Proportionality during Times of Crisis: Precautionary Application of Proportionality Analysis in the Judicial Review of Emergency Measures

Abstract

The need for a standard of review even during times of crisis – proportionality analysis as the ideal and universally applicable standard of review – the problem of excessive stringency of proportionality analysis during an emergency – the principle of in dubio pro libertate as the cause of the problem – adjusting each component of proportionality analysis to the precautionary principle as the solution – the precautionary application of proportionality analysis strikes a more appropriate balance between excessive stringency and insufficient protection of fundamental rights

Introduction

Proportionality analysis has become an essential tool for balancing rights and public interests in many jurisdictions. ¹ In discussing limitations of fundamental rights, many authors highlight the rigorousness and rationality of proportionality analysis ² which makes the resulting decisions balanced (i.e. not biased in favour of individualistic or collectivistic views of the relationship between the individual and the government). ³ Proportionality has been described by its proponents as ‘the central concept in contemporary constitutional law’, ⁴ ‘the foundational element of global constitutionalism’ ⁵ or even ‘the ultimate rule of law’. ⁶

It was therefore not surprising that in a report on government restrictions imposed at the beginning of the Covid-19 pandemic on certain freedoms and other rights, the Venice Commission stressed that any restriction on fundamental rights must be guided by the principle of proportionality. More specifically, emergency measures cannot be excessive (disproportionate) in their stringency or the geographical area they cover. ⁷ It is therefore assumed that national and international courts would apply proportionality analysis not only under normal circumstances but also during a state of emergency.

The problem, however, is that despite the extensive literature on proportionality analysis for normal cases, deeper research into its application in times of crisis has yet to be conducted. ⁸ Building on the experience of the Covid-19 pandemic, which demonstrated that existing standards of human rights protection may come under severe pressure, ⁹ we argue that (a) considering the relevant arguments in favour and against, proportionality analysis is the best standard of constitutional review for times of crisis, provided that (b) each component in its standard structure ¹⁰ is adjusted to the conditions of the emergency.

In the first section, we first argue that a standard of judicial review is required even for states of emergency. In the second section, we discuss proportionality analysis as a universally applicable standard of review. In the third section, we address the problem of excessive stringency in proportionality analysis for times of crisis. In the fourth section, we explain why replacing proportionality analysis with significantly less stringent alternatives (e.g. rational basis test) is undesirable and creates more harm than benefit. Using several judgments issued by various courts in Europe as examples, we identify in the fifth section the main cause of the problem of excessive stringency in the underlying logic of proportionality analysis in normal situations. According to this logic, any uncertainty within the individual components of proportionality analysis should burden the government and its regulation rather than the fundamental rights it interferes with. We refer to this as the principle of in dubio pro libertate and contrast it to the logic based on the precautionary principle, which we argue is far better suited to emergencies. In the sixth section, we discuss the precautionary principle, according to which the uncertainty inherent to any crisis should favour rather than burden the government and its emergency regulations. Finally, in seventh section, we demonstrate with examples of judgments how the individual components of proportionality analysis should be adjusted to the precautionary principle. We also analyse in detail how precautionary application shifts the negative consequences of uncertainty from the government to opponents of the emergency regulation and thereby leads courts to not quashing or declaring an emergency measure or decision illegal unless it is (based on the already available data and evidence) manifestly unsuitable, manifestly unnecessary, or manifestly disproportionate stricto sensu.

The precautionary application of proportionality analysis proposed in this article represents a good balance between two opposing claims: the sufficient protection of fundamental rights during crises on the one hand, and the ability of authorities responsible for averting the crisis to take effective decisions on the other.

The need for a standard of review even for times of crisis

As the article focuses on proportionality analysis in times of crisis, it is first necessary to explain what we mean by ‘crisis’ (or to be more precise, an ‘emergency’) as opposed to ‘normality’. Our preliminary point is that we understand the concept of ‘crisis’ in a narrower sense (i.e. as an emergency situation characterised by an existing or impending disaster that originates outside the legal system, thus excluding, for example, constitutional or political crises from our scope). ¹¹ Since different legal systems define and regulate states of emergency in different ways, and some of them do not recognise states of emergency at all, ¹² we have decided to prefer the substantive definition of emergency to the formal one (i.e. it is not relevant for us whether or not a state of emergency has been formally declared in a particular country).

Thus, we define an emergency using two pivotal and interrelated criteria: (a) the scarcity of relevant information necessary for a proper understanding of the disaster and for predicting its consequences (i.e. epistemic uncertainty); and (b) the severity of the possible negative consequences of the disaster (i.e. seriousness). Since both criteria are continuous rather than dichotomous variables, both ‘emergency’ and ‘normality’ are only ideal approximations (i.e. ends of the scale between which real situations can be placed). ¹³ In fact, no real situation can be defined by absolute epistemic (un)certainty or by non-existent or absolute negative consequences. Consequently, we can only define an ‘emergency’ relatively: as a situation in which the degree of epistemic uncertainty and seriousness is manifestly higher than in normal situations.

