New Zealand Pork Industry Board v Director-General of the Ministry of Agriculture and Forestry [2012] NZHC 888 (3 May 2012)

Last Updated: 29 May 2012

IN THE HIGH COURT OF NEW ZEALAND WELLINGTON REGISTRY

CIV-2011-485-719 [2012] NZHC 888

IN THE MATTER OF an Application for Judicial Review under the Judicature Amendment Act 1972 and Part 30 of the High Court Rules

BETWEEN THE NEW ZEALAND PORK INDUSTRY BOARD

Applicant

AND THE DIRECTOR-GENERAL OF THE MINISTRY OF AGRICULTURE AND FORESTRY

First Respondent

AND THE CHIEF TECHNICAL OFFICER AND BIOSECURITY NEW ZEALAND Second Respondent

Hearing: 24-25 August 2011

Counsel: F M R Cooke QC and J Kaye for Applicant

C R Gwyn, K M Muller and H Sims for Respondents

Judgment: 3 May 2012

In accordance with r 11.5, I direct the Registrar to endorse this judgment with the delivery time of 3:30pm on the 3rd May 2012.

_{JUDGMENT OF WILLIAMS J}

Solicitors:

Jonathan Kaye Law, Wellington

Crown Law Office, Wellington

THE NEW ZEALAND PORK INDUSTRY BOARD V THE DIRECTOR-GENERAL OF THE MINISTRY OF AGRICULTURE AND FORESTRY HC WN CIV-2011-485-719 [3 May 2012]

[1] This case relates to new relatively permissive biosecurity standards proposed by the Ministry of Agriculture and Forestry in 2011 for the importation of raw pig meat. The New Zealand Pork Industry Board, representing pig farmers in New Zealand, says these standards will lead to New Zealand pigs being infected with a virus known as Porcine Reproductive Respiratory Syndrome – a virus that is endemic in most other countries but not found here. The Board says this will devastate the New Zealand industry. MAF says the Board is completely wrong and the risk of incursion is negligible.

[2] The Board applies for judicial review seeking to overturn the new standards.

BACKGROUND

[3] As the name suggests, the Porcine Reproductive Respiratory Syndrome virus, known generally as PRRS, affects pigs. The PRRS virus attacks the pig’s natural disease defence mechanisms and increases its vulnerability to other diseases. It is known to cause abortions and still births in pregnant sows, as well as increased mortality and lighter live weight in young pigs. It is a potential threat to New Zealand’s pork industry if it gets in, though just how big a threat is the subject of some debate. New Zealand is one of five countries in the western world known to be PRRS-free, although I understand that there are also a number of developing countries that remain unaffected.

[4] Between 1995 and 2001, raw pig meat could be freely imported into New Zealand. Over those six years, fifty-four thousand tonnes of fresh, chilled or frozen pig meat was imported in New Zealand from countries in which PRRS was endemic. But in 2001, a Dutch study1 commissioned by the Australian government, suggested that it is possible to transmit PRRS into untainted herds by feeding them raw PRRS infected pig meat. When this study came to the attention of MAF, it promulgated new provisional import health standards (IHSs) under the Biosecurity

_{Act 1993. The new standards allowed only cooked or treated pig meat into}

1 Generally referred to as the Lelystad trial, published in Veterinary Biology in December 2003, and cited as Van der Linden et al 2002.

New Zealand from certain PRRS infected countries.2 They were intended as a precautionary measure. The fear was that pork tainted with PRRS at the time of slaughter in a PRRS-infected country could survive storage and transportation to New Zealand; that scraps of raw meat from those infected carcasses would be generated during food preparation in New Zealand homes, supermarkets or restaurants; and that those still infected scraps would be fed to a local herd.

