|
|
Home
| Databases
| WorldLII
| Search
| Feedback
New Zealand Journal of Environmental Law |
Last Updated: 21 January 2023
173
Letting our Standards Slip?
Precaution and the Standard of Proof
under the Resource Management Act 1991
Greg Severinsen*
Precaution has become a key concept in the field of environmental law. Essentially, taking a precautionary approach means that any uncertainty in our scientific understanding of environmental effects should not be used as an excuse to postpone or avoid taking action to prevent harm. The idea has taken firm root in New Zealand’s legal thinking only in the last two or three decades, and has found expression in a handful of environmental statutes. However, New Zealand’s primary environmental statute, the Resource Management Act 1991, does not refer to precaution in its provisions concerning decision-making on resource consent applications. In the absence of an express statutory reference, the courts have been prepared to read in precaution in a variety of ways in the consenting context. This has been done largely on a case-by-case basis, and has resulted in a range of inconsistent judicial approaches. One of the ways in which precaution has been recognised in the consenting context is in the standard of proof under the Act, particularly when a consent authority is determining the potential effects of a proposed activity. Some cases have required an applicant for consent to disprove the existence of potential effects to a standard greater than the ordinary civil standard, while others have abandoned the concept of proof altogether in favour of the expert predictive judgment of the Court. This article contends that the concept of proof retains a vital role in the fact-finding process, including in the prediction of potential effects. It also questions the need to abandon the ordinary civil standard of proof, because the precautionary flavour of the Act
*BA/LLB (Hons), PhD candidate, Victoria University of Wellington. Email contact: greg. severinsen@gmail.com.
can be provided for better in substantive policy documents rather than through inflexible procedural or evidential rules. The standard of proof need only be enabling, not a guarantor, of precaution. The article concludes that the underlying reason for the complex and confusing dicta that have arisen on the standard of proof is the conflation of two concepts, proof and probability, and that a more coherent approach is possible and desirable.
1. INTRODUCTION
The concept of precaution is one that has found significant traction in the modern era of environmental law. At its simplest, it can be described as the need to be careful where we are uncertain about future effects on the environment — we should not be risking the health of our environment simply because we are not sure if adverse effects will or will not occur — or that “some action is better than inaction”.1 It has been appositely stated that “the core of the principle is that if there is scientific uncertainty, it is ‘better to be safe than sorry’ or ‘take care’ before proceeding without restraint”.2 Uncertainty or a lack of proof of causation should not be used as an excuse to postpone taking action to prevent damage.3
Developed originally in European and international law, and particularly prominent in the fisheries context, environmental precaution has found its way into a number of domestic legal instruments in New Zealand. It now has specific expression in fisheries legislation,4 hazardous substances and genetically modified organisms legislation,5 and most recently in the framework legislation for environmental regulation in the exclusive economic zone.6
The concept of precaution under the Resource Management Act 1991 (RMA) has, as one would expect, received a significant degree of attention. Although the consenting provisions of the Act itself do not contain a specific expression of the need for precaution, courts and commentators have identified and read in a precautionary flavour to the Act in a variety of ways in the consenting process. Thus the presence or absence of a precautionary approach in the RMA has not itself been seen as a serious problem. The real issue is that the courts have diverged markedly in their views on how and when precaution is to be achieved. Academic comment has been equally inconsistent. A number of useful expositions of the case law have been produced, with a particularly helpful and recent summary by Professor Alexander Gillespie. But most commentators have been forced to conclude that, although some broad principles can be observed as emerging from the case law, there remains no one settled approach to precaution in consenting decisions under the RMA.This article takes a narrower yet more optimistic view, in addressing specifically how precaution plays a role in the standard of proof used in resource consent proceedings. It argues that there is in fact a common thread that can be discerned from the case law. Although the legal tests relevant to precaution and the standard of proof have often been described with a degree of ambiguity, the article contends that a unified approach exists that can address the underlying concerns in the judgments and rationalise what the courts have done in practice. This article is part of a wider body of work looking at a range of precautionary mechanisms under the RMA, such as the burden of proof and the parameters around discretionary judgments, and seeks to place the standard of proof in this broader precautionary context. If its contentions appear to fail to safeguard precautionary outcomes, it is not because it sees precaution as being of little value; it is rather because it sees this role as being fulfilled by a legal mechanism other than the standard of proof. Ultimately, the article contends that the standard of proof implied in the RMA is not itself inherently flawed, ambiguous, or unworkable, but rather that it needs to be clarified. More work needs to be done through the subordinate planning and policy development process, especially at the national level, to strengthen precautionary outcomes. It is worth noting that precaution can be achieved at several stages of the planning process (for example, in developing policies, rules, national environmental standards and activity statuses, and these are dealt with briefly).7 However, this article’s focus is on how precaution has been and should be treated in the
context of applications for resource consent, and specifically in the standard of proof to be applied when establishing facts.
2. THE CONCEPT OF PRECAUTION
An analysis of the precautionary approach under the RMA is particularly vulnerable to the dangers of assumption. Commentators may, quite legitimately, have a specific view of what precaution means or requires — particularly because a number of alternative international and domestic conceptions of the idea exist.8 This problem is compounded because the Act itself does not expressly outline what precaution is, or the circumstances in which it can or should be used. Analyses of statutory provisions have therefore tended to be guided by factors that are neither statutory nor consistent, with the accompanying danger that these factors are based on unstated assumptions. Thus some may see precaution as being necessary only where there is a risk of serious or irreversible damage, or where there is a fundamental lack of scientific understanding of general environmental processes.9 Although this article does not take a moral view on what conception of precaution may be “better”, it is vital to outline the assumptions on which the discussion will proceed.
Firstly, in its most general and useful conception, precaution can be seen simply as an approach to risk management that favours environmental protection in cases of factual uncertainty.10 A lack of an express principle in a statute does not automatically translate to a lack of precautionary outcomes; legislation may implement strongly precautionary “approaches” in its operation without referencing the term “precaution” itself.11
The second assumption on which this article proceeds is that precaution is not itself a guide to sound environmental decisionmaking. The concept cannot be a replacement for “sustainable management”, because it can apply only in cases of uncertainty. If the effects of a proposal on the environment are significantly adverse, yet there is no real doubt as to their extent or occur-
rence, precaution can play little or no role. Thus Gillespie reminds us that in relation to any given activity “it is vital to realise that the precautionary principle is temporary”.12 One must be careful not to overreach the usefulness of the concept. If a substantive outcome is perceived to be inadequate for environmental protection, this may be simply because the purpose of the Act has not been applied properly, or that this purpose itself is flawed, rather than any failure of precaution. Yet even in cases of uncertainty, it is important to remember that precaution cannot override an express statutory purpose.The third key assumption that needs to be clarified relates to the kinds of uncertainty with which precaution is concerned. The concept does not apply to any kind of uncertainty. For example, a person cannot point to uncertainty in government policy or the ambiguous language of a legal provision to justify the need for a precautionary outcome. The upshot of this is that precaution applies only to scientific or factual matters,13 and therefore applies only to things that are amenable to the concept, and a standard, of proof (such as the features of an activity, the state of the existing environment, and predictions of effects). Such uncertainties do not include ambiguities in the wording of planning provisions, or in the RMA itself.
One final assumption that is important in the context of domestic legislation is that the international debates over precaution are only of limited relevance. While Gillespie has correctly pointed out that under international law “... a State Party may still choose to accept the risk and proceed as planned, for such a choice is ultimately a political, and not an environmental, one”,14 this is certainly not the case with decisionmaking under the RMA. A risk is only acceptable if it amounts to sustainable management under the purpose of the Act.