Before we explain why proportionality analysis is preferable to alternative standards of judicial review during an emergency, we clarify why such a standard is required at all. It is worth highlighting that emergency measures (and other decisions adopted during an emergency) have two specific functions which are normally unnecessary. Besides the general function of any legislation or legal decision to regulate the behaviour of its addressees, emergency measures and decisions: (a) empower (but also limit) the public authorities in their effort to mitigate and avert a crisis; and (b) are an instrument for restoring normality. In other words, the purpose of these specific ‘emergency’ functions of law is to prevent an emergency from becoming the new dystopic normality in which less protection of fundamental rights and more concentration of public power in the hands of an executive are typical. ¹⁴

This is important for two reasons. First, the ‘emergency’ functions of law assist in determining whether emergency measures truly address an emergency. This is important in the question of whether a standard or precautionary proportionality analysis should be applied in a constitutional review. This is discussed in detail in the seventh section.

Second, these functions of law in emergencies clearly anchor the ideas discussed in this article in a legal (as opposed to extra-legal) approach ¹⁵ to an emergency. We argue that even during an emergency, we still operate within a constitutional and legal framework, not outside the law, as the principle of inter arma silent leges would suggest. We also disagree with the moderate form of the extra-legal approach, which is based on the delineation of areas in which substantial deference is recognised and in which courts not only do not review the discretion of legislature or (more typically) the executive (i.e. primary decision-makers) ¹⁶ but refuse to decide on such cases. ¹⁷ These areas of ‘legal vacuum’ (i.e. without any judicial control) could lead to erosion of the fundamental purpose of the separation of powers, which guarantees that power is not abused to the disadvantage of citizens and their rights. ¹⁸

Moreover, the extra-legal approach cannot be supported even at the level of international law. Although some international human rights treaties include so-called derogation clauses for emergencies, ¹⁹ the derogation of rights is generally limited not only to its scope (i.e. some of the rights cannot be derogated), but also to its intensity (i.e. ‘to the extent strictly required by the exigencies of the situation’ ²⁰ ). These limits logically presuppose that the scope and intensity (or perhaps even proportionality) ²¹ of the derogation would be reviewed by the international courts or other organs even during emergencies. ²²

If we reject extra-legal approaches which prevent (either wholly or in some delineated areas) the judicial review of emergency measures and decisions, the courts must have the power to review the constitutionality of these measures and decisions. It follows that a standard of judicial review for times of crisis is necessary.

The standard structure of proportionality analysis

Having clarified this initial question, several standards of review are now potentially applicable to emergencies. The second logical step in this argument is to explain why proportionality analysis should be preferred over any alternative standard of judicial review.

The structure of proportionality analysis that we use in this article consists of three consecutive components. ²³ The first component of proportionality analysis, i.e. the suitability test (sometimes also referred to as rational connection test), ²⁴ is usually understood as ‘[t]he requirement is that the means used by the limiting law can realize or advance the underlying purpose of that law; that the use of such means would rationally lead to the realization of the law’s purpose’. ²⁵ In some situations (e.g. as interpreted by the Canadian Supreme Court), the literature highlights that the standard is even stricter when it speaks of the means ‘designed to achieve’ the end sought. ²⁶

The second component is the necessity test (or minimum impairment test), ²⁷ according to which we should reject the reviewed regulation as unnecessary if the following conditions are met:

1) the existence of hypothetical alternative means that can advance the purpose of the limiting law as well as, or better than, the means used by the limiting law; 2) the hypothetical alternative means limit the constitutional right to a lesser extent than the means used by the limiting statute. The lesser extent is determined, among others, by examining the scope of the limitation, its effect, its duration, and the likelihood of its occurrence. ²⁸

The rationale of this component is that ‘[w]here a government measure infringes upon a right unnecessarily, then the right has not been optimized’. ²⁹ Hence, only the least intrusive measures can pass the necessity test in its strict form. However, it is also important to allow primary decision-makers a certain margin of discretion under which any regulation would be considered necessary. ³⁰

The third and final component, i.e. the proportionality stricto sensu (the balancing exercise) states that ‘in order to justify a limitation on a constitutional right, a proper relation (“proportional” in the narrow sense of the term) should exist between the benefits gained by fulfilling the purpose and the harm caused to the constitutional right from obtaining that purpose’. ³¹ It is possible that even the least intrusive (i.e. necessary) measure would still not be justified, because it would cause more harm (to constitutional rights) than benefit (in terms of pursuing the public interest). ³² In assessing this benefit or harm, the legal scholarship recognises several criteria, most commonly described by Alexy’s weight formula, which comprises: (a) the intensity of interference with a fundamental right and the degree of fulfilment of the conflicting rights or public interests; (b) the abstract weights of the rights and public interests at stake; (c) an assessment of the uncertainty regarding the epistemic preconditions for the application of the rule in question. ³³ The epistemic uncertainty ³⁴ may concern both empirical questions and normative implications. In other words, for being proportionate in a narrower sense, the benefits of the regulation under review must have a higher abstract value, greater intensity or higher probability than its cost in terms of interference with fundamental rights.

