 As I spell out above, the Court adopted a similar cut- off date with respect finally decided cases in Wolfert and Safert. In Lucky Dev, however, the majority rejected this distinction.

It reasoned that “if events in the past are to be judged according to jurisprudence prevailing at the time when the events occurred, virtually no change in case-law would be possible.”146 While the majority “acknowledges that, at the time of the criminal proceedings against the applicant, there had been an earlier decision relating to double proceedings in Swedish tax matters which had concluded that a complaint concerning similar circumstances was manifestly ill- founded (Rosenquist, cited above),” it held that “ the present case mu st nevertheless be determined with regard to the case- law existing at the time of the Court’s 140 Rosenquist v. Sweden, App. No. 60619/00, Eur. Ct. H.R., Judgment (2004).

141 Sergey Zolotukhin v. Russia [GC], App. No. 14939/03, Eur. Ct. H.R. 2009-I. 142 Sergey Zolotukhin v. Russia ([GC], no. 14939/03, judgment of 10 February 2009, ECHR 2009) 143 Lucky Dev v. Sweden, App. No. 7356/10, Eur. Ct. H.R., November 27, 2014.

144 Id. 145 Id. 146 Id.

27/50 examination.”147 Sweden was thus held accountable for not having seco nd-guessed and departed from ECtHR’s case-law before the ECtHR, in an unpredicted move, decided to do so itself.148 Three judges disagreed . In a joint concurring opinion, Judges Villiger (Liechtenstein), Nussberger (Germany), and De Gaetano (Malta), argued that retroactive application of the Court’s case-law after it has undergone a fundamental change “upsets legal certainty and, more specifically, the interaction between the national courts and the Court.” 149 They argued that “[i]t is disruptive for national courts following the Court’s case -law faithfully to find themselves – without any warning – accused of a breach of the Convention. ”150

Moreover, the judges stressed that national courts “are required to implement the Court’s judgments, but not to anticipate changes in the case -law.”151 In conclusion, the y warned that careful consideration of the retroac tivity problem was needed “in order not to undermine the national courts’ trust in the validity of the Court’s authoritative findings”, and called on scholars for “heightened attention” to the issue.

Of the cases discussed above, Lucky Dev is perhaps the most striking example of how absolute retroactivity may upset the administration of justice, and impose significant economic burdens on state s. The case had structural implications for Sweden and other Contracting States with similar systems. Tax offenders, lawfully convicted in compliance with ECtHR case -law at the time, were now victims of double jeopardy violations and entitled to mass remedies for the parallel imposition of tax surcharges and punishment for tax fraud.

3.5 Preliminary Conclusions This syste matization of cases on retroactivity by the ECtHR allows some preliminary

conclusions. First, in some cases, the Court seems to perceive absolute retroactivity as

147 Id. 148 In Sergey Zolotukhin v. Russia ([GC], no. 14939/03, ECHR 2009, the ECtHR reversed the previous case-law (Rosenquist v. Sweden (dec.), no. 60619/00, 14 September 2004) and held that the notion of “same offence” in Article 4 of Protocol No 7 to the Convention should be expanded to a factual as opposed to a legal comparison of the offences in question.

149 Lucky Dev v. Sweden, App. No. 7356/10, Eur. Ct. H.R., November 27, 2014. 150 Id. 151 Id.

28/50 indispensable to dynamic interpretation of the Convention. This view does not, however, bear out in the Court’s own parctice . As is evidenced by the cases on the recognition of transsexuals, and as will be shown in the subsequent part on SCOTUS ’ jurisprudence, limited retroactivity does not exclude evolving interpretations. Rather, it has been claimed that limited retroactivity facilitates changing interpretations, as it mitigates the way in which a change in case -law upsets legitimate expectations and legal certainty. One might even say that limited retroactivity, or restricting legal effects for the future, logically flows from a mode of interpretation which evolves with time.

A second prelimina ry conclusion is that the Court has been attentive to the pro blems of absolute retroactivity when it affects some settled third-party interests. The string of cases following Marckx on discriminatory inheritance rights strike s a balance between cases which are still open to verdict, and those which were final when the case -law shifted. At the same time, the Court’s majorities have remained deaf to the administrative, economic, and structural difficulties of Contracting States faced with absolute retroactive ch anges in case-law. Moreover, the Court has disregarded faithful reliance by domestic courts on previous ECtHR case-law. The recent and separate opinion in Lucky Dev, however, might

suggest an increased awareness of the problems that are likely to ensue from this stance for the Court’s interaction with domestic courts and Contracting States. The next chapter considers how SCOTUS has approached similar issues of retroactivity and changing case-law.

4 SCOTUS’ Approach to Retroactivity 4.1 Overview While there are fundamental differences between SCOTUS and the ECtHR, the former’s approach to retroactivity is of interest to the nascent discussion in Strasbourg for at least three reasons. First, attempts by SCOTUS since 1960 to limit retroactivity , and even to apply new case-law only prospectively, have been subject to extensive scholarly debate.152 152 Francis X. Beytagh, Ten Years of Non-Retroactivity: A Critique and a Proposal, 61 VA. LAW REV.

1557–1625, 1558 with further references (1975).

29/50 Thus far, however, no one has conducted a comparative analysis to assess the transfer value of these doctrines and discussions for the ECtHR.153 Second, SCOTUS’ doctrine for limiting retroactivity in cases on collateral review is of interest, because in some ways habeas corpus resembles the ECtHR’s review with respect to finality of the reviewed court’s decision. Third, SCOTUS’ evolving interpretation of the U.S. Constitution has been aligned with a common law constitutional approach, 154 which the ECtHR’s “living document”-doctrine arguably resembles. The tension between evolving interpretation and retroactive application is thus not unique to Strasbourg. On this basis, it is of interest to see

how SCOTUS has balanced some of the same concerns. This chapter proceeds as follows. Section 4.2 describes the development of retroactivity doctrines in federal criminal cases, with a particular emphasis on habeas corpus proceedings. Second 4.3 describes retroactivity doctrines in federal civil cases. Section 4.4 describes the temporal aspect of the qualified immunity doctrine. Section 4.5 concludes.