3. A HISTORY OF CONFUSION
Following amendments in 2003 and 2013, planning authorities are directed to implement a relatively weak version of the precautionary principle when preparing policy and planning documents, under s 32(2)(c) of the Act. They must assess the risk of acting or not acting if there is uncertain or insufficient information about the subject matter of those provisions. However, a precautionary principle does not find specific expression in the consenting provisions of the RMA, nor does the Act expressly provide for a precautionary
(or other) standard of proof. The RMA was passed in 1991, whereas the principle did not come to the fore in New Zealand’s environmental thinking until later in the 1990s.15 But arguments over the role of precaution as an “approach” in the RMA are possible at a number of stages of the consenting process. Some have seen precaution as demanding an overall persuasive burden of proof on an applicant, or flexibility in the civil standard of proof according to the impact of an alleged effect. Precaution has also been recognised as playing a role in guiding the overall discretionary judgment of a consent authority, and in the imposition of consent conditions.The way in which precaution influences these steps, notably in the context of the standard of proof when establishing potential effects, has been the subject of conflicting judgments in the courts. As Nolan points out, “it is difficult to discern a common approach to the matter”.16 Others have noted, concerning the use of precaution in the context of proving facts, that “the concepts employed [have] a multiplicity of meanings, or in some cases [are] misused”.17
It is worthwhile taking a moment at this juncture to consider the nature of the consent decisionmaking process under the RMA, to understand the points at which precaution can play a role. This process has been carefully constructed, with each step fulfilling a particular function. For simplicity, in this article the process is treated as one involving a notified and contested application for a discretionary activity, and the consent authority is treated as being the Environment Court (the Court).18
After an applicant has lodged an application and a hearing has commenced, the Court is generally presented with evidence from witnesses called by the parties. Some expert (opinion) evidence may concern factual matters such as the state of the existing environment, the effects of a proposal, or the features of an activity. Other expert evidence, particularly planningbased evidence, may concern nonfactual matters such as the interpretation of objectives and policies. Although the Act does not expressly require allegations of fact to be
proved, it is incumbent in the Court’s role as finder of fact that it makes a determination on factual matters. If a matter requires proof, then there must also be a degree to which, and generally a party by whom, this needs to be proved. These are usually referred to, respectively, as a standard and burden of proof.If an allegation of fact relates to an environmental effect, the Act requires it to amount to either an “actual” or a “potential” effect.19 If such effects are established, the Court is obliged to consider and give discretionary weight20 to them in a broad, overall judgment under s 104B as to whether the proposal fulfils the purpose of the Act (the promotion of sustainable management).21 A number of other matters are required to be considered and given weight in this judgment at the discretion of the Court, not all of which are factual matters. The Act also provides for any “other relevant matter” to be considered in this judgment, under s 104(1)(c). Finally, if the purpose of the Act is met, then the proposal is granted consent (subject to conditions if required). This simplified description of the consent decision-making process reflects the two basic steps of fact-finding and judgment outlined in Ngati Kahu Ki Whangaroa Co-operative Society Ltd v Northland Regional Council: “the Court makes a judgment on such an application after finding facts ...”.22
While this article is concerned with the role of the standard of proof in achieving precaution under the Act (by providing a threshold for the establishment of “effects” which the Court must then have regard to in its overall judgment), it is worth briefly placing this concept within the wider precautionary mechanisms of the Act. Once effects have been established, the Court must have regard to and weigh not only these, but also to a variety of policy documents. A number of these are underpinned by precautionary aspirations. For example, the New Zealand Coastal Policy Statement provides
19 Section 104(1)(a).
an express (yet relatively nonspecific) exhortation to decisionmakers to adopt a precautionary approach where effects on the coastal environment are uncertain.23 While not expressly “precautionary”, riskavoidance measures can also be perceived in the water allocation provisions within the National Policy Statement for Freshwater Management 2014.24 Furthermore, a number of rules and standards, which trigger the need for consent and are thus of relevance before the fact-finding exercise commences, can be seen to be premised on a precautionary approach to risk identification and management. For example, the classification of activities in national environmental standards on contaminated land depends in part on the results of a risk assessment (the greater the risk to human health, the stronger the restriction).25 Also, although it is arguable whether it provides adequate precaution, national standards on the generation of radio frequency fields now address uncertainty by classifying as permitted those telecommunications facilities that do not exceed the maximum exposure levels in specific New Zealand standards.26 Furthermore, standards relating to potable water provide for a degree of precaution by deeming certain activities to be prohibited if likely to make water unsafe to drink.27Finally, it is possible to see a precautionary approach underpinning the Act’s requirement for applicants to provide an assessment of environmental effects (AEE) with a consent application. These provisions do not impose a specific standard of proof to which alleged effects need to be proved or disproved by the parties. Nevertheless, they represent part of a relatively robust approach to risk management by requiring an applicant to present information that is sufficiently detailed to enable informed decisions to be made on the facts.28 In some cases, express consideration must be given by an applicant to risks, alternatives, and safeguards.29 This is consistent with the language of precaution, and is linked to the Court’s powers to decline an application if insufficient information is provided by an applicant.30
29 Schedule 4, cls 6(1)(a), 6(1)(c), 6(1)(d)(ii), 6(1)(e), 7(1)(f ).
30 Section 88(3).
4. STANDARD OF PROOF UNDER THE RMA
Although formal admissibility rules are virtually nonexistent under the RMA,31 and the Court has a power to “receive anything in evidence that it considers appropriate to receive”,32 this should not be confused with its responsibility to determine facts. The Court cannot rely on its broad ability to hear evidence to avoid deciding on the truth of a factual matter, or simply to direct that all admissible evidence be considered under the overall broad judgment in s 104B. Intrinsic to the Court’s role as fact-finder is to determine a single set of facts from what may be equally admissible yet conflicting allegations. This is an exercise in proof, rather than the narrower concepts of reliability or admissibility.
The correct standard of proof is often not characterised as an issue of precaution, because it is a wider concept than precaution. For example, determining a fact relating only to the present, or ascertaining what has already occurred, has no conceptual relevance to precaution, because one can only be cautious about the outcomes of future events.33 Thus it is now settled that “disputed facts in issue in proceedings under the RMA have to be proved on the balance of probabilities” where a fact is limited to the past or present.34 However, the correct standard of proof accompanying the Court’s fact-finding role where allegations concern future events is an issue that has significant implications for precaution, and continues to be debated.
4.1 Precaution and the Standard of Proof under the RMA: Two General Approaches
Professor Gillespie has identified four contenders for the appropriate standard of proof when establishing future effects in any given precautionary approach. These are (from weak to strong): a “scientific suspicion”, “reasonable grounds for concern”, “balance of probabilities”, and “beyond reasonable doubt”.35 The latter two formulations will be familiar as the ordinary civil and criminal standards of proof. The conventional civil standard is well understood, as described by Denning J in Miller v Minister of Pensions:36
31 Shirley Primary School v Christchurch City Council [1998] NZEnvC 394; [1999] NZRMA 66 (EnvC) at [141]. 32 Section 276(1)(a).
It must carry a reasonable degree of probability, but not so high as is required in a criminal case. If the evidence is such that the tribunal can say: “We think it more probable than not,” the burden is discharged, but, if the probabilities are equal, it is not.