Even though proportionality analysis and especially the balancing in the third component is sometimes criticised for denying a special status to fundamental rights ³⁵ and reducing them to mere public interests, ³⁶ we argue that the standard structure of proportionality analysis should be understood as a universal standard of review and also be applied by the courts during an emergency. This is so because the rationale behind this structure, as interpreted by one of its most important proponents – Robert Alexy, ³⁷ understands fundamental rights as legal principles ³⁸ and thus guarantees their protection to the highest possible degree, even when it is reasonable (or even necessary) to limit them because of collision with public interests or other fundamental rights. Hence, we agree with Klatt and Meister, who argue that proportionality analysis allows a reflection of the special status of rights and their higher abstract weight compared to other public interests. ³⁹ We also agree with Barak, who notes that proportionality is a legal framework that must be filled with content, whereby ‘the component of proportionality stricto sensu may well incorporate the notions of “rights as trumps”’. ⁴⁰

Understanding fundamental rights as principles which should be protected to the fullest extent possible is even more important during an emergency, when the institutional guarantees of those rights, for example the separation of powers, and checks and balances, are often weakened in the name of crisis management and when fear or panic in society might blunt the vigilance of citizens.

The problem of excessive stringency

One strong argument against the application of proportionality analysis during an emergency is that it represents a too strict standard of review. ⁴¹ In times of crisis, this standard may prevent primary decision-makers from taking sufficiently effective decisions to mitigate or avert the crisis. In other words, if the courts applied proportionality analysis to emergency measures and decisions, they would be ‘trespassing on legislative territory’ ⁴² since ‘balancing between competing interests is the legislator’s job’. ⁴³

To support this argument, let us examine examples of judicial decisions in which the courts quashed or declared unconstitutional emergency measures as disproportionate. The first case concerned a ban on public demonstrations in Switzerland, which was adopted in the wake of the Covid-19 pandemic. The Communauté genevoise d’action syndicale lodged a complaint against the ban, and the case was brought before the European Court of Human Rights. ⁴⁴ By a narrow majority of 4 votes to 3, the Court found a violation of Article 11(1) of the European Convention (freedom of peaceful assembly). The first ground for finding a violation of the Convention, but not relevant to the analysis here, was the specific situation in Switzerland concerning the impossibility of effective (in this case, timely) judicial review of executive measures. ⁴⁵ However, the Court not only found a violation of Article 11 on procedural grounds (due to the absence of a national remedy) but also focused on the measure’s disproportionality. It therefore examined the nature of the restriction on the fundamental right on the merits. The nature of the prohibition, namely because it was a prohibition without exception, long-lasting and backed by a criminal sanction, and also because, for example, the physical presence of employees in offices and factories was not restricted in any way, whereas demonstrations outdoors were prohibited, and because it interfered with an important freedom guaranteed by the Convention, ⁴⁶ led the majority of the judges to the conclusion that Switzerland exceeded the ‘certain’ margin of appreciation that governments have in adopting decisions during an emergency. ⁴⁷

We leave aside a substantive assessment of whether the conditions set by the Court were actually met and focus only on the basic characteristics of the argument. The government’s margin of appreciation in the decision was influenced only insignificantly by the exceptional circumstances in spring 2020, when many doubts existed about the nature of the risks arising from Covid-19 and the means of combating the disease (i.e. by the epistemic uncertainty). ⁴⁸ Instead, the Court focused on the nature of the total ban in comparison with other activities that were not banned. Not surprisingly, this line of reasoning in proportionality analysis led to a finding of a violation of the European Convention.

Similar reasoning has also been used by other courts in reviewing anti-pandemic measures. In Germany, for example, in their judgment of 30 July 2020, the Supreme Administrative Court (Verwaltungsgerichtshof) of Baden-Württemberg in Mannheim ruled that the obligation to test slaughterhouse employees twice a week was contrary to the principle of proportionality. ⁴⁹ The main reason for this was the existence of possible alternative measures, namely the possibility of granting exemptions from testing in individual cases (e.g. an employee who does not come into frequent contact with other employees) or ordering testing only in justified cases (e.g. an employee returning from holiday). ⁵⁰

The final case we highlight was decided by the Bavarian Supreme Administrative Court (Bayerischer Verwaltungsgerichtshof) in Munich, which ruled that it was illegal to ban parties and barbecuing in public places. ⁵¹ The Court identified a problem with this regulation in that the ban concerned all parties and barbecuing regardless of the number of people present. In its judgment, the Court stated that the act of barbecuing itself was not problematic as a risk of spreading Covid-19. A large crowd, however, would be problematic, where individual participants do not observe social distancing and where it would be almost impossible to trace the participants if the disease were spread from this event. ⁵²

One could disagree with these judgments since they are genuinely strict and leave too narrow a space for the discretion of primary decision-makers and their attempts to mitigate and avert a crisis. In line with this argument, it might be appealing to favour the less stringent rational-basis test known from American constitutional law ⁵³ or its alternative in the form of the unreasonableness doctrine in British public law based on the Wednesbury case. ⁵⁴

The downsides of alternative standards of review

In this section, we argue and explain why alternative standards of review such as those introduced above cause more harm than benefit. The aforementioned examples clearly show that the stringency of proportionality analysis did not arise from its standard structure (i.e. by the existence of the three components described above), as the courts did not always need the third component (i.e. proportionality stricto sensu) to declare the reviewed emergency measures illegal. The true problem was that the courts did not reflect the specifics of the emergency, and consequently were too stringent when they interpreted and applied the individual components of proportionality analysis. Hence, changing the standard structure of proportionality analysis (i.e. replacing it with significantly less stringent standards of review, especially omitting the proportionality stricto sensu test) is not necessary.