4.2 Criminal Cases During an age of progressive interpretation of the Constitution’s criminal procedural guarantees and expansion of the right to habeas corpus, the Supreme Court, under Chief Justice Warren (1953-1969), adopted a doctrine of selective non- retroactivity. Some have explained this doctrine as a pragmatic solution to the potential difficulty of having to grant habeas relief to any person finally convicted who could claim that although lawful at the time, their proceedings now fell short of new ly announced constitutional requirements.155 The solution would be a general discretionary standard to limit the temporal effect of new case-law.

The first case in the movement toward this solution was Linkletter v. Walker, 381 U.S. 681 (1965).156 At issue in Linkletter was whether the landmark case in Mapp v. Ohio,157 which 153 Atrill, supra note 18 at The author discusses merely the prohibition against retroactive criminal law and not adjudicative retroactivity as such.

154 David A. Strauss, Common Law Constitutional Interpretation, 63 UNIV. CHIC. LAW REV. 877–936 (1996). 155 LOW, supra note 35 at 831. 156 Linkletter v. Walker, 381 US 618 (1965). 157 Mapp v. Ohio, 367 U.S. 643 (1961).

30/50 overturned Wolf v. Colerado 158 and made the exclusionary rule under the Fourth Amendment applicable to the States through the Due Process Clause , should apply retroactively to cases finally decided prior to Mapp.159 The Linkletter Court concluded that the Constitution “neither prohibits nor requires retrospective effect”, and that the question of retroactivity should be determined on a case -by-case basis.160 The Court did not limit this holding to final cases on collateral review, but asserted a “general power to make its decisions non-retroactive” – even fully prospective.161 The Court set forth three factors to be considered when deciding whether new case-law applied retroactively: “the prior history

of the rule in question, its purpose and effect, and whether retrospective operation will further or retard its operation”. 162 The Court concluded that the Mapp rule did not apply retroactively, because the deterrent purpose of the rule would not be served by release of convicts due to past police misconduct, and that such “wholesale release” would disrupt “the delic ate state- federal relationship” and “tax the administration of justice to the outmost.” 163 Two years later, in Stovall v. Denno, 388 U.S. 293 (1967), the Supreme Court reformulated the criteria under the Linkletter to the following: “(a) the purpose to be served by the new standards, (b) the extent of the reliance by law enforcement authorities on the old standards,

and (c) the effect on the administration of justice of a retroactive application of the new standards.”164 The Stovall Court held that new case-law would always apply to the case-atpoint, thus ex cluding pure prospectivity. 165 The three-part test would determine whether new case-law applied retroactively to other cases.

The Linkletter/Stovall standard met with “vehement criticism” , as the Court would later put it.166 To some, the non- retroactivity jurisprudence suffered from “doctrinal confusion and inconsistency”. 167 Others argued that non-retroactivity violated Article III of the 158 Wolf v. Colorado, 338 U.S. 25 (1949).

159 Linkletter v. Walker, 381 US 618 (1965). 160 Id. 161 Fallon and Meltzer, supra note 28 at 1740. 162 Linkletter v. Walker, 381 US 618 (1965). 163 Id. 164 Stovall v. Denno, 388 U.S. 293 (1967). 165 Id. 166 Teague v. Lane, 489 U.S. 288 (1989).

167 Haddad, supra note 35 at 1062 with further references.

31/50 Constitution, as it gave the Court a “law-making function”.168 Conservative justices on the Court worried that the discretion to limit retroactive effects of a new rule, made it too easy to change and expand the Constitution. 169 Yet others argued that the standard produced unequitable results, as it denied some prisoners of past and pending proceedings newly announced Constitutional rights .170 Paul Mishkin argued that all new announcement of case-law should be fully retroactive on direct review, but not retroactive in the extraordinary form of review through habeas corpus.171 “Upon reflection”, Justice Harlan became convinced of Mishkin’s point that new case-law should be given full retroactivity in cases pending di rect review, but not in cases on

collateral review.172 In two dissenting opinions in Desist v, United States, 394 U.S. 244, 256 (1969) and Mackey v. United States, 401 U.S. 667, 675 (1971) , Harlan proposed to distinguish between cases pending direct review a nd habeas corpus. 173 He reasoned that non-retroactivity would normally be justified in habeas corpus cases, because one principal purpose of the review was to incentivize lower courts to apply the Convention correctly.

He explained that the threat of habeas serves as a necessary additional incentive for trial and appellate courts throughout the land to conduct their proceedings in a manner consistent with established constitutional standards. In order to perform this deterrence function, the habeas court need not […] necessarily apply all "new" constitutional rules retroactively. In these cases, the habeas court need only apply the constitutional standards that prevailed at the time the original proceedings took place.

Harlan’s reasoning would later prevail. In Griffith v. Kentucky, 479 U.S. 314 (1987) , the Supreme Court under Chief Justice Rehnquist , established full retroactivity for all “criminal cases pending on direct review” at the time when new case-law is pronounced.174 168 AMERICAN TRUCKING ASS’NS V. SMITH, 496 U.S. 167 (1990), supra note 45 at 2343 (Scalia, J.

concurring in the judgment). 169 Fallon and Meltzer, supra note 28 at with references to cases. 170 Mishkin, supra note 35. 171 Id. 172 Desist v. United States, 394 U.S. 244 (1969). 173 Desist v, United States, 394 U.S. 244, 256 (1969) (Harlan, J., dissenting opinion) and Mackey v. United States, 401 U.S. 667, 675 (1971) (Harlan, J., opinion concurring in judgments in part and dissenting in part).