However, a substantial body of case law has seen this civil standard as inadequate in the context of predicting future effects under the RMA. The Court in Long Bay-Okura Great Park Soc Inc v North Shore City Council described the problem as being concerned with “not just ... probabilities of more than
0.50 — the common law standard of proof — but ... to probabilities of less than0.50 (which, traditionally, common law Courts normally disregard in relation to ‘facts’”.37 In other words, entirely disregarding lowprobability effects cannot achieve precaution. There have been two broad and divergent approaches by the courts to this problem. Some decisions have considered that precaution can only be adequately implemented under the RMA by altering the normal civil standard to make proof more difficult for an applicant, in cases where there is uncertainty as to whether an effect will occur. In other words, an applicant should be required to prove to a degree greater than the balance of probabilities that an alleged effect will not occur. In promoting this approach, such cases have been careful to stress that this does not result in a criminal standard of proof being required.38 This tradition was established most cogently by the Planning Tribunal in McIntyre v Christchurch City Council.39 The Planning Tribunal there considered that proof was necessary to establish potential effects, and that “... the evidence must satisfy us of the fact (ie that there will or will not be such an effect) on the balance of probabilities and having regard to the gravity of the question”.40 This approach can be labelled as one involving a sliding scale of proof, where the standard depends on the potential impact of an effect.
A number of subsequent cases have generally endorsed this approach.41 However, other cases have taken divergent views or used different phraseology. Thus while in Telecom v Christchurch City Council the Environment Court
40 At 105–106.
41 It is worth noting here that although some of the cases identified in this article (such as McIntyre) pre-date the 2003 amendments to s 32 of the RMA (which first imposed an express requirement for planning authorities to consider risk where information is uncertain), the cases were decided in the consenting context. Thus their comments concerning the correct approach to the standard of proof in that context remain unaffected by the amendments. Furthermore, s 32 requires only an express consideration of the risk of
concluded that “it is clear beyond any doubt that in evaluating debates about scientific matters such as this the Court can only be guided by the view it forms of the preponderance of the scientific evidence”,42 the Environment Court in Shirley Primary School v Christchurch City Council departed from this view:43
... there is no one standard of proof, if that phrase is of any use under the Act. The Court must simply evaluate all the matters to be taken into account under s 104 on the evidence before it in a rational way, based on the evidence and its experience; and giving its reasons for exercising its judgment the way it does.
Here, the Court abandoned the concept of “proof ” altogether when looking at the establishment of future effects, seeing a discretionary “evaluation” or “judgment” as a more useful tool. This approach was supported in Director General of Conservation v Marlborough District Council, where the Court held that assessing future effects is:44
... not simply a matter of fact but of judgement. As we pointed out in the Interim Report, simple assertions as to the standard of proof break down under the complexity of the probabilities involved.
Such a patchwork of approaches has prevented any coherent legal test from emerging. Commentators have been reduced to generalisations such as “patterns suggest that precautionary measures are possible if they are based upon reasonable evidence”,45 that a threat needs to be “based on sound science”, or “reasonable grounds for concern” are required.46 Generalisations are necessary to describe or reconcile divergent dicta, but the consequence is that the precision of any legal test has been lost.
The idea described in McIntyre that a civil standard of proof can slide according to the gravity or potential seriousness of consequences has been a persistent
acting or not acting, and as such does not alter the correct standard of proof under the Act for determining the existence of an activity’s specific effects.
46 At 372.
one.47 Even Shirley, while seeming to reject a sliding standard of proof in favour of a discretionary evaluation, recognised that a balance of probabilities standard is inadequate because “the Court is required to evaluate beyond the balance of probabilities (ie 5050) where the risk (even if low) is of high potential impact”.48The solution of a sliding scale is an intuitively attractive one, because it stamps a precautionary flavour on an Act that is at least partly concerned with the protection and preservation of the environment. However, a sliding or flexible standard also presents significant difficulties. First, it is arguably a contradiction in terms to claim that a decisionmaker can apply both a balance of probabilities standard while at the same time having regard to the gravity of the question. If the seriousness of a question in practice demands a higher standard of proof, this standard cannot realistically continue to be characterised as one on the balance of probabilities. Second, there are uncertainties over the degree to which the standard of proof can change, the extent to which gravity affects the standard, or whether these questions can be determined in an objective or consistent way. Third, unless a criminal standard of proof is contemplated, a sliding scale is still incapable of ensuring that potential effects of extremely low probability are taken into account.49 In light of these formidable challenges, it is not surprising that the Environment Court in Shirley rejected the very concept of “proving” future effects, preferring an evaluative approach. Nolan and Somerville have seen this counterapproach in Shirley as a helpful and sensible interpretation,50 while the Court in Long Bay considered that there is “no such thing as a standard of proof for future events”.51
However, it is suggested here that the real issue with McIntyre may not be with the appropriateness of proof as a concept, as concluded in Shirley, but rather with the identification of the subject of proof in the sliding scale approach. Essentially, a misinterpretation of the subject of proof has led the courts in cases like McIntyre to see a sliding scale as necessary to achieve precaution, while the same error has led to the drastic abandonment of proof in the Shirley tradition.
In essence, the courts have been compelled to recognise that the RMA implements a precautionary approach, because it requires potential effects of low probability to be considered by a consent authority when exercising its overall broad judgment. This is clearly a requirement of s 104(1)(a) of the Act. Thus a balance of probabilities standard has been seen as inconsistent
51 Long Bay, above n 37, at [321].
with the wording of the Act, because it would mean that only effects that are “more likely than not” would become a mandatory consideration for the Court. A similar assumption was made also by the House of Lords in the different context of Davies v Taylor, where it was stated that the “balance of probabilities... can be restated as the burden of showing odds of at least 51 to 49 that such andsuch has taken place or will do so”.52 By introducing in the RMA a standard of proof dependent on the gravity of the question, lowprobability effects can be considered as long as this is justified by their potential impact. For example, the Court may be only 5 per cent convinced that an effect will occur, but a 5 per cent standard of proof may be seen as sufficient to establish this effect if its potential impact is sufficiently grave.
This logic proceeds on the significant assumption that the thing needing to be proven is that an effect will or will not occur, and that this is then deemed to be proven if a certain standard of proof is met.53 For example, in Long Bay, the Court held that when confronted with an effect of 16.67 per cent probability, that effect would be disregarded if a balance of probabilities standard were imposed.54 If this assumption is correct, then the ordinary civil standard of proof would certainly fail to implement the precautionary approach of the Act.
However, the assumption is not beyond challenge. Three logical components of the fact-finding process can be discerned in a legal context. Firstly, there needs to be an alleged fact. Secondly, that fact needs to be proved by a party (or at least accepted by the Court as proved). Thirdly, the fact must be proved to the requisite standard. The second and third components relate to the burden and standard of proof respectively, and have been the subject of extensive judicial comment. However, the significance of the first component has been largely overlooked, even though it has been implicitly recognised from the earliest judgments on the issue.55 The subject of proof has a significant impact on the extent to which fact-finding enables precautionary outcomes.
4.3 Determining the Subject of Proof: Potential Effects
In its role as a finder of fact, the Court is obliged under s 104(1)(a) of the Act to determine the existence of “actual” and “potential” effects. These effects, once proved, are then weighed in the overall broad judgment under s 104B. The definitions of these terms are therefore crucial in determining the nature of the matters that are required to be proved to the Act’s standard of proof.
Neither an “actual effect” nor a “potential effect” is itself defined in the Act, although the definition of “effect” in s 3 includes specific examples of potential
effects: any potential effect of high probability and any potential effect of low probability which has a high potential impact.The examples of potential effects given in s 3 clearly contemplate that a potential effect is one that is less than certain to occur, and one that can have a range of potential impacts. But it is less clear whether the definition is exhaustive. To qualify as “potential”, does an effect have to be either of low probability and high impact on one hand, or of high probability on the other?
