Should the penal norm be non-retroactive in an unfavorable sense? The yes, the affirmative answer, is so obvious in an order of democratic legitimacy that the question is even surprising because of its doubtful nature. It is a “yes, of course”: the unfavorable non-retroactivity of criminal law is a postulate that is as closely linked to legal certainty and as essential to the principle of legality as it is to the democratic rule of law.
Things become more complicated when we ask the same question with respect to jurisprudential changes in pejus. It is not the criminal statement that changes, but the way in which the courts understand it, and this results in there being a penalty where there was none, or in the penalty being higher than the one that had been imposed. There are many examples here, and some are intuitively more sympathetic or unfriendly to the thesis that such a thing can be done by the courts, and this can cloud our path towards the right answer, towards systematic and principled thinking.
The first one I propose is that of soft drugs coming to be considered hard. Amphetamine was initially classified as a drug that does not cause serious damage to health, but this changed jurisprudentially in the 1990s, when it was classified as a drug that does have such effects. With the current regulation, the penalty for trafficking would go from one to three years of imprisonment to three to six years of imprisonment (art. 368). Is the person sentenced to the more serious penalty right (constitutional reason) in that he had a legitimate expectation that the crime he committed was lighter, or is it rather the judge who understands that the new interpretation is a more reasonable reading of the criminal sentence and in that sense as and even more foreseeable than the first?
We will understand the judge better in the well-known example of marital rape in the United Kingdom, which gave rise to the ECHR C. R. v. the United Kingdom, of 22 November 1995, which denied European protection to the appellant who claimed that until his conviction it had been understood that such sexual assault was unpunished. As in the first, in the third example the jurisprudential change has a great influence on the imposable penalty: at one point the Supreme Court understood that the single action that intentionally causes several deaths should not be treated as an ideal concurrence of crimes but as an actual concurrence; if they were homicides of the basic type of Article 138 (if they were murders, the specific concursal rule of Article 140. 2 PC that leads to revisable permanent imprisonment), the penalty would go from twelve and a half to fifteen years (higher penalty in its upper half) to a penalty of twenty years (sum of the penalties with the limit provided for in Article 76.1.a PC). The last example sparked a lively controversy in Germany in the 1990s, when the Supreme Court decided to lower the blood alcohol level reflecting an absolute inability to drive (from 1.3 gr/liter to 1.1 gr/liter). And it did so retroactively.
What is the question?
As so often in Criminal Law, the questions are actually two, here well intertwined. The first is dogmatic, de lege lata: does our legal system prevent unfavorable retroactive jurisprudence? The second is political, of criminal policy and constitutional policy: should our system prevent unfavorable retroactive jurisprudence?
The questions are so interrelated because there is no express constitutional declaration in this regard, as there is for normative statements, for “non-favorable punitive provisions” (art. 9.3 CE), for “legislation” that establishes “crimes, misdemeanors or administrative infractions” (art. 25.1 CE). It will be necessary to interpret such an abstract text as the Constitution, and in this interpretation of what it is, what it should be in the light of the constitutional values and principles will have a great deal of weight.
The yardstick
The political question is: what is better for our system; that unfavorable jurisprudence be retroactive or not? What does “better” mean in this context?
It is better if we gain more in terms of constitutional values and goods. Criminal rules – and rules governing the temporal validity of rules that describe crimes and provide for penalties are criminal rules – should be evaluated against that bar in two senses. They are good if they protect constitutional goods. Rules that criminalize sexual assault are good if they protect sexual freedom. Let us agree to call this first way of measuring “efficacy”. The second is that the State plays with fire (with prison) when it creates, interprets and applies criminal rules, so that all these activities must be subject to certain basic principles if we do not want the cure to be worse than the disease. Let us call this “efficiency”. In their essential contents, principles act as trump cards: such contents cannot be undermined for efficiency reasons (unless essential contents of principles collide, which is equivalent to two trump cards on the game mat).
The thesis of pejorative retroactivity
The classic strategy in our legal system is that of unfavorable jurisprudential retroactivity. Classic in that it belongs to the continental legal tradition; classic in that it responds to what our courts were doing before the STEDH Del Río Prada, with the endorsement of the Constitutional Court, including for this case of the so-called Parot doctrine, which I will refer to later.
It is important to highlight the assumptions of this model.
- The first is that the same statement admits at least two reasonable readings and that what happens is that for the prosecution of a given case, the one that is worse for the accused than the one that had been used up to that moment is chosen as the most appropriate. If this second unfavorable interpretation for the convicted person is not reasonable (semantically, axiologically and methodologically reasonable, according to the jurisprudence of the Constitutional Court: STC 137/1997) there is no problem or, better, the problem is solved by another route: the decision must be annulled for violation of another of the manifestations of the principle of legality, the binding of the judge to the criminal law (or mandate of typicality). What happens here is that the judge does not follow the statement, but invents it.