174 Griffith v. Kentucky, 479 U.S. 314 (1987).

32/50 The Court held that the “nature of judicial review” deprived courts of the legislative prerogative to regulate the temporal effects of new rules, and that selective application of new rules violated the principle of treating equally situated parties the same.175 In Teague v. Lane, 489 U.S. 311, the Court endorsed the second principle of Harlan’s opinion; non-retroactivity in habeas corpus proceedings with the narrow exception of decriminalized conduct or watershed changes in case- law.176 The non-retroactivity rule in habeas corpus cases would be operationalized as a threshold question, where t he case would be dismissed if the petitioner relied on a new rule or asked the court for a new rule.177

The Teague Court adopted Harlan’s rationale for this limitation, arguing that the central purpose of the extraordinary remedy of habeas corpus was to supply “a necessary incentive for trial and appellate judges throughout the land to conduct their proceedings in a manner consistent with established constitutional principles.”178 4.3 Civil Cases The development of retroactivity doctrines in civil cases has to some extent followed criminal law. Today, SCOTUS applies new case-law retroactively to cases pending direct review, but not to cases that were finally decided at the time. I will briefly trace the se jurisprudential developments in the following.

As was the case with retroactivity in criminal cases, Justice Harlan would early on signal the course the Court would eventually take two decades later. In United States v. Estate of Donnelly, 397 U.S. 286 (1969), he argued for full retroactivity in pending cases, and nonretroactivity in finally decided cases.179 He reasoned as follows:

The impulse to make a new decisional rule nonretroactive rests, in civil cases at least, upon the same considerations that lie at the core of stare decisis, namely to avoid jolting the expectations of parties to a transaction. Yet, once the decision to abandon precedent is made, I see no justification for applying principles determined to be wrong, be they constitutional or otherwise, to litigants who are in, or may still come to, court. The critical factor in determining when a new decisional rule 175 Id. at 323.

176 Teague v. Lane, 489 U.S. 288 (1989). 177 Id. 178 Id. at 306. 179 United States v. Estate of Donnelly, 397 U.S. 286 (1969).

33/50 should be applied to a transaction consummated prior to the decision's announcement is, in my view, the point at which the transaction has acquired such a degree of finality that the rights of the parties should be considered frozen. Just as in the criminal field, the crucial moment is, for most cases, the time when a conviction has become final, see my Desist dissent, supra, so, in the civil area, that moment should be when the transaction is beyond challenge either because the statute of limitations has run or the rights of the parties have been fixed by litigation and have become res judicata. Any uncertainty engendered by this approach should, I think, be deemed part of the risks of life.

These lines did not convince the Court at the time, however. In Chevron Oil Co. v. Huson, 404 U.S. 97 (1971) , the Court adopted a three -part test resembling the Stovall -test for determining whether a new rule of judge-made law should apply retroactively to cases pending direct review. 180 Under this test, the Court would consider: (i) that the decision indeed establishes a new rule of law; (ii) examine the rule’s history, purpose, and effect to ascertain whether retroactive operation would further or retard its operation; and (iii) weigh the inequity imposed by retroactive application. 181 While the Chevron Oil test is still applied by several state courts, the Supreme Court would depart from this standard in three

taxation cases in the 1990s .182 At issue in the se cases was whether taxpayers could claim refund of state tax paid under a statute later declared unconstitutional. In the first, American Trucking Associations, Inc. v . Smith, 110 S. Ct. 2323 (1990), the Court held that a tax-payer was not entitled to a refund (or any other retroactive relief) for taxes paid before the law-changing decision.183 Four judges applied the Chevron test, and four others dissented in favor of deciding the case “under the best current understanding of the law”. Justice Scalia concurred in upholding the statute, but sided with the dissenters on retroactivity.184 His declaratory grounded objections to retroactivity are discussed above.

In the second case, James M. Beam Distilling Co. v. Georgi a, 501 U.S. 529 (1991), five justices voted in a plurality opinion for the proposition that a new Constitutional interpretation should apply retroactive ly to all other cases still pending review .185 In the 180 Chevron Oil Co. v. Huson, 404 U.S. 97 (1971).

181 Id. 182 Kay, supra note 35. 183 American Trucking Ass’ns v. Smith, 496 U.S. 167 (1990). 184 Id. (Scalia, J., concurring). 185 James M. Beam Distilling Co. v. Georgia, 501 U.S. 529 (1991).

34/50 third and final case, Harper v. Va. Dep’t of Taxation, 509 U.S. 86, 97 (1993) , the Court prohibited “selective temporal barriers to the application of federal law in noncriminal cases”, and established that a new rule “is the controlling interpretation of federal law and must be given full retroactive effec t in all cases still open on direct review and as to all events, regardless of whether such events predate or postdate our announcement of the rule.”186 4.4 Qualified Immunities Doctrine Finally, under the qualified immunities doctrine for damages claims agains t government officials, the Court has limited the application of new changes in case-law to older facts.

In Harlow v. Fitzgerald, 457 U.S. 800 (1982), the Supreme Court held that government officials “performing discretionary functions are shielded from l iability for civil damages insofar as their conduct does not violate ‘clearly established’ statutory or constitutional rights of which a rea sonable person would have known” at “the time [the] action occurred”.187 The Court reasoned that “[i]f the law at that time was not clearly established, an official could not reasonably be expected to anticipate subsequent legal developments, nor could he fairly be said to “know” that the law forbade conduct not previously identified as unlawful.” This doctrine thus allow s government officials to “reasonably anticipate

when their conduct may give rise to liability for damages.”188 It balances the “public interest in deterrence of unlawful conduct and in compensation of victims”, with the need to “avoid excessive disruption of government”.189 A central question under this doctrine is when the law is “clearly established”. The Court has required that the “unlawfulness must be apparent”.190 In other words, “ the contours of the right must be sufficiently clear that a reasonable official would understand that what he is doing violates that right.” 191 The test operates as a threshold question to bar suits seeking to vindicate newly interpreted constitutional rights.192 186 Harper v. Virginia Dept. of Taxation, 509 U.S. 86 (1993).

187 Harlow v. Fitzgerald, 457 U.S. 800 (1982). 188 Id. 189 Id. 190 Anderson v. Creighton, 483 U.S. 635 (1987). 191 Id. 192 Fallon and Meltzer, supra note 28 at 1753.