A number of cases have assumed that these categories are the only valid ones to establish a potential effect. Thus the Court of Appeal in Dye v Auckland Regional Council stated that:56
[a] high probability of occurrence is enough to qualify the potential effect as an effect, whereas a potential effect which has a low probability of occurrence qualifies as an effect only if its occurrence would have a high potential impact.
The Environment Court in Baker Boys v Christchurch City Council also noted that the Act has “an interesting omission from the definition of other potential effects, including, for example potential effects with a 50% probability of occurring and average impact”.57 However, the Court in that case went on to recognise that the concept of an “effect” is inclusive and the definition in s 3 does not in practice preclude the validity of other categories of effect.58
The issue has been the subject of obiter comment from the Supreme Court, where Richardson J favoured an exhaustive definition:59
It is not surprising, then, that to achieve balance, the framers of the legislation restricted consideration of potential effects by the qualifiers imposed by s 3. ... [These] limit consideration to “(e) Any potential effect of high probability; and (f ) Any potential effect of low probability which has a high potential impact”.
It is respectfully suggested that the Supreme Court erred in reaching this conclusion. It is significant that the Court itself expressed its obiter comments with a degree of tentativeness.60 A potential effect is not itself defined in the Act, and the use of the term “includes” in s 3 signifies a non-exhaustive definition of “effect”, including within the category of a potential effect. The two examples of potential effect given in the section are more usefully seen as a signal that both potential impact and probability of occurrence are important components
59 Westfield (New Zealand) Ltd v North Shore City Council [2005] NZSC 17, [2005] 2 NZLR
597 at [183].
60 At [184].
of a potential effect, rather than imposing two exhaustive categories into which a potential effect must fit.If this were not the case, the omission of effects of medium probability and impact noted in Baker Boys would be concerning. It could undermine the protective focus of the Act. Furthermore, it would be verging on the absurd for the Court to have to consider effects of low probability and high potential impact, while not having to consider effects of medium probability and high potential impact. This is reason enough to conclude that the definition of an effect (and potential effect) in s 3 is not exhaustive. The overall broad judgment required under s 104B is also deliberately holistic and inclusive in nature.61 Section 3 should therefore not be seen as constraining the definition of a potential effect as used in s 104. The most persuasive conception of a potential effect is one that encompasses effects of any probability and any potential impact.
4.4 Actual Effects vs Potential Effects
Having addressed the nature of a potential effect, it is left to determine the nature of an “actual” effect, before turning to the standard to which these effects need to be proved. These are the two kinds of effect that must be considered by the Court under s 104(1)(a) if they are proven, thus they are the subjects of proof to which a standard needs to be applied. Parliament seems to have intended that the terms have different meanings; if they did not, then it would have been unnecessary to refer to them separately in s 104(1)(a). The Environment Court in Maniototo v Central Otago District Council found the distinction between actual and potential as “rather puzzling because in a sense all the effects of a proposed activity are potential rather than ‘actual’ at the time of consideration by the consent authority”.62 Similarly, in Waring v Tasman District Council the Court concluded that “all future effects are potential effects, because they have not occurred yet”.63 Thus the difference cannot realistically be one of present and future. The only plausible distinction is one of probability. This was recognised in Baker Boys where the Court concluded that unless an effect is inevitable, it is to be classed as a potential effect.64 In Maniototo, an actual effect was considered to be one that is “likely”.
There are obvious difficulties with the idea that a future effect can be inevitable. No such certainty can ever exist in reality. A threshold of being “likely” is also unhelpful, because it lacks precision. However, although it is significant that Parliament provided separate terms to describe effects of different probability, the point at which a potential effect becomes an actual effect is ultimately irrelevant. This is because there is no distinction in how they are treated under s 104(1)(a); the requirement is simply that regard be had to both. It is more useful to conceive of the two not as separate concepts, but rather as useful labels for the ends of a single spectrum of probability, and as a reminder to decisionmakers to consider all effects within the bounds of this spectrum.The significance of this approach for the standard of proof is that it clearly identifies the subject that is required to be proved. In providing for a spectrum of probability, the Act recognises that a party does not have to prove that an effect “will” or “will not” occur. Such certainty in future effects is conceptually impossible,65 thus Parliament could not have intended that a party should have to prove it. Therefore it is suggested that any above zero per cent probability of occurrence, together with an above zero potential impact, form the elements required for an alleged effect to amount to one recognised by the Act. By requiring the Court to consider effects of any probability, the Act is enabling of a precautionary approach.
5. PROBABILITY AND PROOF UNDER THE RMA
5.1 Conceptual Differences
It is worth pausing at this point to reflect on the fact that, in the context of effects, the subject of proof (an above zero scientific probability) and the degree to which it needs to be proved (a legal standard of proof ), are emerging in this discussion as separate concepts. Dr Somerville has presented a distinction between scientific and legal proof, in that the former is predictive while the latter is generally postdictive.66 Similarly, the Environment Court in Long Bay pointed out that the finding of fact by a court is not a scientific exercise.67 This perceived limitation of the law has resulted in the use of the language of risk prediction or evaluation rather than the language of proof or fact-finding.68 It
has also meant that the focus has been on how risk is established (guided partly by an evaluation of policy considerations) rather than how potential effects are established.69However, it is suggested here that the role of legal proof may be wider than simply establishing events that have already occurred. Simply because proof at law is generally postdictive, it is not necessarily constrained in this way. While the purpose of probability is to determine, scientifically, how likely a future effect is to occur, the purpose of legal proof can be seen as the pursuit of truth
— to assess the degree to which any allegation of fact is correct.70 Thus while a probability of occurrence of over 50 per cent means that an effect is more likely to occur than not, this can only be true if it is accepted that this allegation of probability is proved to the required standard.
An assertion of the probability of an effect occurring will usually be accompanied by expert evidence. An expert witness might claim that an effect will or will not occur, but it is just as likely that she or he will provide a probability or range of probabilities of occurrence either in quantitative (“40 per cent”) or qualitative (“unlikely”) language. Irrespective of the probability that is alleged, the Court retains a vital role in determining the degree to which the evidence alleging it is persuasive. In other words, the Court needs to determine the truth of an alleged probability. Unlike an assessment of probability, this is not a scientific or empirical process reliant on expert evidence.71 It is a fundamental role of the Court as finder of fact, and one that can encompass a wide range of considerations, including the relative reliability, persuasiveness, confidence and qualifications of a witness.72 This reflects the fact that proof at law is not solely concerned with empirical observation and cannot be separated from the judgments of human beings, and that there is a disconnect between the basic functions of law and science.73
A formidable hurdle to this reasoning appears to be that future effects are, conceptually, incapable of proof. Something that has not yet occurred cannot be shown to be “true” or “untrue”, because it is not yet a “fact”. Only present and past events can reasonably be treated as facts and thus amenable to proof in the present. As the Environment Court pointed out in Director General of Conservation v Marlborough District Council:74
The ... hypothesis is not easily described as an assertion of fact since it is about future possible effects. We remind ourselves too that effects are defined by the RMA to include “effects of low probability and high potential impact”. That means that such questions are not necessarily “proved” on the balance of probabilities ...