- The second presupposition is the minimum respect for the mandate of taxativity of the criminal statement. Its possibilities of interpretation are limited and knowable, under penalty of constitutional nullity prior to that of an alleged unconstitutionality due to unfavorable retroactivity. The four basic manifestations of the principle of legality must be analyzed in their logical order: law, precision, judicial binding, unfavorable non-retroactivity. We ask ourselves whether the postulate of non-retroactive application was respected if a sufficiently restrictive law was applied.
A) It could be thought that doing things this way, applying to the case the new unfavorable interpretation that the court considers better, is more correct in the first sense of fairer in that it is more effective. And this is not trivial, since Criminal Law is dedicated to the protection of the most important individual and social goods. If we tolerate unfavorable jurisprudential retroactivity, we will be able to punish marital rape, as opposed to the previous unjust impunity; or punish with the appropriate severity (as real and not ideal concurrence of crimes) those who throw a grenade and kill several people; or severely punish the sale of a new drug whose effects on health must be considered serious, according to the new expert assessment.
B) The downside of this way of administering criminal justice is legal certainty. The convicted person will invoke that he trusted that the interpretation of the precept would be different and that this trust was reasonably supported by the fact that this was the way in which the judges had been acting. “I think it is very good that now they consider that conducts such as mine are punishable or more punishable, but I did not know it and could not foresee it. Apply this way of reading the criminal precept from now on”.
Note that this is a principled, efficiency objection. Unfavorable jurisprudential retroactivity would be a form of unfavorable penal retroactivity contrary to the principle of legality and to the fundamental value of security that this principle seeks to preserve.
The reply from the ranks of this thesis is the following. The loss of security is as notorious as it is not decisive, and it also has a final remedy.
a) For the first, for the relativization of the damage to security, it is said that this legal security is guaranteed by our system with the requirement that the interpretations of the criminal statements be foreseeable insofar as they are reasonable. The new harsher reading of the precept was a possibility that the citizen had to count on; in fact, it is the one that is now affirmed as the most reasonable. This is sufficient to preserve his safety, according to the European Court of Human Rights. What Article 7 requires is that the addressees of the law should be able to know that they run a real risk of incrimination. This is what is known in the Anglo-Saxon world as the thin ice doctrine: it is a matter of warning skaters that the ice lake may break.
Moreover: as García Amado has highlighted (“Sobre el principio de legalidad penal y su alcance de la previsibilidad como componente de la legalidad”, 2016), if the rule that unfavorable jurisprudence can be retroactive governs, the citizen can know that his conduct can be analyzed and subsumed from unfavorable interpretations alternative to the one the courts are making.
b) In addition, the possible remaining insecurity has a lenitive: if the defendant acted in the mistaken belief that he did not commit a crime, his error, whether of prohibition (he believed, for example, that marital rape was not a crime) or of type (he believed that the substance with which he trafficked did not cause serious damage to health), will make his guilt less, or non-existent, or for a more minor conduct.
The thesis of pejorative non-retroactivity
Is this way of doing things better, which as I said one line above is for us the traditional one, or should we opt for the alternative, as suggested by several important authors (Winfried Hassemer, Víctor Ferreres, Guillermo Oliver) and especially as at least to some extent suggested by the European Court of Human Rights in the Del Río Prada v. Spain judgment of October 21, 2013? Is it better that only conduct carried out after the corresponding interpretative change has taken place can be punished in peius and certainly not in the case on the table of the court and which is the one that initiates the interpretative change (it could not be punished as rape, or as trafficking in hard drugs, or as actual concurrence of crimes, in the examples that support us in this exposition)?
Note the new balance. The decisions for the case are not the best, the most effective, and neither are they nor all those that judge conducts carried out prior to the change in jurisprudence. In exchange, the legal-criminal security of citizens remains completely untouched, who would be assured not only that they will not be punished if the criminal law in force does not say so according to any of its possible reasonable interpretations, but also that they will not be punished if their conduct is not subsumable in the law as the courts have been interpreting it.
Furthermore, in defense of this thesis, it is possible to minimize the prevention deficit insofar as the sentence proclaims precisely urbi et orbe the harshest penalty or punishment “from now on”. Until today it was not known that marital rape was punishable, but it is known that it is punishable from today, even if the sentence in which this is stated is only declaratory and in practice acquittal.