The problem, however, with such less stringent standards of review is that they may not sufficiently protect certain rights under restrictive emergency measures. One can imagine such measures that may not be unreasonable to avert a crisis, but they may interfere fundamentally with individual rights. An illustrative example is the strict lockdowns applied in China until the end of 2022 under a dynamic Covid-zero strategy. ⁵⁵ These measures would probably pass a rational-basis test, but they interfere with fundamental rights in an extreme manner. ⁵⁶ This is because the less stringent standards of review typically exclude the third component of the standard structure of proportionality analysis (i.e. the proportionality stricto sensu) and are thus not capable of capturing even the most absurd (extremely disproportionate) infringements of fundamental rights when they can be interpreted as broadly reasonable means to somewhat legitimate governmental objectives (or public interests).

Finally, attempts to create and apply new review standards generally undermine citizens’ legal certainty. This can be counterproductive, especially during a crisis generally characterised by a high degree of overall uncertainty in society. Any new standard of review created by a court in deciding a particular case will raise two issues. First, uncertainty will arise in the disputing parties as to which of the already existing standards of review will be applied in their case, especially in borderline situations where the choice of an appropriate standard of review is not entirely clear. Second, uncertainty as to whether the court will create and apply yet another newly created standard of review, especially in cases which do not fall into one of the previously established categories of situations with already assigned appropriate standards of review.

To summarise this argument, we agree that applying proportionality analysis as a standard of review during an emergency might be too restrictive to primary decision-makers and consequently jeopardise the purpose of the emergency measures issued to mitigate and avert the crisis which necessitated the emergency in the first place. Unlike normal situations, crises often require timely and decisive action from primary decision-makers to be averted. Primary decision-makers have only a narrow window to gather the relevant data, run important analyses and balance competing interests before they are urged to issue a decision. It follows that the legislature and executive are often forced to adopt emergency measures or other decisions under conditions of uncertainty concerning both the risks posed by the existing or impending disaster and the consequences of the emergency measures they are adopting. Given a certain degree of uncertainty and seriousness (which varies between different emergencies and also within each emergency as it evolves), it is objectively more difficult for empirical premises to be based on unquestionable facts supported by scientific evidence. Consequently, an emergency always establishes a more or less modified context in which primary decision-makers should be afforded a correspondingly greater degree of discretion than under normal circumstances. ⁵⁷

However, such discretion (although broad) cannot be unlimited, even in these extraordinary circumstances. Klatt and Schmidt, for example, recall the famous judgment of the German Federal Constitutional Court, which found section 14(3) of the German Aviation Security Act unconstitutional. According to this law, a hijacked plane with passengers can be shot down by a public authority if the plane could be used as a weapon and if its destruction is the only effective way of preventing this threat. The Court quashed the relevant provisions of the law because it believed the act would give public authorities too broad discretion. Since no one may know for certain whether the crew or passengers of the plane will eventually regain control and therefore whether the plane will actually be used as a weapon, we are dealing with a high degree of epistemic uncertainty. The problem, however, was that the Act a priori (automatically) entrusted public authorities with unlimited and uncontrolled discretion to shoot down the aircraft; it thereby excessively interfered with the right to life and dignity of the people on board. ⁵⁸

In the context of an emergency, the key task then is finding an acceptable degree of public discretion in the face of epistemic uncertainty, i.e. how to strike an appropriate balance between the effectiveness of primary decision-makers and the protection of fundamental rights. Applying proportionality analysis as if a normal situation existed would limit discretion too extensively and thus prevent primary decision-makers from effectively averting the crisis. They would be allowed to act only when it is very likely that the measure or decision would lead to the intended objective (i.e. when the epistemic uncertainty is almost non-existent). In the aforementioned case, it would mean shooting down the hijacked aircraft just a few seconds before it struck the skyscrapers in the city centre. However, abandoning proportionality analysis and replacing it with any of the aforementioned alternatives would leave primary decision-makers with almost unlimited discretion and consequently jeopardise fundamental rights beyond an acceptable margin. Primary decision-makers would be allowed to act whenever a reasonable assumption exists that the measure or decision would lead to the intended objective (i.e. when the epistemic uncertainty is almost absolute). In our case, it would mean shooting down the aircraft as soon as reasonable suspicion exists that it has been hijacked.

Hence, a proper (balanced) solution to this issue rests in maintaining the standard structure of proportionality analysis but adjusting how the courts apply its individual components rather than altering or replacing it entirely.

The principle of in dubio pro libertate for normal circumstances

Before any adjustment in the application of proportionality analysis, it is necessary to show how it functions under normal circumstances. If we examine the aforementioned examples of judicial review of emergency measures, we can conclude that they were all driven by the same (and in normal circumstances perfectly legitimate) underlying logic: in a situation of uncertainty over whether the regulation under review is suitable, necessary or proportionate stricto sensu, the courts favour fundamental rights over governmental objectives (and interests to adopt such a regulation). Let us refer to this logic as the principle of in dubio pro libertate.