Fundamentally, proof and facts are concepts that have utility only in the present. For example, even though one may think it relatively simple to show that the sun will rise in the east tomorrow — and this is, of course, very likely
However, this does not lead automatically to the conclusion reached in Shirley, Long Bay and other cases, and also implicitly in Cross on Evidence,75 that the effects of a proposal under the RMA are not facts and can be assessed only through a subjective evaluation rather than proof. While a potential effect concerns future events, it is not itself a future event. A potential effect is comprised of two basic components: an assessment of probability of occurrence, and an assessment of probability of impact.76 Any assessment of probability is an exercise that is bound to the present, not to the future. Once an event occurs, it ceases to have potentiality and crystallises into certainty
The corollary of an approach in which future effects are seen as being not amenable to proof is that past and present facts are not conceptually amenable to probability. As the Environment Court pointed out in Long Bay, probabilities are not inherent qualities of events.78 A present fact or a past event is either true or it is not, because its probability has already crystallised when it occurred.79
Any uncertainty that arises is not due to probability of occurrence in a predictive sense, but rather due to the degree to which a decisionmaker is convinced of the truth of an assertion. Thus one cannot claim that there is a 50 per cent probability that an apple is red; one can only say that she or he is 50 per cent satisfied that this is so. An assertion that someone is 50 per cent satisfied that there is a 60 per cent chance of the sky being blue is contradictory, because both are expressions of proof.In the context of potential effects, although probability and proof both address uncertainty, they perform separate functions. A standard of proof is designed to assess the value and persuasiveness of assertions, evidence and witnesses, not to alter a probability described by an expert within that assertion or evidence. Probabilities must necessarily be expressed through the assertions of witnesses, yet proof remains essential to determine whether that assertion is or is not persuasive. In contexts where scientific or factual predictions of effects are not as pronounced (as in most civil proceedings concerned with the resolution of disputes arising in the past), there is little objection to using the language of probability to describe proof, because it can be useful shorthand for the degree to which one is convinced. It is only in scenarios like the consenting process, where proof and probability need to be considered separately and against different yardsticks, that it becomes useful to draw a distinction both in practice and semantically.
5.2 The Approach of the Courts to Proof and Probability
Under the RMA the courts have often, albeit in most cases unknowingly, proceeded on the basis that proof and probability are separate concepts. The distinction is, in fact, one that is largely intuitive in the fact-finding process. Thus in Baker Boys the Court noted that “[w]e find that it is likely that up to 85% of the supermarket traffic will use the central (eastern) access/egress point
...”.80
Here, the Court clearly recognised that it had to be satisfied (“we find that it is likely”) of an alleged probability (85 per cent), rather than simply being satisfied (to any standard) of the existence of a future effect. Similarly in Shirley, the Environment Court stated that “... the word ‘potential’ shows that it is not proven actual effects that need to be considered but also scientifically possible effects established to our satisfaction ...”.81 Here, the concepts of probability (scientific possibility) and proof (to the satisfaction of the Court)
are distinguished. Similar practical distinctions between probability and proof can be found in a number of other cases.82However, despite the implicit judicial acceptance of a distinction between proof and probability, the conclusions drawn by the courts have tended not to recognise a difference explicitly. This can be seen in cases reflecting both the Shirley and the McIntyre approach. Thus although the Court in Director General of Conservation v Marlborough District Council distinguished between the concepts of evidential and belief probabilities, it required that it be satisfied that an effect “will” occur, rather than being satisfied of a chance of occurrence.83
In Shirley, the Court assumed that proof fulfils the same function as probability. It thus saw a balance of probabilities standard as being inappropriate, because the Act requires potential effects of less than 50 per cent probability to be taken into account.84 It was held, citing the Privy Council decision in Fernandez v Government of Singapore, that there is nothing in the law that requires a risk to be ignored if it is less than 50 per cent likely to occur.85 The Court in Shirley considered an example of a highimpact effect, which would be unacceptable to ignore:86
... if your doctor says you have cancer and a 17% (1 in 6) chance of dying within the year you might consider the chance of dying is high ...
The concept of probability (contained in the 17 per cent chance) was here remedied directly by a change to the standard of proof. However, this conclusion disregards a vital ingredient of the example given — that the probability is still one asserted by an expert. If the doctor in question possessed, for example, a doctorate in English literature from the University of Antarctica, or was clearly biased, one might be disinclined to believe her or him. The probability asserted would likely not be proved to even the most lenient of standards, even though the probability itself would be considered to be relatively high when looked at in isolation from its source. One American judge has gone so far as to point to “the myth of the objective scientist”.87 As in Shirley, the line of cases following McIntyre has also assumed that the subject of proof is the existence of a future effect rather than any above zero probability of its occurrence.88
although in the context of planning provisions rather than the consenting context, also required proof that an effect “will” occur: Wratten, above n 7, at 173.
This assumption is also evident in the more general civil context of Davies v Taylor. The House of Lords there recognised that the law in some contexts can and does impose lower standards of proof than the balance of probabilities, and drew a distinction between the mathematical nature of probability and the more impressionistic nature of witness credibility.89 However, the likelihood of a future event was treated only as one aspect of the standard of proof, rather than an assertion of fact that itself needed to be subjected to proof. Furthermore, while the decision in Westfield (New Zealand) Ltd v Hamilton City Council concerned the standard of proof in relation to planning provisions rather than a resource consent application, it highlights the general reluctance of the New Zealand courts to recognise a conceptual difference between proof and probability in formulating legal tests for the standard of proof. It was held that:90
... it is a contradiction in terms to say that the Court was required to determine “on the balance of probabilities” whether provisions “could” give rise to consequential effects. The possibility that something “could” happen is clearly a lower threshold than the probability that it will occur. The tests are mutually exclusive.
Here, the concept of proof is seen as an expression of scientific probability. It is true that a court cannot impose a test that has two different probabilistic thresholds for the same alleged effect. In light of this, the decision does not appear to appreciate that the term “will” is itself a threshold of probability (100 per cent) that is equally as inconsistent with a 51 per cent threshold as the term “could” (above zero per cent). Just as one cannot at the same time prove that there is both an above zero and an above 50 per cent chance of occurrence, one cannot at the same time prove that there is both a 100 per cent and a 50 per cent chance of occurrence. If, however, it is accepted that proof and probability are separate concepts with different purposes (one scientific, one judicial), the contradiction between both the terms “will” and “could” on one hand, and “balance of probabilities” on the other, disappears. Conceptually, one is equally capable of proving, on the balance of probabilities, an almost certainty of something occurring (that it “will” almost definitely occur) or disproving a zero per cent chance of something occurring (that it “could” occur).
However, the correct threshold of these two options (“could” and “will”) is in fact determined clearly by the Act. The definition of a potential effect requires that all effects that “could” occur (that is, those of low probability) are considered under the Court’s overall broad judgment. But this does not remove
the separate requirement to prove potential effects. The Court, in assessing evidence, in some way needs to be satisfied as to the truth of any assertion of probability, in order to decide between what may be contrasting views of witnesses.Finally, it is worth noting that in Clifford Bay, Judge Jackson took a positive step in recognising that it is possible to distinguish between the concepts of proof and probability. Unfortunately, he then rejected that approach:91
It is possible, of course, to avoid the problem by stating: “we find that on the balance of probabilities there is a 10% (say) risk that a certain effect will happen”. We have four observations about that: (1) Both statements about probabilities are about our ignorance and uncertainty. They are not assessing different things ...
The three observations that followed were all based on the assumption that this first observation was true.92
5.3 Proof and Probability: Identifying the Causes of Conflation
Although the McIntyre and Shirley traditions represent very different responses to a perceived problem in the standard of proof, the problem itself stems from the same misinterpretation of the correct subject of proof. The underlying reason for this widespread conflation of the concepts of proof and probability may be twofold. Firstly, it is tempting to see the commonality in language used to describe proof and probability as meaning that the concepts are interchangeable. A standard of proof is usually couched in the language of probability, such as the balance of “probabilities”, or more “likely” than not. As such, it is tempting to see the concept of probability as adding nothing to a standard of proof that already addresses chance and likelihood.