[5] Another particularly relevant event occurred in 2001. New Zealand adopted the World Trade Organisation’s 1995 Agreement on the Application of Sanitary and Phyto-sanitary Measures (known as the SPS agreement). One purpose of the SPS agreement is to reduce the ability of WTO members to use biosecurity measures as a cover for the maintenance of illegal trade barriers.3 In part this purpose is achieved by requiring State Parties to produce scientific evidence supporting the need in biosecurity terms for any protection measures they impose.4 In the light of this, New Zealand’s provisional exclusion of raw pig meat imports carried with it an obligation to undertake a more detailed assessment of PRRS risk within a reasonable time.5 MAF had then to be satisfied on good scientific grounds that the measure remained necessary for New Zealand’s biosecurity and that it was the least trade restrictive option available. Since 2005, MAF has been engaged in that review process.

[6] In April 2011, ten years after it was imposed and six years after the commencement of the review, the Director-General of MAF relaxed the ban – subject to certain import conditions that I will explain below. The new 2011 standards now allow the resumption of raw pig meat importation.

[7] The New Zealand Pork Industry Board (I will refer to this organisation throughout this judgment as NZPIB) is opposed to this decision and wants it overturned. NZPIB fears that raw pig meat imports will inevitably cause PRRS to be introduced into New Zealand leading to the large scale infection of New Zealand

_{herds and widespread damage to the New Zealand pork industry. In this application}

2 _{They were the US, Canada, EU and Senora State in Mexico.}

3 _{See SPS Article 2(3).}

4 _{SPS Article 2(2).}

5 SPS Article 5(7).

for judicial review, NZPIB challenges the Director-General’s decision. For its part, MAF says NZPIB vastly overstates the PRRS risk and simply wants to keep pig meat sourced from countries such as Canada, Mexico, and the US out of New Zealand. Pig meat from these countries can significantly undercut the local New Zealand product on price.

[8] It is now necessary to return to 2001 and step through the processes MAF went through in its transition from the 2001 ban to the 2011 partial relaxation of the ban.

THE NEW PORK IHS

2001 review

[9] In 2001, after the provisional import ban was imposed, officials completed a limited risk analysis, more properly described as a quantitative release assessment. It concluded that the risk of PRRS transmission through consumption of chilled or frozen pork was “more than negligible”. An external technical review of the draft assessment was commissioned in mid-2002. Three international PRRS experts were instructed. The reviewers disagreed with the assessment’s conclusion. They considered that the risk of PRRS infection was negligible and that the Lelystad trial

and a later concurring study by Magar and Larochelle in Canada,6 should not be

relied upon. They suggested that the trial conditions in those studies were too different from the real life conditions in PRRS infected countries where pigs are slaughtered, stored and exported, to be of any assistance in assessing transmission risk to New Zealand.

The process under Part 3 Biosecurity Act

[10] As prescribed in the Biosecurity Act 1993, MAF relaxed the ban on raw pork imports by promulgating a new pork IHS. The process for doing so was long and

_{contested. And it was rendered even more complex than usual because of a}

6 Magar and Larochelle 2004.

significant law change mid-process. That law change and its effect on the process is a central focus of this case. It is necessary therefore to describe in some detail the process and the statutory drivers that led to the new IHS.

[11] The core idea in Part 3 of the Biosecurity Act (the part relating to importation of goods posing a threat to New Zealand’s biosecurity)7 is effective risk management. Inherent in this concept is an acceptance that completely removing risks may not always be possible or even necessary for New Zealand’s biosecurity. It may be sufficient to manage biosecurity risks down to an acceptable level. Just what is acceptable in this context will, of course, often be highly contested.

[12] Section 16 sets out the purpose of Part 3 in these terms:

The purpose of this Part of this Act is to provide for the effective

management of risks associated with the importation ... of risk goods.

[13] This purpose is imported directly into s 22: the provision by which new IHSs are prepared and released. Section 22(1) makes it clear that the purpose of an IHS is also effective risk management.