35/50 4.5 Summary In sum, this brief overview has shown that the U.S. Supreme Cour t limits retroactive application of new case-law in several aspects. One key takeaway for the purposes of this paper is the emphasis put on whether retroactivity would have prevented or deterred unwanted actions in the past. In criminal law, the Court applies new changes retroactively to all cases pending direct review, but normally refuses to apply or develop new case- law when reviewing finally decided cases under habeas corpus. In civil law, the Court applies new case-law retroactively to pending cases, b ut not to finally adjudicated cases. In civil damages cases against government officials, the Court assesses whether the action

infringed “clearly established law” at the material time. Do SCOTUS’ various approaches to retroactivity, and the debate these approaches have spurred amongst scholars, have any applicability for the ECtHR? I argue that they do. In my view, the distinction between cases pending direct review and finally decided cases on collateral review is of particular relevance to the ECtHR. Moreover, the distinction between interpretation and redressability offers a functional disentanglement of the more disruptive aspects of case-law departures, which the ECtHR should consider going forward.

5 A Non-Retroactivity Approach for the ECtHR? 5.1 Preview of the Argument I propose that the ECtHR should adopt a similar distinction as SCOTUS in Teague, mutatis mutandis, to preclude retroactivity in cases that were finally decided by domestic courts in accordance with ECtHR case-law before the change in case-law took place, and where the situation complained of is not ongoing. This distinction would further the principles of primacy for domestic implementation and subsidiarity of ECtHR supervision, which were recently emphasized by all 47 Contracting States in the Copenhagen Declaration.193 Under this rule, new changes in case- law would apply retroactively to cases pending before domestic courts at the time of the change, and to cases that were finally but incorrectly

193 Copenhagen Declaration, supra note 13.

36/50 decided with respect to existing ECtHR case-law. In pending cases, domestic courts would be in position to decide the matter in acco rdance with new ECtHR-case law. In wrongly decided cases, institutional considerations for deference to primary domestic implementation would not apply, while individual interests in corrective justice would favor the possibility for retroactive revie w. Moreover, I propose that when the Court departs without warning from settled and domestically implemented case-law, it should do so with ex nunc and prospective effect, and deny remedies with retroactive effect. For ease of understanding, a flow chart of the proposal, at the time of change, could look like this:

Under this scheme, plaintiffs would be guaranteed the leve l of protection which the ECtHR demanded when the case was decided by domestic courts. Domestic courts would be required to interpret the Co nvention in compliance with ECtHR’s case -law, without running the risk of being defaulted for doing so. The ECtHR could intervene in ongoing situations now deemed to v iolate the Convention. And, the Court could depart from previous case-law, but for the future, without upsetting the “shared responsibility” with domestic courts which is vital to the “proper functioning of the Convention system” and “efficient protection of human rights”.194 There are several wrinkles with this proposal , however. I w ill address some of these

wrinkles, and trade -offs, in the following. Section 5.2 discusses the worry that nonretroactivity would preclude dynamic interpretation. Section 5.3 discusses the distinction between pending cases before domestic courts and the ECtHR. Section 5.4 tests the 194 Id.§ 6.

37/50 proposed rule against the underlying considerations for and against retroa ctivity, more broadly. Section 5.5 responds in some measure to the objection that a case-by-case standard would be preferable to a per se-rule. Section 5.6 concludes.

5.2 Would Non-Retroactivity Preclude Dynamic Interpretation? One objection to non- retroactivity is that it would hamper dynamic interpretation of the Convention as a “living instrument”.195 Recall Lucky Dev, where the majority claimed that “virtually no change in case-law would be possible” if “events in the past are to be judged according to jurisprudence prevailing at the time when the events occurred”.196 I argue that this argument is flawed.

Cases from SCOTUS, and some cases from the ECtHR ,197 show that courts can change their case-law while limiting the temporal effects of such changes. Non-retroactivity can even be seen to facilitate jurisprudential departures. One critique made against nonretroactivity doctrines in the U.S. was precisely that it made it too easy for courts to change the law. 198 On this view, non- retroactivity allowed judges to mitigate the practical implications of de partures from precedents which w ould otherwise weigh against overruling. Others have argued for non-retroactivity for the same reason, as it reduces the transaction costs related to changing a rule , thereby solving a path- dependence problem

which could otherwise have hampered timely development of the law. However, it might be that non-retroactivity, in practice, would hamper evolution of caselaw to the extent it discourages complaints. Some U.S. commentators have warned that non-retroactivity removes incentives to litig ate, thus frustrating the development of caselaw.199 Yet, this concern should not be overstated. First, the rule proposed would only limit retroactivity in cases where domestic courts have applied previous case- law correct. This leaves out the large number of cases where domestic courts fail to do so, or where the case presents a new question under the Convention. Moreover , since the Court as proposed 195 Lucky Dev v. Sweden, App. No. 7356/10, Eur. Ct. H.R., November 27, 2014.

196 Ibid. 197 See for instance Christine Goodwin v. The United Kingdom [GC], App. No. 28957/95, Eur. Ct. H.R. 2002-VI. 198 Stephens, supra note 35 at 1565 with further references to Scalia. 199 Ghatan, supra note 35.

38/50 could overturn previous case -law by finding a violation ex nunc , incentives to complain would still be significant.200 In sum, the concern that non- retroactivity prec ludes dynamic interpretation is not convincing.

5.3 The Distinction Between Pending Cases Another objection to the non- retroactivity rule proposed above is that it does not include pending cases before the ECtHR at the time of the change. On this view, new case -law should apply retroactively also to cases that were pending before the ECtHR when the shift occurred, as they would be in the same situation as the case which served as vehicle for the change. This view finds support in the Court’s case- law, by which the Court treats other complaints that were pending before it at the time of departure in the same way as the successful complainant because they are in “exactly the same position”.201 This argument, however, assumes equa lity across relevant dimensions. Yet, the heterogeneity of the