There is, however, an important additional problem, which is both principled and systematic. Deeply rooted in the democratic invention of the division of powers, in our system the judges are “subject only to the rule of law” (art. 117. 1 CE), which means, following now the Organic Law of the Judiciary: “they are independent with respect to all judicial and governing bodies of the Judiciary”; that judges and courts are forbidden “to correct the application or interpretation of the legal system made by their inferiors in the judicial hierarchy, except when they administer justice by virtue of the remedies established by law”; and that neither “may judges and courts […] issue instructions, of a general or particular nature, addressed to their inferiors, on the application or interpretation of the legal system that they carry out in the exercise of their jurisdictional function” (art. 12).
The judge is guided by what is fully proven to have happened and by the law, and in his understanding of the law he is not bound in doing justice to the case by his own precedents or those of other judicial bodies.
Against the thesis of non-retroactivity it has also been argued that it makes the penal system harsher and ultimately more insecure. Harsher, again according to García Amado, because judges may prefer harsher interpretations that do not tie their hands for the resolution of future cases. The harsher interpretations will always be exceptable for the specific case through more benevolent readings for the accused (there would be no main problem in suddenly classifying ecstasy as a soft drug) while those more favorable to the defendant would be unalterable for the specific case (ecstasy could not be qualified as a hard drug for the case in question).
The system of non-retroactivity may paradoxically make the criminal law more insecure to the extent that it serves as a network for indeterminate criminal types. The legislator may tend to vagueness and vagueness if he knows that the insecurity it generates will be remedied by jurisprudence; on the other hand, the thesis of retroactivity will make the lack of taxativity unbearable, and therefore unfeasible.
What is to be done?
It is not easy to decide in favor of one or the other way of doing things. Hence the dilemma in the title. And hence I could have written this entry in the form of Rayuela, so that, depending on the reading order, two alternative conclusions can be reached, to the reader’s liking. In essence, having ruled out that there is a relevant deficit of prevention in the thesis of non-retroactivity, the thesis of unfavorable retroactivity is the most consistent with judicial independence and the thesis of non-retroactivity, with legal certainty.
With the scale that I specified before to measure which is better – which is more “constitutional”, if I may say so – I am inclined towards the second thesis: that of non-retroactivity.
The thesis of unfavorable retroactivity would be a perfect thesis if its presupposition (taxativity of the normative enunciation) and its final corrector (the error as a mitigating or exonerating factor) worked. But neither one nor the other is the case. The presupposition is rigor with the taxativity mandate: that the normative statement has only a limited and knowable number of readings. But this is not always so; rectius: it is not infrequently not so. I now recall STS 484/2015, of 7 September, on cannabis clubs, in which the Plenary painfully lamented that “we must now examine” a criminal type “of unclear contours and profiles, almost unbridled” (FD 4). And I emphasize above all the understandably deferential doctrine of the Constitutional Court in relation to respect for the criminal legislator’s mandate of determination. We allow the criminal legislator an unbearable imprecision precisely because the case law makes it bearable, because the concreteness of the statement is reasonably feasible by virtue of the jurisprudential context (SSTC 68/1989, 89/1993). Therefore, as Víctor Ferreres has lucidly explained (“El principio de taxatividad en materia penal y el valor normativo de la jurisprudencia”, 2002), there is no doubt that the unfavorable non-retroactivity of jurisprudence serves such a fundamental value as legal certainty and is in line with the tolerance of constitutional jurisprudence for vague criminal types precisely by virtue of its jurisprudential clarification. And therefore he is absolutely right when he points out the following: “when faced with the complaint that the law is vague, the Constitutional Court refers to the jurisprudence. Therefore, when faced with the complaint that an unfavorable change in case law has been applied retroactively, can the citizen be told that only the law is relevant? Where does that leave us?”.
The damage to legal certainty would not only come from such open statements, but also from the confidence that a very stable and long-lasting interpretation of the criminal type reasonably arouses. This is the Parot case. In Hassemer’s words, “situations are observed in which the legal community has a greater knowledge of the content of criminal jurisprudence than of the content of the criminal law and also trusts the status of the jurisprudence” (“Fundamentals of Criminal Law”, 1984).
And for these cases of serious damage to legal certainty, recourse to error does not always come to the rescue. Unfortunately. This would be valid for the case of marital rape. The relief would be excessive if, in the case of drugs, it leads to impunity: the crime actually committed (hard drugs) was committed with an error of type. Perhaps this consequence could be saved by way of interpretation, as an error on an aggravating circumstance, but in any case there is no remedy by way of error to the insecurity suffered in the case of Parot or in the case of ideal competition that becomes real.