[14] Section 22 provides for a MAF officer called a Chief Technical Officer (CTO) to prepare an import risk analysis8 (IRA). This is the document that provides the scientific support for the IHS to which it relates. The CTO must then consult on that or on the draft IHS itself,9 and then recommend that the Director-General issue an IHS if this is considered necessary.10 CTOs, as the name suggests, are people

_{with subject matter expertise. They are appointed by the Director-General in light of}

7 The things to be effectively managed are “risk goods”. They are defined in the Biosecurity Act as follows:

risk goods means any organism, organic material, or other thing, or substance, that (by reason of its nature, origin, or other relevant factors) it is reasonable to suspect constitutes, harbours, or contains an organism that may—

(a) cause unwanted harm to natural and physical resources or human health in New Zealand; or

(b) interfere with the diagnosis, management, or treatment, in New Zealand, of pests or unwanted organisms

8 Obliquely referred to in s 22(7) as “a document that analyses or assesses the risks associated with the goods”.

9 _{Section 22(6).}

10 Section 22(1).

that expertise.11 It is the CTO who does most of the legwork under s 22 while the

Director-General makes the final decision.

[15] The core of s 22 is subsection (5). By its terms, the CTO must consider the following matters, before recommending a new IHS to the Director-General:

(a) the risk that organisms (as defined in the Act) will be brought into

New Zealand;

(b) the effect of such importation on people, environment and economy; (c) New Zealand’s international obligations;

(d) any other matters considered by the CTO to be relevant to the over- arching purpose of effective management of importation risks.

Import risk assessment 2006

[16] An import risk assessment on raw pig meat imports was prepared under the leadership of Howard Pharo, Manager – Animals Risk Analysis, at MAF. Once completed, the assessment was reviewed by six offshore reviewers. They were Dr Scott Dee of the University of Minnesota; Dr Jeffrey Zimmerman of Iowa State University; Dr Lis Alban, a veterinary epidemiologist from Denmark; Professor Collin Wilks from the University of Melbourne; Spirodon Doudonounakis from the Ministry of Rural Development and Food in Greece; and Torben Grubbe, Animal Health Advisor at the Danish Veterinary and Food Administration.

[17] Once reviewed, the IRA was released in July 2006. It concluded:

1. There is a low likelihood that chilled or frozen pig meat from a country with endemic PRRS will harbour the virus when imported into New Zealand.

2. Since cooking inactivates PRRS virus, and since pigs are the only species susceptible to this organism, effective exposure would require the feeding of uncooked pig meat to pigs in New Zealand.

11 Section 101.

Although scraps may be generated from imported pig meat at several points during its preparation for human consumption, the feeding of raw meat to pigs is illegal under the 2005 garbage feeding regulations. It is concluded that an exposure pathway would exist only on pig farms that were not complying with the garbage feeding regulations.

3. If pig farms in this country did become infected with PRRS through the illegal feeding of uncooked imported pig meat, the likelihood of spread to other pig farms would be low as long as standard biosecurity practices were observed.

4. If PRRS virus were introduced into New Zealand, the consequences would be significant on affected farms, particularly in breeding units.

[18] The assessment also concluded that in the unlikely event of incursion, the most likely candidates for first infection were pigs from small backyard operations. This, it was argued, was because commercial pig farmers complied with regulations that banned the feeding of raw pig meat to pigs whereas non-commercial, backyard operators were less likely to comply with, or even know about, the ban on such feeding. At the time, it was considered most unlikely that pigs from backyard operations would find their way into commercial herds. In an affidavit filed in this

application, Mr Pharo said this reasoning was confirmed in a subsequent study.12

Draft IHS 2007

[19] In reliance on the IRA, new draft pork IHSs were released in November

2007. The crucial change from the previous IHS was that raw pig meat would be admitted into the country provided it was subject to:

PREPARATION as consumer-ready cuts packaged for direct retail sale, not including minced (ground) meat, not including the head and neck, not exceeding 3kg per package, with the following tissues removed: axillary, medial and lateral iliac, sacral, iliofemoral (deep inguinal), mammary (superficial inguinal), superficial and deep popliteal, dorsal superficial cervical, ventral superficial cervical, middle superficial cervical, gluteal and ischiatic lymph nodes; and any other macroscopically visible lymphatic tissue (i.e. lymph nodes and lymphatic vessels) encountered during processing.