Court’s case-load, arising from 47 Contracting States, and the substantial delays caused by the Court’s backlog, suggest that retroactivity of new case-law for all pending ECtHR cases would not necessarily generate more evenhanded results.202 A middle position sometimes heard is that the Court should apply changes retroactively to pending ECtHR cases at the time, but not to finally decided domestic cases that were not pending before the ECtHR whe n case-law shifted. This was the solution proposed by the three minority judges in Lucky Dev. 203 This distinction, however, would produce an anomaly whereby changes would a pply retroactivity to the oldest group of cases pending before ECtHR at the time, but not to more recent cases where a complaint had not yet been

lodged in Strasbourg. 200 See Fallon and Meltzer, supra note 28 at 1806. 201 Lucky Dev v. Sweden, App. No. 7356/10, Eur. Ct. H.R., November 27, 2014, with further references to Dovguchits v. Russia, no. 2999/03, § 24, 7 June 2007; Redka v. Ukraine, no. 17788/02, § 25, 21 June 2007;

Rizhamadze v. Georgia, no. 2745/03, § 27, 31 July 2007; Ştefanescu v. Romania, no. 9555/03, § 20, 11 October 2007; and Vanjak v. Croatia, no. 29889/04, § 32, 14 January 2010. 202 Statistics of the ECHR - Analysis of statistics, Thematical statistics, Facts and Figures, https://www.echr.coe.int/Pages/home.aspx?p=reports&c=#n1347956867932_pointer (last visited Apr 20, 2018). The Court’s backlog peaked 151,600 in 2012. Per 2018 it is at 56,250 cases.

203 Lucky Dev v. Sweden, App. No. 7356/10, Eur. Ct. H.R., November 27, 2014.

39/50 I argue that non- retroactivity of departures from case-law for the law-changing decision, would solve some of these difficulties. As the Court did in Goodwin and I, it could find a violation ex nunc , but deny retroactive remedies before the date of the law -changing judgment. This would assure equal treatment between the complainant in the law-changing decision and others going forward. Moreover, it would be consistent with the principle of subsidiarity and primary domestic implementation. It is widely recognized that the ECtHR is not a fourth instance court, but a subsidiary “safeguard for individuals whose rights and freedoms are not secured at the national level”. 204 Non-retroactivity for pending cases

decided in accordance with previous case-law would be taking these “limits imposed on its action” seriously.205 In addition, it would allow the Court to dispose of those cases in its back-log which would not have resulted in findings of violations had they been decided within a reasonable time, and focus its attention on cases where domestic authorities and courts have not secured Convention standards.

The next section discusses a rule on non-retroactivity with respect to the competing policy interests in play, more broadly. 5.4 Assessing the Underlying Principles 5.4.1 Considerations Advanced for Retroactivity I expound first on considerations often advanced for retroactivity, namely corrective justice, efficiency, and equality, before I turn to considerations advanced for nonretroactivity, legal certainty, reliance, finality, and lack of deterrence.

5.4.1.1 Corrective Justice A first concern against non-retroactivity is that it would deny individuals the benefit of new judge-made rules. We ma y call this an interest in corrective justice, or an accurate application of the law. When the Court interpret s rights in new, expansive ways with absolute retroactivity, it gives individuals the benefit of remedies for past state action or 204 Copenhagen Declaration, supra note 13§ 28.

205 Centro Europa 7 S.R.L. and Di Stefano v. Italy [GC], App. No. 38433/09, Eur. Ct. H.R. 2012, § 197.

40/50 inaction w hich the Court has come to view as a rights violation. As such, absolute retroactivity has been championed to enhance individual rights. Along these lines, Katia Lucas-Alberni claims that “prospective overruling, even in order to preserve the expectations of a respondent State in exceptional cases, is at variance with the function of the [ECtHR]”.206 She contends that “protection of fundamental rights does not lend itself to any compromise, and nor indeed does the protection machinery”.207 There is nothing inherent , however, in absolute retroactivity that maximizes individual rights. Retroactive changes may as well restrict a previous interpretation, thus granting individuals fewer rights than before. After Lucky Dev, for instance, the Grand Chamber in

A and B v. Norway changed its case-law on the double jeopardy rule again, to allow for the kind of double tax fraud sanctions that the Lucky Dev Court held violated the Convention.208 Tulkens et al. point to other examples in ECtHR case-law where the Court has restricted , rather than expanded, rights. 209 Presumably, proponents of absolute retroactivity would not favor retroactivity of restrictive case-law. The argument that absolute retroactivity is to be preferred over non- retroactivity because it f urthers human rights protection is thus restricted to situations where the new interpretation does in fact expand rights. As such, the objection is substantially outcome-driven and should not be

dispositive for the retroactivity problem per se. Yet, to the extent new interpretations do expand rights and freedoms in the Convention, and most often to date that has been the case with ECtHR’s case-law, non-retroactivity would deny individuals and companies the additional protection which the ECtHR has later come to view as required by the Convention. This objection to a rule on non-retroactivity would presumably be more or less important depending on the type of right (to be discussed infra).

A related objection against non-retroactivity which seems less dependent on outcome, is the Dworkian view that judges ought to decide cas es pursuant to their best understanding of the law, regardl ess of when the facts occurred or when the issue was resolved by 206 LUCAS-ALBERNI, supra note 70 at 383.

207 Id. at 382. 208 A and B v. Norway [GC], App. Nos. 24130/11, 29758/11, Eur. Ct. H.R. 2016. 209 Tulkens and Van Drooghenbroeck, supra note 16 at 280 with further references to case-law.

41/50 domestic courts. This best-rule thesis presupposes that there is only one correct answer, not two best rules at different points in time.210 The problem of absolute retroactivity in the context of the E CtHR, however, is that of two, not one, possible rules. Domestic courts, deciding the case in accordance with existing ECtHR case-law, apply what is thought to be the best rule at the time. They resolve the case in accordance with how the ECt HR would have judged it then. Had the domestic court decided the case differently, it would have violated the obligation to interpret the Convention in accor dance with the ECtHR’s caselaw. The choice would have been between complying with rule X, or infringing the

obligation to interpret the Convention correctly by adopting a new rule Y . Insofar as domestic courts should apply the case- law of the ECtHR faithf ully, and leave it to the ECtHR to develop the Convention, it seems that the best rule for domestic courts is to apply rule X.

5.4.1.2 Efficiency and Equality Another consideration “generally viewed as favoring retroactivity” is efficiency.211 On this view, non-retroactivity of new ECtHR case -law would hinder efficient realization of human rights protection, because Contracting States would only be required to prevent violations going forward from the law -changing decision, and not also remedy similar actions or ina ctions in the past. This concern should not be overstated, however.