A typical way of expressing this principle in proportionality analysis is through the weight formula created and popularised by Robert Alexy. The formula includes: (a) the intensity of intervention on fundamental rights and the degree of fulfilment of conflicting rights and public interests; (b) their abstract weights; and (c) the degree of reliability of empirical and normative premises (i.e. epistemic uncertainty). ⁵⁹ According to Alexy, a strict relationship exists between the categories of intensity of intervention and epistemic uncertainty and leads to formulation of the second law of balancing. According to this rule, ‘the more heavily an interference with a constitutional right weighs, the greater must be the certainty of its underlying premises’. ⁶⁰

If the courts apply proportionality analysis under normal circumstances, then the primary decision-makers (i.e. the legislature and the executive) carry the burden of proof in each component of its standard structure. The primary decision-maker is thus required to argue and prove that: (a) the adopted law or decision is suitable for achieving a legitimate purpose; (b) no other equally effective and less restrictive alternative is available for achieving this purpose; and (c) the benefits in achieving this purpose outweigh the interference with fundamental rights by this law or decision. Uncertainty in any of these points should result in the primary decision-maker losing the case. In other words, the more intensively the courts apply the principle of in dubio pro liberate in their review, the less room there is for democratic or expert decision-making (discretion) by the primary decision-maker. ⁶¹

Indeed, it was the principle of in dubio pro libertate which led the courts in the aforementioned cases to their ultimate decisions. In the cases of public demonstration and barbecuing, the courts correctly found that while the restriction on rights under the reviewed anti-pandemic measures were certain, their benefits to public health were often only assumed according to the knowledge at that time. The courts on the one hand knew that both bans significantly interfered with fundamental rights, but on the other hand they were not (and at that time could not possibly be) certain of how these bans would serve the purpose of protecting public health against Covid-19.

Similarly, in the cases of barbecuing and testing employees, the courts knew with certainty that the measures under review were significantly more restrictive on fundamental rights than their hypothetical alternatives (i.e. banning only larger gatherings which could become overcrowded, and mandatory testing of employees under specific circumstances only, such as when the employee returns from holiday). However, the courts could not possibly be certain in how effective those less intrusive alternatives would be in protecting public health against Covid-19.

Finally, specifically in the case of public demonstrations, the court could not possibly be certain whether a ban on public demonstrations would be suitable for protecting the public health against Covid-19 at all, especially when people could freely meet (and thus infect) each other in indoor areas such as offices and factories.

This demonstrates that the principle of in dubio pro liberate, although necessary and desirable to adequately protect fundamental rights in normal cases, often leads to an overly strict application of proportionality analysis in emergency cases because the courts cite uncertainty (inevitable in such cases) as a strong argument against any emergency measures and decisions taken by primary decision-makers.

The precautionary principle for emergencies

We therefore suggest that proportionality analysis during times of crisis should be governed by a different underlying principle, namely the precautionary principle, which states the opposite: if it is uncertain whether the emergency measure or decision under review is suitable, necessary or proportionate in the narrower sense, the precautionary principle favours the governmental objective (to mitigate and avert the crisis) over fundamental rights.

The precautionary principle (Vorsorgeprinzip, i.e. ‘the principle of taking care before we act’) was first applied in German environmental protection policy in the 1970s, ⁶² and has been widely adopted in international environmental law ⁶³ since the Rio Declaration on Environment and Development, which states: ‘In order to protect the environment, the precautionary approach shall be widely applied by States according to their capabilities. Where there are threats of serious or irreversible damage, lack of full scientific certainty shall not be used as a reason for postponing cost-effective measures to prevent environmental degradation’. ⁶⁴ Although the precautionary approach ‘stands in opposition to the idea – previously widely accepted – that measures to address environmental threats ought only to be adopted where there exists firm scientific evidence establishing a likelihood of harm’, ⁶⁵ a consensus has been reached on its basic premise that it is desirable to take preventive action to avoid environmental damage. ⁶⁶

Today, the precautionary principle is applied in areas which extend beyond environmental protection, ⁶⁷ and it is understood as ‘a principle of public decision-making that requires decision-makers in cases where there are “threats” of environmental or health harm not to use lack of full scientific certainty as a reason for not taking measures to prevent such harm’. ⁶⁸ To analyse this principle in more detail, we can say that it should be applied especially in situations characterised by a sufficient risk of harm.