In this situation I believe it is preferable to shield legal certainty with respect to the penalty, with respect to freedom, because it belongs to the heart of the constitutional guarantees. Judicial independence cannot be traced back to the content of a fundamental right. Or certainly not if such independence serves to increase the penalty. And it would not be so much violated as qualified by the submission of the judge to the legal system, to the law understood and applied in a constitutional system.
Difficulties and limits
The retroactive application of the law is clear because its moment of entry into force is clear. There is absolutely no parallel clarity in relation to a deviation from an interpretative line that generated expectations of continuity due to its stability. What is this stable jurisprudence that changes? In systems such as ours, where there is a higher body for the unification of doctrine, it would seem at first glance that only its case law, and not that of lower courts, should be taken into account. The question is when it can be said that the way of interpreting exists as stable, something that will only be clear if it is expressly stated by that judicial body, as our Criminal Chamber does with its Plenary Decisions (which normally have the vocation of stabilizing a line that until then was disputed) for the future. In other cases, it is worth asking what constitutes the stability that must be ensured: how much time and repetition it requires and how much dissent it is capable of withstanding.
What is meant by the above is that the non-retroactivity of unfavorable jurisprudence is in fact an exception to a system in which the judge, in his independence, can decide for the case the legal reading he deems most just, even if it is not the most favorable to the defendant. And it is an exception for reasons of legal certainty: only when that decision, although reasonable, is unforeseeable due to the uncontroversial soundness of previous interpretations and applications. And therefore, with Strasbourg: yes, there are unforeseeable but reasonable interpretations of criminal sentences.
Does the answer depend on the role of the judge?
There is no link between the thesis of retroactivity and the conception of the judge’s work as “cognitive-decisional”, as a finder of the true meaning of a norm already in force. Nor is there a link between the thesis of non-retroactivity and the understanding of the judge as a kind of successive legislator, who attributes meaning “to a significantly inert formula”, so that, in the words of Francisco Laporta, in reality “the rule in the strict sense is the product of interpretation, it is born with it” (“The Parot doctrine and the theory of Law”).
Honestly said, although the opposite would suit my position of non-retroactivity: the conception of the judge as a successive regulator does not lead to unfavorable non-retroactivity. Given that he would be a second normative agent and that he is in any case severely limited by the first, reasons of efficiency and judicial independence could justify limiting the mandate of unfavorable non-retroactivity to the legislator. In fact, this addressee is the only one expressly mentioned in the Constitution.
De lege lata: is unfavorable retroactive jurisprudence prohibited today?
In the Del Río Prada case, on the Parot doctrine (which understood that the reduction of sentence for work should be computed on the initial sentence and not, as had been interpreted, on the final sentence, thirty years in the most serious cases), the European Court does, and in fact imposes, what I have argued here that should be done: to prohibit in exceptional cases unfavorable retroactive jurisprudence for elementary reasons of legal certainty. And it understands that the foreseeability of the interpretation and application of the rule depends not only on the reasonableness of the interpretation itself, but also on its historical consolidation. Thus, it should be emphasized that there could be reasonable but unforeseeable interpretations and therefore inapplicable to the case.
In relation to the new pejorative interpretation that constituted the so-called “Parot doctrine”, the Constitutional Court categorically understood that the changes in jurisprudential criteria and their conformity with the Constitution could not be analyzed from the perspective of the violation of the right to criminal legality, but from the perspective “of the principle of equality in the application of the law (art. 14 CE)”, of the “voluntary application of the law”, of the “voluntary application of the law” and of the “voluntary application of the law”. 14 CE)”, of “selective voluntarism” or of “unmotivated departure from the consolidated applied criterion maintained until then by the jurisdictional body whose resolutions are challenged, which constitutes the essence of the inequality of application according to our jurisprudence” (STC 40/2012, FJ 6).
Then came Strasbourg (STEDH of October 21, 2013) and understood that what is transcendent for the analysis of respect for the principle of legality is “whether the new interpretation was reasonably foreseeable for the plaintiff”, and that for this it was necessary to take into account the way in which the rule “had been applied”, which in the present case was otherwise more favorable to her consistently by the prison or judicial authorities over many years” (§ 115) and with “nothing to indicate the existence of a discernible trend in the evolution of case law” in the direction of the new interpretation (§ 117).
In conclusion, is unfavorable criminal case law irretroactive in our law today? Yes, if the new interpretation in peius is unforeseeable, which will not be affirmed because of its unreasonableness (since this would annul the interpretation for another reason: because of the lack of binding the judge to the law), but because of its own novelty and the full stability of the most favorable interpretation: because of the lack “of a discernible trend in this regard in the case law”, as stated with Strasbourg.