Contracting States would at no point be in a position to disregard the Convention. Rather, they would be required to comply with how the ECtHR currently interprets the Convention. A third objection to non-retroactivity is that it produces inequality.212 Non-retroactivity is said to treat similarly situated plaintiffs differently, based on arbitrary circumstances such as when the transaction occurred or when the facts of dispute were finally decided in court.

Yet, the assumption that retr oactivity promotes equality, and non- retroactivity produces inequality, is too simplistic . As Lehn observes, equality is a n inherently “neutral principle”.213 Similarly situated parties can be treated equally by non-retroactivity as well 210 Roosevelt, supra note 17 at with further references.

211 Fisch, supra note 35 at 1088; Tulkens and Van Drooghenbroeck, supra note 16 at 268. 212 Stephens, supra note 35 at 1561. 213 Lehn, supra note 54 at 580.

42/50 as by retroactivity. Take cases with opposing private interests, like civil disputes, 214 or, in the context of the ECtHR, complainants challenging state action or inaction that intersects with conflicting third party interests. One such example is discussed above, namely cases where retroactive application of new (expansive) discrimination law would favor plaintiffs over third party relatives, or their successors, while non-retroactivity will favor third party interests over the p laintiff.215 This argument could be made more broadly, however. As Martti Koskenniemi puts it; “[e]very increase of the right to freedom of somebody is felt as an encroachment on the right to security of someone else. And so on.”216 Arguably, most

cases before the ECtHR could be said to involve some sort of balancing of opposing individual interests.217 Retroactive application of new case-law in a child welfare case, for instance, might favor the biological parent over the adoptive parents, or even the child, or vice versa. Retroactive application of new case- law in an eviction case could favor the neighbors over the evicted, or vice versa.

Moreover, inequality arises because the Court changes its case- law, not becaus e of the temporal limitations per se. The three judges writing separately in Lucky De v alluded to this point when they wrote that “every change in the case- law will inevitably bring abo ut situations of inequality as ‘new’ applicants are treated differently from ‘old’ ones. This inequality cannot be a voided wherever the dividing line is drawn.” 218 To illustrate this point, consider the following hypothetical: A complaint with similar facts as Lucky Dev comes before the ECtHR at the same time Lucky Dev’s case is decided by the highest Swedish court. For identification, let us call our imagined complainant Unlucky Dev. Like

the Swedish Court, t he ECtHR would presumably reject her claim that there had been a violation of the double jeopardy rule, as it did in its previous decision on a similar case in Rosenquist. When the ECtHR later changes its interpretation in Zolotukhin, Unlucky Dev would not have had a second shot at the ECtHR. Under Article 35 (2) b, the Court “shall not deal with any application” that is “substantially the same as a matter that has alr eady 214 Fisch, supra note 35 at 1085.

215 See section 3.3.2. above. 216 Martti Koskenniemi, Rights, History, Critique, in HUMAN RIGHTS: MORAL OR POLITICAL? , 55 (Adam Etinson ed., 2018). 217 Koskenniemi, supra note 216. 218 Lucky Dev v. Sweden, App. No. 7356/10, Eur. Ct. H.R., November 27, 2014.

43/50 been decided by the Court”. She would not have been compensated for tax surcharges levied on her, or the conviction served for tax fraud. Contrast Lucky Dev, and recall that Lucky Dev’s case was decided domestically at the same time that the ECtHR, in our hypothetical, decided the other case. When Lucky Dev lodges a complaint with the ECtHR, after the case-law change in Zolotukhin, the ECtHR applies this new law retroactively. On this basis, Sweden is found to have violated t he Convention, and Lucky De v, unlike Unlucky Dev, receives compensation for the double sanctions imposed on her for tax fraud.

This hypothetical illustrates that a bsolute retroactivity, just as non -retroactivity, may produce unequal results, based on when the facts occurred or when they were decided. Neither can fully produce equality, because the difference in treatment is caused by the change in case-law itself.

In sum, t he argument t hat retroactivity produces more individual justice, efficiency, and equality is far fr om watertight, and to some extent determined by outcomes. This observation should caution against placing too much weight on these concerns either way.

I turn next to discuss considerations in favor of a non-retroactivity rule. 5.4.2 Considerations Advanced in Favor of Non-Retroactivity 5.4.2.1 Legal Certainty The central principle underlying non-retroactivity is legal certainty.219 Lord Rodger calls it a “fundamental principle of most legal systems” that the “legal consequences of situations should be judged according to the law in force at the time.” He views legal certainty as part of a “wider principle that people should be judged by their contemporaries and by contemporary standards and beliefs, rather than many years later by people with perhaps very different standards and beliefs.”220 One manifestation of legal certainty is reliance. 221 Non-retroactivity furthers reliance by

safeguarding legitimate expectations from unpredictable changes in the future. Whi le 219 JURATOWITCH, supra note 36. 220 Alan Rodger, A time for everything under the law: some reflections on retrospectivity., 121 LAW Q.

REV. 57–79, 63 (2005).

221 Steiner, supra note 72 at 2.

44/50 individual complainants may rely on new ECtHR case- law, this form of reliance is not at issue here insofar it is formed after the material facts and after the final domestic court decision. The reli ance interests we consider are instead those of third -parties, states, or domestic courts. Where a case is between an individual com plainant and the Contracting State, without directly implicating third party interests , absolute retroactivity could nonetheless indirectly challenge the situation of other parties whose interests the State’s position aims to protect. In some situations, m oreover, the disruptive consequences of absolute retroactivity on states can be far reaching. 222 Challenges against taxation laws or

administrative tax penalties, such as the Lucky Dev case, can have particularly widespread ramifications for state finances i n the event of a retroactive finding of a violation. Here, retroactivity could upset settled state finances or undercut settled budgetary or public interests. In view of how retroactivity in tax cases may seriously affect public finances, Eva Steiner observers that “there is a general consensus amongst [national] jurisdiction that these claims should be dealt with prospectively only.” 223 Fallon and Meltzer argue more generally that “a remedial calculus should attend to the potentially disruptive consequences of [retroactive] remedies.” 224 Finally, absolute retroactivity could undermine “the trust of