To determine whether the risk of harm is sufficiently high, the uncertainty created by the risk should be evaluated. The risk can be thought of as ‘a situation or event in which something of human value (including humans themselves) has been put at stake and where the outcomes are uncertain’. ⁶⁹ Uncertainty relates to several aspects of risk. Andreas Klinke et al. state that uncertainty includes the uncertainty of:

variability … of individual responses to an identical stimulus among the individual targets within a relevant population, measurement errors or uncertainties of modelling, indeterminacy resulting from a genuine stochastic relationship between cause and effect(s), lack of knowledge resulting from ignorance or … exclusion from external influences. ⁷⁰

Factors related to uncertainty include complexity, defined as the ‘difficulty of identifying and quantifying causal links between a multitude of potential candidates and specific adverse effects’, ⁷¹ and ambiguity, understood as a ‘variety of (legitimate) interpretations based on identical observations or data assessment’. ⁷² Another important factor is the seriousness of a risk, namely, the extent of impending harm to a protected interest (e.g. human health or the environment). ⁷³

It follows that the less we know about a given situation (i.e. uncertainty) and the greater the potential harm to the protected interests (i.e. seriousness), the more ground for applying the precautionary principle, which is the ‘rational and prudent response in a face of uncertainty’. ⁷⁴ Since the very essence of an emergency (as defined in the first section) is a higher degree of uncertainty and seriousness (objectively caused by the existent or impending disaster) compared to the normal situation, the reason to apply the precautionary principle in these situations is especially persuasive.

In line with our definition of emergency and normality as ideal ends of a scale rather than realistic situations, we emphasise that it would be a misperception to understand the in dubio pro libertate principle and the precautionary principle as mutually exclusive. On the contrary, in any real-life situation, both principles apply to a certain extent, depending on the uncertainty and seriousness of the situation. In other words, the higher the degree of uncertainty and seriousness (i.e. the further we move on the scale from normality to emergency), the more significant the role of the precautionary principle. Emergency can thus be operationalised as a situation in which the precautionary principle (by a certain margin given by the uncertainty and seriousness) prevails over the principle of in dubio pro libertate (i.e. the situation is placed closer to emergency than to normality on the normality/emergency scale).

To continue with our argument, we suggest that the standard structure of proportionality analysis should be applied even during times of crisis (i.e. in situations operationalised above), but each of the three components in this structure should be (to a degree depending on an intensity of a crisis) adjusted to the precautionary principle to better reflect the epistemic uncertainty given by that crisis. We refer to this form of proportionality analysis as precautionary application of proportionality analysis. We believe that this solution removes the objection that proportionality analysis is too strict and makes it impossible to respond adequately to a crisis characterised by a high degree of uncertainty. At the same time, it also preserves the structure of proportionality analysis and ensures that individual rights are not encroached beyond that which is necessary, and that rights and public interests are balanced.

This solution is preferable to the alternative proposed by Meßerschmidt, who suggested that ‘the precautionary principle has no precedence over the principle of proportionality. … Nevertheless, the precautionary principle may prevail against the proportionality test.’ ⁷⁵ This would mean that the proportionality principle and the precautionary principle would be applied separately and then balanced against each other. We believe that it is safer to transparently adjust the parameters of the individual components of proportionality analysis to emergencies in advance rather than to apply the proportionality analysis as if a normal situation existed and then challenge and modify its results because they may appear too strict. In other words, the precautionary principle can and should be integrated into the proportionality test itself, rather than being balanced against it (and potentially prevailing against it). We believe that any ex post modification of the results should be rejected since it could mean ‘watering down the principle of proportionality’ ⁷⁶ and thereby weakening the level of fundamental rights’ protection in emergencies.

Precautionary application of proportionality analysis

This section focuses on specifically how the precautionary application should adjust the use of individual components in proportionality analysis.

Suitability test

In a suitability test, the problem of applying this component of proportionality analysis during an emergency lies in the novelty of the crisis and the related unfeasibility of clearly determining the extent to which a particular measure will contribute to a legitimate aim. In normal situations, a consensus stems from the principle of in dubio pro liberate that marginal contribution alone is insufficient. ⁷⁷ However, in exceptional situations, which should be governed by the precautionary principle, the required degree of certainty on the suitability of a measure should be reduced. ⁷⁸ To pass the suitability test, the measure or decision under review no longer needs to be suitable. On the contrary, it is sufficient for it to not be manifestly unsuitable. Consequently, the negative consequences of uncertainty in the ability to contribute to the legitimate aim shift from the proponents of the measure or decision under review (i.e. a primary decision-maker) to its opponents.

Put differently, the courts should not quash or declare illegal any emergency measure or decision unless it is very likely (based on the already available data and evidence) that it cannot contribute to the legitimate aim in any way. It follows that with more reliable and abundant data (i.e. less epistemic uncertainty), the easier it should be to prove that a measure is manifestly unsuitable and thus the stringency of the suitability test increases. In the aforementioned public demonstrations case, if the court used the precautionary application of proportionality analysis, it would have to consider the ban of public demonstrations suitable unless it was very likely that it would not at all contribute to the protection of public health against Covid-19. This would have been extremely difficult at the beginning of the Covid-19 crisis since little to no information was available on how quickly the virus spread and what could be regarded a safe distance between people to prevent transmission of the disease. An extreme, manifestly unsuitable alternative to the analysed case would be, for example, banning public demonstration and protest in a virtual space, which enables people to participate from the safety of their homes.

After several months of gathering scientific data, however, proving little to no contribution of the ban to public health would become easier, and thus the uncertainty which initially favoured the primary decision-maker’s discretion would become a less significant factor. As the degree of uncertainty decreases over time, the measures which would have passed a suitability test (as not manifestly unsuitable) at the beginning of the crisis would be more likely quashed or declared illegal. Consequently, the varying degree of what is ‘manifestly unsuitable’ guarantees the highest possible protection of fundamental rights at any given moment, based on the current level of epistemic uncertainty and seriousness of a given emergency.