Secondly, in practice, words are often used that contain a subtle probabilistic element, yet are not recognised as carrying this meaning. For example, it may not be immediately apparent that there is a difference between the phrases “the probability of X occurring is more likely than not” and “I consider it more likely than not that X will occur”. However, in the first phrase, the word “is” is an expression of proof, and hides an important point. It shows that there is no less than a 100 per cent satisfaction that the alleged probability is correct. The phrase could equally be translated to: “I am 100 per cent convinced that the
probability of X occurring is over 50 per cent”. The “100 per cent convinced” is an expression of proof, whereas the “over 50 per cent” is an expression of probability. In the second of the above phrases, the word “will” also tends to mask an important meaning. It shows that the thing being proved is that the effect will occur, not that there is any lower probability of an effect occurring. In other words, the phrase could equally read: “I am 51 per cent satisfied that the probability of something occurring is 100 per cent”. This is a very different test to that in the first phrase.Ultimately, proof remains important irrespective of the probability asserted, and probability remains important irrespective of the degree to which evidence is proved. A change in the standard of proof should not serve to remove effects of extremely low probability from mandatory consideration in the Court’s overall judgment, nor should it allow consideration of alleged highprobability effects where the evidence is unpersuasive. This approach is sensitive to the concern of the Privy Council in Fernandez, because it does not demand that even very lowprobability effects are ignored.93
5.4 Recognising Objective Proof and Subjective Evaluation
Simply because the Court may choose not to use the language of “proof ” or “balance of probabilities” does not necessarily mean it has rejected proof in favour of a subjective judgment.94 Often the fact-finding process is simply and legitimately characterised as one involving a qualitative preference of one witness’s evidence over another. Common formulations roughly translatable to proof on the balance of probabilities include “on balance”, “the more likely position”, findings “in favour of those expert witnesses” or “on weighing all before us”.95
Even if the Court declines to accept the evidence of any party, this does not necessarily mean that it has exercised a subjective judgment in lieu of proof. Although a finding of fact must be based on the evidence before the Court, this is not to say that the particular allegations from any given witness must be accepted.96 A “true” probability may be taken by averaging the assertions of equally persuasive witnesses, or by adopting the view of evidence commissioned by the Court itself.97
Recognising a low threshold of probability for potential effects and retaining a robust standard of proof provides an appropriately precautionary gateway to ensure that all genuine assertions of potential effects are assessed against the wide range of matters in s 104(1). Only at this point is it appropriate for the Court to consider how potential effects should be treated, by using its judgment. Thus despite the broad discretion that the Court has in assigning weight to various matters, the proof of potential effects remains important at the fact-finding stage (even if this is not characterised as traditional “fact” finding). Proof provides an objective threshold of persuasiveness below which alleged effects cannot be given weight in the Court’s judgment,98 and above which effects cannot lawfully be ignored by the Court or assigned an unreasonable degree of weight.99 Thus while the language of risk can undoubtedly be helpful, it is useful to separate out its components. If risk is perceived as comprising only the probability and potential magnitude of an effect, then this understanding of risk is amenable to objective proof. However, if policy considerations guide how risk is established, then the exercise results in a legal ruling of an evaluative or normative nature.100 It is suggested that this understanding of risk is useful in the Court’s overall judgment when weighing the positive and potential adverse effects of an activity.101 The normative quality of judicial decisionmaking in general102 is heightened in the resource management context, where the decisionmaker wields substantial discretion. However, a normative understanding of risk does not belong in the fact-finding stage where the objective identification of potential effects should occur using a consistent and predictable standard.103 What we “know” about a risk (risk identification) should be distinguished from what we choose to do with that knowledge (risk management).104 Simply because a policy seeks to avoid a particular kind of effect should not influence a decision as to whether an activity may or may not cause that effect.105
In one sense, it is possible to characterise the evaluative role of the Court as objective rather than subjective. Such descriptions have some truth, in that the Court in exercising its expert judgment clearly does not simply do as it pleases. It is guided by evidence, policies, and the purposes and principles of the Act itself. However, the use of the term “subjective” in this article is not intended to indicate a complete judicial freedom of choice. It is used rather as a descriptor to distinguish the factual objectivity of proof from the inherently greater subjectivity involved in expert discretion, where a choice must be made in the context of normatively charged questions. Such choices may still legitimately be characterised as objective, in that they are not determined by the personal prejudices of decisionmakers, but the very existence of discretion as to what “should” happen indicates a degree of objectivity less than that found in the context of factual proof. Such objectivity has substantial value in establishing how likely a potential effect is to occur.
6. DETERMINING THE APPROPRIATE STANDARD OF PROOF FOR POTENTIAL EFFECTS
It has been suggested above that the better method of establishing the existence of potential effects is via the proof of probabilities rather than recourse to the Court’s subjective judgment. It has also been suggested that the conflation of the concepts of proof and probability has led the courts in a number of cases, in the McIntyre tradition, to abandon the normal civil standard of proof as being inadequate for precautionary outcomes. However, these conclusions do not necessarily mean that the Act actually requires the implementation of a balance of probabilities standard. It would be possible to argue that the Act requires proof to some other standard, even if the thing to be proved is accepted as being an above zero probability. There are several contenders for such a standard: beyond reasonable doubt; a sliding scale according to the gravity of the question; a standard associated with a shifting evidential burden; and the balance of probabilities.
6.1 Proof Beyond Reasonable Doubt
Conceptually, it would be possible for the standard of proof in the RMA to be one of beyond reasonable doubt. Facts, including those alleging a probability of occurrence, can in theory be proved to any standard. However, two factors militate against such an approach. Firstly, the criminal standard is not one that is supported by the scheme of the Act or case law. A positive feature of the ordinary civil standard is that it applies equally to all parties (in all cases but those where matters are very finely balanced). In contrast, a standard of beyond
reasonable doubt could apply only to one party. For this to be precautionary, it would require the standard to apply only to a party alleging the absence of a potential effect. By itself there may be little practical problem with this. However, the necessary corollary of this standard is that a much lower standard of proof (any reasonable doubt) would in practice apply to those alleging the existence of a potential effect. It would be unprincipled to require the Court to accept the truth of evidence alleging a lowprobability effect where much more persuasive evidence suggested that there were no chance of such an effect. The courts have thus consistently rejected the applicability of a criminal standard.106 Secondly, a criminal standard is not necessary to achieve precaution if it is recognised that the subject of proof is any above zero probability of occurrence. A different standard is justified in the criminal context to address a presumed imbalance between the resources of the state and the private person, but also because there is often an immediate correlation between the finding of fact and a conviction under the law. No such protections are necessary in the RMA context. Although there may often be an imbalance in the resources of applicant and submitter, the general concern of sustainable management is to weigh development and protection; neither one is assumed to be more important than the other. In contrast, in the criminal context, the law has long considered it more important to safeguard the liberty of a potentially innocent accused than to facilitate convictions. In other words, it is considered preferable to acquit guilty offenders than to convict innocent persons.107 In the consenting process, the consequences of a finding of fact are neither as drastic nor immediate. Proven potential effects of low probability are simply required to be weighed in an overall broad judgment alongside other factual and nonfactual matters. Their existence need not result in consent being declined or even in the imposition of onerous conditions, and their absence need not result in a grant of consent or a lack of adequate conditions. One American commentator’s assertion that the party with a burden to disprove the existence of effects will inevitably lose, because it is impossible to prove a negative, does not recognise that findings of fact under the RMA usually do not lead to an automatic legal outcome.108 Thus an imposition of a criminal standard on an applicant when disproving potential effects would be an excessive response, and one not justified by the balance and
discretion inherent in the Act.
at 376.