the national courts in the reliability and persistence of the Court’s case -law”.225 The toll absolute retroactivity places on the interaction with domestic courts will be discussed more in detail infra. A second manifestation of legal certainty is finality. To Stephens, finality is the “notion that there must be an endpoint to litigation or to the possibility of litigation”.226 It underlies a rule of non- retroactivity, b ecause the retroactive application of new ECtHR case- law would revive domestic court decisions that are final. Finality interests shield the status quo from later disturbance, and promote the institutional authority of a court judgment. Finality is thus tied to the doctrines of res judicata, in that l egal positions gain protection because

time has passed. Some might protest that the mere passage of time should not be dispositive with regard to retroactivity. T his argument, however, fails to appreciate that pass age of 222 Id. at 19. 223 Id. at 19. 224 Fallon and Meltzer, supra note 28 at 1793.

225 Lucky Dev v. Sweden, App. No. 7356/10, Eur. Ct. H.R., November 27, 2014. 226 Stephens, supra note 35 at 1565.

45/50 time is dispositive on the merits in many circumstances, for instance under statutes of limitation, or preclusion. Some have even argued that retroactivity should be treated as a mere problem of preclusion.227 Two comments are warranted with respect to the ECtHR and finality. First, passage of time is for practical reasons perhaps more significant here than elsewhere, since the Court’s backlog is such that it takes years from the time a complaint is lodged until the Court decides on the admissibil ity and/or merits. It is telling that the Court’s ambition, as established in the Brighton Declaration in 2012, was to decide a case within two years of communication. Far from achieving this ambition, however, it is not unusual for cases to

take between five and ten years to be decided. This means th at absolute retroactivity can infringe interests of legal certainty that relate to facts arising more than a decade before the ECtHR’s decision. Second, the concern for finality interests is p erhaps less straightforward in the context of an international court than within a domestic court system. As individuals and legal entities may bring a complaint over a final domestic court decision before the ECtHR within six months of the domestic court’s decision, one m ight argue that the finality of the domestic court’s decision is only relative. Yet, this view runs up against the general acknowledgment that the ECtHR is not, and should not act as , a court of fourth instance. 228 Therefore, a

complaint lodged with the ECtH R does not affect the finality of the domestic decision. Moreover, inasmuch as absolute retroactivity takes on an erga omnes effect, it would disrupt not only complaints before the ECtHR, but also the finality of other final domestic court decisions that have solved similar issues on the basis of the same departed reasoning in ECtHR case-law.

5.4.2.2 Interaction with Domestic Courts Related to this concern for legal certainty is the worry that absolute retroactivity upsets the interaction between domestic courts and the ECtHR, thereby infringing the “shared responsibility” for safeguarding the Convention. The judges writing separately in Lucky 227 Lehn, supra note 54 at 585.

228 Centro Europa 7 S.R.L. and Di Stefano v. Italy [GC], App. No. 38433/09, Eur. Ct. H.R. 2012 § 197; Copenhagen Declaration, supra note 13 at § 6.

46/50 Dev cautioned that it is “disruptive for national courts following the Court’s case -law faithfully to find themselves – without any warning – accused of a breach of the Convention.”229 They argued that domestic courts are “required to implement the Court’s judgments, but not to anticipate changes in the case law.” Absolute retroactivity, in their view, could “undermine the national courts’ trust in the validity of the Court’s authoritative findings”.230 The Draft Copenhagen Declaratio n also emphasized this concern that “national authorities can apply and enforce the Convention principles at domestic level and rule with certainty on the situations submitted to them without running the risk of

subsequent disavowal.” 231 To solve this tension, the Draft Declaration proposed that the Court take positions that are “stable and coherent”. In its Opinion on the Draft Copenhagen Declaration, the Court’s responded that there is “no formal doctrine of precedent in the Convention jurisprudence”, and insisted that it may depart from previous case -law when there are “strong reasons” to do so. 232 A rule on non- retroactivity promises to solve this fundamental disaccord between Contracting States and the Court without curtailing the Court’s prerogative in developing the Convention. Non-retroactivity would meet the same end of avoiding disruptive changes in domestic implementation, while preserving the room

for dynamic interpretation. 5.4.2.3 Lack of Deterrence Finally, it is worth pausing to note that one’s view on whether non-retroactivity furthers or hampers the interaction between the ECtHR and domestic courts is part of a larger question on the purpose of the ECtHR’s human rights supervision. Is it to ensure accurate application of the evolving standards of the Convention to individuals, or is it to ensure that domestic courts apply the Convention correctly at the time of their review? As I spell out above, the corrective justice approach supports retroactivity only where new case- law expands rights. The principles of primacy for domestic courts and subsidiarity for the ECtHR suggest that ensuring correct domestic implementation should take priority. If the

purpose of ECtHR supervision is to supervise that domestic court s apply the Convention 229 Lucky Dev v. Sweden, App. No. 7356/10, Eur. Ct. H.R., November 27, 2014. 230 Id. 231 Draft Copenhagen Declaration, supra note 19 at 57.

232 European Court of Human Rights, supra note 20.

47/50 faithfully, non- retroactivity would properly confine the review to the relevant legal circumstances at the time. The deterrence function of the ECtHR review is not served by inserting error in domestic proceedings that complied with authoritative understandings of the Convention when conducted.