We would also like to emphasise that the (less stringent) precautionary application of proportionality analysis should be used only to review the measures and decisions designed to mitigate or avert a crisis. Other measures or decisions (even though adopted during an emergency) should be reviewed using the normal proportionality analysis governed by the principle of in dubio pro libertate. In this sense, the burden of proof whether the measure or decision under review genuinely targets crisis mitigation remains with the primary decision-makers.

Necessity test

In the necessity test, the restrictions on rights caused by the measures or decisions under review are required under normal circumstances to be narrowly tailored to a legitimate purpose. ⁷⁹ More specifically, it must be shown (in accordance with the principle of in dubio pro libertate) that an equally effective yet less restrictive alternative measure or decision is not available. The aforementioned cases of barbecuing and testing employees show that it is difficult to argue for the strict necessity of measures taken during times of emergency. There is too little knowledge and time (i.e. too much uncertainty) to conduct all the appropriate analyses which would prove that any hypothetical alternative measure or decision is either less effective or at least equally intrusive to fundamental rights as the measure under review.

Hence, we argue that, as with the suitability test, the courts should instead follow the precautionary principle and quash or declare illegal only those emergency measures or decisions which are manifestly unnecessary, thus leaving sufficient space for the discretion of primary decision-makers. ⁸⁰ Again, as a result of the precautionary application, the negative consequences of uncertainty shift from the primary decision-makers to the opponents of the measure or decision under review. Hence, in the cases of barbecuing and testing employees, the courts should consider the measures under review necessary unless it was very likely (based on the already available data and evidence) that less intrusive hypothetical alternatives, (e.g. permitting public barbecues but requiring a safe distance of three meters between people; mandatory testing of employees in specific situations only) would truly be at least equally effective in protecting the public health against Covid-19. As with the suitability test, it can be assumed that more reliable data on Covid-19 would be gained over time (e.g. that the Covid-19 virus is unlikely to be transmitted at distances of more than two meters between people), reducing the level of uncertainty and thus the discretion of primary decision-makers, and the stringency of proportionality analysis would thus gradually increase. Consequently, the measures would have had a much higher chance of passing the necessity test (as not manifestly unnecessary) at the beginning of the Covid-19 pandemic than after several months or years had passed, and thus, because of the varying degree of what is ‘manifestly unnecessary’, the fundamental rights would have been protected to the maximum extent possible at any given moment, based on the current level of epistemic uncertainty and seriousness of a given emergency.

The consequence of the precautionary application of the first two components of proportionality analysis (i.e. suitability and necessity tests) is an important change from normal proportionality analysis. The first two components of proportionality analysis are often referred to as threshold criteria. This means that from the court’s perspective no balancing is done at this stage (i.e. the relative importance of the legitimate purpose being pursued and of the fundamental rights being restricted does not play a role), but rather a search for a threshold where the regulation under review passes the test. ⁸¹ However, due to the changes (typically decrease) in overall epistemic uncertainty over time, both thresholds become fluid, as they have to be determined in each individual case by balancing the degree of uncertainty, the seriousness of the crisis, and intensity of interference with fundamental rights.

Proportionality stricto sensu test

The final component in proportionality analysis, which is usually understood as its core, is the proportionality stricto sensu test, which attempts to balance conflicting rights and public interests. ⁸² The context of an emergency affects only some of the basic criteria included in Alexy’s weight formula. ⁸³

The intensity of interference of the measures and decisions under review with fundamental rights and the reliability of the empirical and normative premises concerning this interference are generally unaffected by emergency circumstances. Using the cases of public demonstration and public barbecuing as examples, the level of certainty in the intensity of the measures’ interferences with the freedom of peaceful assembly and the freedom of movement and personal autonomy remains the same regardless of whether we ask these questions under normal or emergency situations.

The abstract weights of fundamental rights and public interests also remain unchanged. This is because the characteristic feature is abstractness, which requires looking beyond the specific context of each individual case. However, it can be expected that rights and interests jeopardised during an emergency (e.g. right to life, public health, or national security) are among the central values protected by constitutions and international treaties, and therefore their abstract weights will often be particularly high.

It follows that the context of an emergency will significantly affect the reliability of the empirical and normative premises concerning the degree of fulfilment of rights and public interests which the emergency measures and decisions under review seek to protect. Again, using the aforementioned cases of public demonstration and barbecuing as examples, it is very difficult (or almost impossible) to assess how much and how likely bans on public demonstration or barbecuing will fulfil the fundamental right to life or the interest in protecting public health, especially with the limited data available at the beginning the pandemic.