6.2 Proof on a Sliding Scale: Flexibility of Application
Cases such as McIntyre, Transpower and Contact Energy have adopted the concept of a sliding scale of proof, according to the gravity of the question, to address the perceived inability of an ordinary civil standard to facilitate precautionary outcomes.109 This reflects a more general concern in the law that the normal civil standard is generally most appropriate only if the costs of error are equal for each party.110 In resource management proceedings, it is not always clear whether the costs of error are equal, because the magnitude of the cost depends on the particular proposal in question, and because costs are as likely to fall on the wider “public” environment as they are to impact on the parties to proceedings. The idea of the sliding scale can be seen as seeking to address the changeable nature of the cost of error in the consenting context.
Two broad approaches to the sliding scale can be discerned. One of them sees the standard as identical to the ordinary civil standard. Thus the High Court in Royal Forest and Bird v Buller District Council held that there is a need “to have regard to the particular context” and the “seriousness of the consequence” in determining whether an alleged fact has been proved, but:111
... so long as these considerations are observed, the standard of proof is unaltered. There is no separate and special standard of proof which falls somewhere between the criminal and civil standards.
However, such a view may be unhelpful in practice. If the ultimate standard does not actually change, it is difficult to see what practical difference it can make to have regard to the potential impact of an effect, unless it is simply an exhortation to investigate a matter in more detail.
Some cases have generally emphasised that “flexibility” in the civil standard simply means that the strength of evidence must be greater for a fact-finder to be convinced on the balance of probabilities, but that the standard itself does not change. Notable is the decision of the Supreme Court in Z v Dental Complaints Assessment Committee, in a nonresource management context.112 In the context of disciplinary (and thus quasi-criminal) proceedings, this flexibility has been well established.113 However, with respect, it is submitted that a
at 106.
113 At [107].
person’s “natural tendency to require stronger evidence” where allegations are serious cannot fit comfortably within the ordinary civil standard in the resource management context, even if it is not elevated to a formal “legal proposition”.114 Allowing for this “natural tendency” would rely on seeing the strength of evidence and the extent to which a fact-finder is persuaded as separate concepts, given that the latter is seen as immovable and the former as flexible. However, if evidence is accepted as being stronger, in practice this means that a fact finder is more convinced of the truth of an assertion. A higher standard of proof has in fact been met in such cases. In tending to require stronger evidence for a standard of proof to be met, the Court is in reality requiring that it be persuaded or convinced to a greater degree. To characterise it otherwise is to ignore thereality of the exercise.
6.3 Proof on a Sliding Scale: Altering the Standard
An alternative approach to the sliding scale, as applied in Transpower, is to concede that it can alter the ordinary civil standard from the balance of probabilities to somewhere closer to the criminal standard.115 In light of the Supreme Court’s decision in Z, this is unlikely to remain sound law. However, even if it is accepted that the standard of proof could genuinely change under the RMA, this understanding of the sliding scale remains unsuitable under the Act.
Firstly, the mischief being addressed in cases like McIntyre is the inability of an inflexible civil standard to ensure that low-probability effects are considered in the Court’s overall broad judgment. This mischief can be addressed better by the recognition that probability is a different concept from proof, and that potential effects have a low probabilistic threshold. This is because, unless a criminal standard of proof is contemplated (which is unsustainable for reasons discussed already), a sliding scale is otherwise incapable of ensuring that potential effects of extremely low probability are taken into account. In other words, if proof and probability are considered to be doing the same thing, the Act contemplates that at least in some cases (where the magnitude of effect is extremely high) the criminal standard must apply. Thus a sliding scale is unnecessary, and ultimately unable, to facilitate precautionary outcomes as envisaged in the Act.
Secondly, a standard of proof that is not the ordinary civil or criminal standard may be difficult to apply in practice. Despite the widespread use of mathematical language to describe proof, the concept of proof is not particularly mathematical. This was hinted at by the House of Lords in Davies
114 At [102], [105].
115 Transpower, above n 38, at 21.
v Taylor, where Lord Simon of Glaisdale stated, in relation to the prediction of future events, that “... mathematical theory can have in general only marginal significance. So the law ordinarily proceeds to treat probability according to certain easily understood standards.”116Mathematical expressions of the criminal and civil standards are useful tools for describing what is, in reality, a limited range of intuitive findings. The reduction of the criminal and civil standards to percentages is successful not because proof is always capable of mathematical expression, but because it happens to reflect loosely what is intuitive to human decision-making in cases of doubt.117
In contrast, other standards of proof are not likely to be practically useful, because they are less intuitive. Confusion would remain even if such math ematical expressions were reduced to nonnumerical descriptions such as “likely”, “probable” or “very convinced” because they are still descriptions of a threshold that is not particularly intuitive.
Thirdly, although the adoption of different standards is possible, and has occurred in other jurisdictions, it would be difficult to implement in the New Zealand legal system where historically it has not played a role.118 It would be likely to result in inconsistency in fact-finding under the RMA, which would be exacerbated further if the standard were to change depending on the gravity of the question.
Fourthly and finally, adopting a sliding scale of proof under the RMA that is dependent on gravity also requires circular reasoning. What happens in practice if there is a factual dispute as to whether a potential effect has a high or a low potential impact? The Court is required to apply a standard of proof to all alleged effects, including their potential impact. Yet in order to determine the applicable standard of proof to apply under a sliding scale it first needs to determine the “gravity” of the question. To determine the gravity of the question, it first needs to determine whether a potential impact is high or low. To determine this, an appropriate standard of proof needs to be applied. The Court in Clifford Bay recognised that determining the gravity of a question is difficult because one must first somehow determine for whom the question has gravity, in a context where considerations may pull in different directions.119
6.4 A Standard of Proof Associated with a Swinging Evidential Burden
The Environment Court in Shirley saw the expert judgment of the Court as being more appropriate than proof in establishing potential effects. Therefore it was considered that there existed no persuasive burden on any party to prove potential effects.120 However, the Court recognised that there is an evidential burden of proof, and this burden rests on the party which would fail if it did not produce further evidence.121 In other words, an evidential burden swings between parties; as soon as one party adduces evidence in support of its assertion, the burden shifts to the others to adduce more persuasive evidence suggesting otherwise.
Wherever a burden of proof exists, it must be associated with a standard of proof. A persuasive burden in the civil context generally requires an over 50 per cent standard of proof, whereas an evidential burden is typically associated with a much lower standard of proof (although an exact formulation has not been settled).122 A standard of proof accompanying an evidential burden in civil cases, where this is not also accompanied by a legal burden, has been described as requiring only that evidence is adduced that leaves the mind of a reasonable fact-finder in “a state of equilibrium”.123 There is no need actually to persuade or convince the fact-finder of the truth of the assertion being supported.
If it is accepted that the Court needs to exercise its judgment in determining the existence of potential effects, as concluded in Shirley, there can be no real objection to the use of a swinging evidential burden. A standard associated with this burden is not important where proof is not determinative of whether a fact is established. Thus it has largely procedural and practical rather than substantive legal value. It simply provides a framework for the presentation of evidence, and a signal to parties that they may need to produce stronger evidence if they are to persuade the Court to exercise its judgment in the way desired.