The position of the U.S. Supreme Court is of interest in this regard. Based on views that retroactivity fails to deter conduct th at was lawful at the time of action, SCOTUS limit retroactive application of new case- law in three situations. First, in federal habeas corpus cases, new case-law is not applied retroactively to review state court decisions, save in two very limited circum stances, because it would not further the “deterrence function” of incentivizing “trial and appellate courts throughout the land to conduct their proceedings in a manner consistent with established constitutional standards.” Second, in cases pending direct review with respect to exclusion of evidence obtained in violation of the 4 th

Amendment, new case- law is not applied retroactively to searches that took place in objective reliance on binding precedent at the time.233 The reason is that new law could not have deterred actions that were, by good faith, lawful at the time. Third, in cases pending direct review with respect to qualified immunity, the Supreme Court limits liability insofar as the officer in question relied on clearly established law at the time. The reason is again that retroactivity of new law would not have deterre d the action in question. Fallon and Meltzer argue that the argument for corrective individual justice carry “diminished force when the government can plausibly assert that, in view of a rule’s newness, its violation

involved no fault on the part of government officials.” 234 In such situations, there is “less need to impose a ‘penalty’ to deter future misconduct”.235 Similarly, with respect to the ECtHR, retroactive application of new case -law to overturn what was at the time correct understandings of the law , would not deter future rights violations, but faithful implementations upon which the proper functioning of the human rights machinery is based. The ECtHR, rather than defaulting, should incentivize good faith 233 Davis v. United States, 564 U.S. 229 (2011).

234 Fallon and Meltzer, supra note 28 at 1793. 235 Id. at 1793.

48/50 application of the Convention in accordance with ECtHR case -law by domestic courts. Non-retroactivity would allow it to do so. On this basis, I claim that legal certainty, interaction with domestic courts, and lack of meaningful deterrence, strongly favor non-retroactivity. It is not obvious, however, that a doctrine on non-retroactivity should take the form of a bright-line rule as opposed to a caseby-case standard. The next section addresses this question.

5.5 Bright Line-Rule or Discretionary Standard? The distinction proposed above between cases pending before domestic courts at the time of the shift in case -law, and cases that were finally decided b y domestic courts in accordance with ECtHR case-law, has the hallmarks of a bright -line rule. Such a rule has the benefit of simplicity and predictability, but the ease with wh ich it could be applied comes with costs. A competing view could be that since the p olicy interests that favor retroactivity or non-retroactivity will be more or less present in different types of cases, a more discretionary or relativized standard would produce more equitable and tailored results.

The discretionary tests set out by the U.S. Supreme Court’s in Linkletter/Stovall and Chevron Oil show how a non-retroactivity standard could be operationalized and applied. The Linkletter/Stovall and Chevron tests, for federal habeas corpus and civil cases respectively, considered first the newness or unpredictability of the new rule, its nature and purpose, and the potential disruptions on the administration of justice of retroactive application. These factors, I would argue, capture substantial concerns against retroactivity in a given case, and are as such suited to solve the issue on a case-by-case basis.236 However, the inherent flexibility in such case- by-case tests undermines legal certainty,

equality, and invite unpredictable noise. As discussed above, t his critique about inequity was frequently made against the Linkletter/Stovall and Chevron Oil tests. Moreove r, a 236 See also Fallon and Meltzer, supra note 28.

49/50 discretionary standard presupposes that a court pronounces on whether new case -law applies retroactively or not in each case, which encourages litigation without providing further guidance t o authorities or individuals on how to solve cases that are n ot brought before the court . Thus, while the benefit of a discretionary standard would be that retroactivity can be tailored to the case at hand, a significant cost is precisely that it must be decided by courts case by case. Given that the ECtHR has the c apacity to decide approximately 2000 cases per year, 237 and receives between 5 0.000 and 65.000 yearly applications,238 I would argue against standards that require the Court to decide more cases.

5.6 Conclusion On balance, I conclude that the proposed rule of non- retroactivity better balances the relevant considerations in play, and better enables the shared responsibility between the ECtHR and the domestic courts in safeguarding rights and freedoms in the Convention, as compared to the disruptive approach to abs olute retroactivity which the ECtHR have pursued in many cases so far. This is not to say tha t a rule on non- retroactivity comes without trade-offs. As I spell out above, there are competing concerns for corrective justice, equality, and efficiency in play. Yet, these concerns do not, in my view, outweigh the overall benefits of non-retroactivity.

6 Final Remarks To summarize, the paper has discussed whether the ECtHR ought to adopt a doctrine on non-retroactivity to mitigate the disruptive effects of cha nges in case-law. It has analyzed the inconsistencies in the Strasbourg Court’s approach to the retroactivity problem, and contrasted it with a description of how SCOTUS has approached parts of the same problem.

237 Draft Copenhagen Declaration, supra note 19. 238 Statistics of the ECHR - Analysis of statistics, Thematical statistics, Facts and Figures, https://www.echr.coe.int/Pages/home.aspx?p=reports&c=#n1347956867932_pointer (last visited Apr 20, 2018). The Court received 53,500 applications in 2016, and 63,350 applications in 2017. It delivered 993 judgments in 2016, and 1,068 judgments in 2017.

50/50 Based on these descript ions, I have proposed a rule on non- retroactivity which would distinguish between cases pending before domestic courts at the time of the shift in ECtHR case-law, and cases that were finally and correctly decided at the time. Moreover, I have proposed that when the ECtHR depar ts from case- law, it should do so ex nunc , without granting retroactive remedies.

To test the soundness of my proposed non- retroactivity rule, I have approached it from different angles. First, I have considered whether it would hamper dynamic interpretations, and whether it ought to include pending cases before the ECtHR. Second, I have discussed how it bears out with respect to underlying considerations, such as justice, equality, legal certainty, inter- court interaction, and incentives. Finally, I have c onsidered the likely objection that a rule on non-retroactivity would be better framed as a discretionary standard as opposed to a bright line -rule. I have concluded that although the standards formulated by SCOTUS in Stovall and Chevron Oil capture substa ntial concerns underlying the

retroactivity problem, they are too flexible and invites unnecessary disputes. In the debate about the future of the ECtHR human rights machinery, I argue that a doctrine on non- retroactivity provides a functional middle ground between Contracting States weary of unpredictable and disruptive case-law from the ECtHR, and the Court’s “living instrument”-doctrine. Consistent with recent years’ reform declarations in Brighton239 and Copenhagen,240 it would defer the primary domestic implementation to domestic courts, and enable the ECtHR to act as a su bsidiary supervisor of the rights and freedoms in the Convention.

239 Brighton Declaration, High Level Conference on the Future of the European Court of Human Rights, 2012. 240 Copenhagen Declaration, supra note 13.