The issue of epistemic reliability (uncertainty) is specifically discussed by Klatt and Schmidt, who suggest that among the various options for assessing the intensity of interference with a fundamental right, the option that is potentially the most intense, i.e. the most pessimistic, ⁸⁴ should be selected as the reference criterion. Consequently, in contrast to an optimistic option, the potentially most intrusive interference with rights is considered, thereby protecting the essence of those rights against the most intrusive restrictions. However, this approach contains a problem, which the authors correctly point out: ‘the most pessimistic classification involves the risk of incapacity [of the primary decision-makers] to act’. ⁸⁵ Therefore, according to the authors, it is necessary to combine an assessment of the most pessimistic cases with the most probable ones. ⁸⁶ Thus, it is not enough to consider the most pessimistic option if it would also be extremely unlikely.

When applying this rule to the entire balancing exercise (i.e. to both sides of the equation), we can say that the conflicting right or public interest that we seek to fulfil also has its most pessimistic variant (i.e. that the regulation in question was not adopted and did not protect the aforementioned rights and interests at all) and at the same time the most probable variant of harm to the conflicting right or public interest which could be caused by not adopting the concerned regulation. However, this rule cannot be successfully applied during an emergency because, as explained above, the epistemic uncertainty caused by the crisis is too high for the court to even identify the most probable pessimistic variant not to mention its application as the reference criterion in the balancing exercise. In other words, it would be extremely difficult to determine the seriousness and likelihood of the danger to public interest which collides with fundamental rights by not adopting the emergency measure. This is why the two criteria of ‘emergency’ are intertwined: the less we know about the nature of the existing or impending disaster (i.e. the greater the uncertainty), the more severe the consequences we should expect if no emergency measures are adopted (i.e. the greater the perceived seriousness). Similarly, determining the benefit of the measure in terms of protecting fundamental rights which collide with public interest would also be challenging. ⁸⁷

Hence, the only means of avoiding the vast majority of (if not all) emergency measures being quashed or declared illegal due to the fact that their benefits are too uncertain, we suggest the precautionary application of the proportionality analysis also for the proportionality stricto sensu test. As in the two aforementioned tests, precautionary application would shift the negative consequences of uncertainty from the primary decision-makers to the opponents of the emergency measure or decision under review. An emergency measure or regulation should therefore pass the test unless it is manifestly disproportionate rather than proportionate. Hence, in the same manner as the two previous tests, the courts should consider an emergency measure or decision proportionate unless it is very likely (based on the already available data and evidence) that the harm it causes to the fundamental rights it interferes with outweighs its benefits in terms of the protection of colliding fundamental rights or public interests. The stringency of the test would once again gradually increase as more reliable data become available to the courts over time.

Examining the cases of public demonstration and barbecuing for the last time, both measures would most probably pass the precautionary application of proportionality analysis as not manifestly disproportionate stricto sensu at the beginning of the pandemic, since the courts would have limited information about the contagiousness and mortality rate of the Covid-19 disease. Therefore, it could not be proven that the harm to fundamental rights affected by the ban outweighs the benefits of the measures under review (perhaps the virus spreads rapidly at these events and consequently kills thousands of people). However, after several months (i.e. after people receive vaccination or people infected with the disease become more immune), the courts could exclude this pessimistic option as extremely unlikely and reach the opposite conclusion of the balancing exercise and quash or declare the emergency measures in question illegal, thus again using the varying degree of what is ‘manifestly disproportionate stricto sensu’ to protect fundamental rights to the maximum extent possible at any given moment based on the current level of epistemic uncertainty and seriousness of a given emergency.

Conclusion

In this article, we argue that proportionality analysis should be applied even during times of crisis, but provided that each component of its standard structure is adjusted to the precautionary principle. This proposed solution strikes a more appropriate balance between two undesirable extremes.

The first extreme is to review emergency measures and decisions as if under a normal situation. This leads to excessively stringent review and the vast majority of emergency regulations being quashed or declared illegal since their (certain) negative effects generally outweigh their (uncertain) benefits. Consequently, the primary decision-makers would be left incapable of averting the crisis.

The opposite extreme is to abandon the application of proportionality analysis and replace it with a lower standard of review which generally excludes the proportionality stricto sensu test from its structure. As a result, the vast majority of emergency measures or decisions would pass the review since they generally have a rational connection with the aim of mitigating or averting the crisis. This would give the main decision-makers almost unlimited discretionary powers and deprive citizens of adequate protection of their fundamental rights.

The aim of this article was to theoretically describe and explain the general logic behind the precautionary application of proportionality analysis which makes our findings broadly applicable. The proposed adjustment of the three components of proportionality analysis allows adaptation of the stringency of proportionality analysis to the variations in crisis severity and fluctuating degree of epistemic uncertainty. The resulting adjusted proportionality analysis can be safely applied in judicial review under any type or stage of emergency. A downside of this approach, however, is that we did not include specifics related to individual countries and their legal systems into our consideration. Future research should therefore focus on how states of emergency or crisis management are regulated in individual countries and how precisely the general logic described in this article should be applied to those countries.

Another limitation of this article is that we have not addressed the procedural aspects of proportionality analysis. Given the different consequences of epistemic uncertainty caused by application of the precautionary principle, it is possible that the burden of proof for the individual components of proportionality analysis shifts compared to normal circumstances. Future research should therefore examine which party to a dispute needs to argue and prove what and to what extent (i.e. the burden of proof) within each component of proportionality analysis in order to be successful before a court.