However, the concept of a swinging evidential burden is problematic if the existence of potential effects needs to be proved (as in McIntyre), rather than evaluated using the Court’s judgment (as in Shirley). This is partly because the standard associated with an evidential burden is uncertain. At its most basic, an evidential burden can be characterised as requiring that a party simply point to evidence supporting the existence of the fact alleged. Yet in North Canterbury
121 At [121].
Gas Ltd v Waimakariri District Council, the standard mentioned was simply that there be a “real risk”.124 In Ngati Kahu v Northland Regional Council, fact finding needed to be based on evidence of “probative value”.125 In Shirley itself the Court required only a “scintilla” of evidence to shift the burden.126
Despite this problem of consistency, one may conclude that a party is obliged simply to provide evidence that meets a basic level of probity, and is more persuasive than the evidence adduced against it. A more fundamental issue, however, is that meeting an evidential standard in lieu of a persuasive standard may be incapable of establishing a potential effect in a reliable or substantive way.127 If an application were non-notified or no submissions were lodged, a consent authority should not be compelled to accept as fact an alleged effect (or lack of ) simply because an applicant pointed to probative evidence tending to support the assertion. An applicant should be required to satisfy the Court to an acceptable standard that a potential effect does not exist, even if no other party exists to challenge the assertion with competing evidence.128 Furthermore, even if multiple parties produced competing evidence, a consent authority should not be obliged to accept the evidence of any one party as “proved” simply because it is more persuasive than another’s. This reflects the inquisitorial and administrative nature of the consenting regime, and the fact that its primary purpose is not simply the resolution of private disputes. Thus although the Court in Ngati Maru Iwi Authority v Auckland City Council required that a scintilla of evidence be “probative” to satisfy an evidential standard, it also stressed that an assertion must still also be proved to a “satisfactory standard” — namely, the balance of probabilities.129In practice, evidence in contested proceedings may be likely to meet the ordinary civil standard of proof if no probative rebuttal is offered in response. However, it is in theory possible for all evidence under consideration to prove unpersuasive. Thus if proof is recognised as retaining an essential role in the fact-finding process, as in the McIntyre tradition and as supported here, the concept of an evidential standard in isolation is untenable.
As with the idea of a sliding standard, the use of an evidential standard is also unnecessary to ensure precaution if the concepts of proof and probability are recognised as distinct. This is because it is likely to be no more difficult
to prove on the balance of probabilities that there is an above zero chance of occurrence, than to point to a scintilla of probative evidence suggesting that an effect will occur.
6.5 Proof on the Balance of Probabilities
Proof on the balance of probabilities appears to be the only viable standard by which potential effects can be established under the RMA. It is logically appealing, easily implemented, and well understood. It has none of the unintuitive features of a sliding scale somewhere between the civil and criminal standards, and has none of the subjective uncertainties or circular reasoning associated with setting a standard reflecting the “gravity” of an effect.
The ordinary civil standard is not itself inconsistent with a precautionary approach to fact-finding under the Act, because proof is not directly concerned with the scientific uncertainty of future events. Rather, it concerns the reliability and persuasiveness of witnesses and evidence. Scientific uncertainty is addressed already by the Act’s recognition that the subject of proof, a potential effect, incorporates a low threshold of probability. A robust civil standard of proof simply recognises that more persuasive or reliable allegations of probability should not be rejected in favour of less persuasive or reliable allegations. This is an appropriate approach in a regime where economic and environmental concerns exist in a legitimately fine balance.
7. CONCLUSION:
A CONSISTENT AND PRINCIPLED APPROACH TO PRECAUTION IN THE STANDARD OF PROOF
The approach of the courts to most questions related to precaution under the RMA has been inconsistent. This article has sought to propose a simplified and principled understanding of precaution in the standard of proof under the Act that, in most cases, is capable of rationalising what the courts have done in practice. Simplicity and consistency in the law is important, as Panckhurst J noted in Royal Forest and Bird v Buller District Council: “... the thinking in at least some ... cases is quite elaborate and, it seems to me, of a nature likely to excite grounds of appeal ...”.130
It is suggested that the use of an amended “precautionary” standard of proof under the Act is untenable. The conflation of the concepts of present and future and of proof and probability has led the courts down divergent and confusing paths. The rather optimistic statement in Ngati Kahu that the McIntyre and
Shirley approaches are “consistent in dealing with the precautionary approach” cannot realistically be sustained.131 The McIntyre tradition has recognised the operation of an altered or flexible standard of proof. However, this is problematic in that the concept of proof is targeted at the persuasiveness of evidence, rather than the scientific probabilities with which the concept of precaution is concerned. Any standard of proof short of a criminal standard appears to be inconsistent with the very low probabilistic threshold in the Act’s definition of a potential effect, if proof and probability are not distinguished, yet a criminal standard is clearly inappropriate in the inquisitorial and administrative setting of the RMA.The line of cases in the Shirley tradition has responded to a conflation of proof and probability by rejecting proof as a means of establishing the existence of potential effects, in favour of an evaluative exercise of discretionary judgment. This approach has the potential to undermine the objectivity essential in the Court’s role as a finder of fact, upon which its subsequent discretionary judgment relies.
By distinguishing between proof and probability, and recognising any above zero probability of occurrence as the correct subject of proof for potential effects, the balance of probabilities emerges as the only workable standard of proof under the Act. In assessing alleged effects in the fact-finding process, the Court is obliged to act in a precautionary manner by admitting lowprobability potential effects to its overall broad judgment. At the same time, it retains the essential ability to reject, using an intuitive and workable standard, evidence that is unreliable or unpersuasive. In practice, the recognition of an ordinary civil standard is unlikely to imperil the Act’s focus on protection, because it would not be difficult for an opponent to prove that there is any above zero probability of an effect occurring. Even where scientific uncertainty of potential effects arose from a fundamental absence of scientific knowledge or understanding of environmental processes, in most cases it would be enough for an opponent to adduce evidence of this absence. Such evidence would not be speculative or necessarily fail to meet a civil standard of proof, because proving on the balance of probabilities that there is a general absence of understanding should itself be sufficient to show at least a non-negligible chance of an effect occurring.
This article has addressed the role of precaution in the standard of proof under the RMA. Other articles address, in the same spirit, the issues of pre caution in the burden of proof, the overall broad judgment of the Court, and consent conditions.132 Ultimately, the contention of this body of work is that the
RMA can, without express amendment, provide an appropriately precautionary framework under which consenting decisions are made. The standard of proof and the subject of proof to which this standard apply are important parts of the wider precautionary decisionmaking framework, but are enabling rather than determinative of precautionary outcomes.The mechanics of the Act suggest that the area most ripe for substantive precautionary change is in the development of flexible subordinate planning documents, which influence the evaluation of policy options in a consent authority’s overall broad judgment and imposition of consent conditions, rather than in the rigid operation of procedural rules like the standard of proof.133 After all, science “does not deliver decisions”.134 This flexibility is appropriate, given that precaution does not always demand that consent be declined if potential effects exist — it simply requires some form of proportionate action to be taken to address uncertainty (for example, in the imposition of conditions).135 Dr Royden Somerville’s call in 2002 for the development of national policy statements to address environmental risk in a substantive way remains equally relevant today.136 The Environment Court should not be permitted to transform an element of judicial procedure like the standard of proof into a policy creation tool.137
However, it is imperative for lawyers and policymakers to remember that, while a legal regime addressing environmental uncertainty is essential, the greatest gains for our planet and for humanity are to be achieved by eliminating uncertainty in the first place.138 Above all, we must be active in striving to understand the Earth and our place within it. Appreciating the correct role of the standard of proof in resource management consent proceedings can help to focus attention where it is most valuable: in the development of substantive policies that stem from robust scientific research and wellfounded risk management decisions.
Judgment under the Resource Management Act” (2014) 14 Otago L Rev (publication pending).
137 At 154.
138 See BD Goldstein “Risk Management will not be Improved by Mandating Numerical Uncertainty Analysis for Risk Assessment” (1995) 63 U Cin L Rev 1599 at 1608.
NZLII:
Copyright Policy
|
Disclaimers
|
Privacy Policy
|
Feedback
URL: http://www.nzlii.org/nz/journals/NZJlEnvLaw/2014/7.html