12 The Pearson Report 2008.

[20] The practical effect of this would be that cuts of raw pig meat could be imported into New Zealand provided they were in packages smaller than 3kg and visible lymphatic tissue had been removed.

[21] The thinking behind this was that any residual risk of PRRS incursion could be effectively managed by reducing the quantity of waste pork off-cuts entering the waste stream and finding their way into pig bins.

[22] As required by s 22(6), the CTO consulted with stakeholders, including NZPIB, on the draft. NZPIB made it clear that it considered these proposed measures to be ineffective and based on poor science. Following consultation, unchanged provisional IHSs were issued in April 2009.

SECTION 22A – THE RIGHT TO CHALLENGE THE DIRECTOR-

GENERAL’S SCIENCE

[23] In the interim (December 2007 to be precise), the Court of Appeal had made its decision in the National Beekeepers case.13 In that decision, a unanimous Court held that honey could not be imported from Australia unless separate approvals under each of the Biosecurity Act and the Hazardous Substances and New Organisms (HSNO) Act 1996 were obtained. This was because of the risk that this honey could bring with it a new “passenger” organism known as P. alvei. Such

importation, the court held, would potentially breach both the Biosecurity and HSNO Acts.

[24] In response, Parliament enacted legislation to simplify matters. The Biosecurity Amendment Act (No. 2) 2008 removed the requirement for HSNO consent in these circumstances but added a further independent check procedure to

_{the Biosecurity Act. This was done by inserting a new s 22A.}

13 National Beekeepers of New Zealand v Chief Executive of Ministry of Agriculture and Fisheries

[2007] NZCA 556.

[25] The commentary on the Bill as reported back to the House from the Primary Production Select Committee assists in understanding the reason for s 22A. The Committee said:14

... New Zealand must also maintain its biosecurity standards while meeting our

international trade obligations.

...

We agree with submitters that the process for assessing the evidence should be transparent and trustworthy. We note that most import health standards are developed in co-operation and collaboration with the appropriate sector.

We recommend the insertion of new clause 5A [becomes s 22A], requiring MAF to develop a process for an independent panel to review whether MAF had adequate regard to the scientific evidence in cases where significant concerns have arisen during the consultation process on a draft import health standard.

...

The Director-General will be required to respond formally to the recommendations of the review panel and to have regard to those recommendations when making a final decision on the import health standard. (my emphasis)

[26] Section 22A thus introduced a new participatory process into the promulgation of IHSs. The section provides for an Independent Review Panel (IRP) to be appointed by the Director-General to “review whether, in developing an import health standard, there has been sufficient regard to the scientific evidence about which a person consulted under s 22(6) has raised a significant concern.” The focus then is the science. The process for appointing and running an IRP had to be

prescribed by notice in the Gazette.15 A comprehensive notice setting out the

procedure for the appointment, conduct and reporting of an IRP was duly gazetted in

June 2008.

[27] Once an IRP report is completed pursuant to a s 22A reference, the Director- General must:16

(a) receive the report;

_{(b) take into account the findings and recommendations of the IRP;}

14 _{At page 2.}

15 _{Section 22A(1).}

16 Section 22A(3).

(c) as soon as is reasonably practicable determine the “issue in dispute”;

and

(d) give reasons for such determination.

[28] Thus, the IRP process can be engaged after the CTO consults with stakeholders in relation to a proposed IHS under s 22(6) if a stakeholder has a significant concern in relation to the scientific evidence upon which the CTO relies. The Director-General must determine the issue in dispute before deciding whether to promulgate the IHS under debate. The IRP process is designed to introduce additional transparency into the Director-General’s deliberations, but it is advisory only. The Director-General must still make the final decision on the science and on what standards are consistent with effective risk management.

IRP appointed

[29] On 9 April 2009, NZPIB triggered the s 22A process. The underlying contention was that the draft standard had significantly underestimated the risk that allowing raw pig meat imports – even in 3kg packages – would lead to the arrival of PRRS into New Zealand with devastating effect on the pork industry here. In a comprehensive submission filed by NZPIB with the Director-General at the same time as its request for an IRP, NZPIB identified nine ‘contentious scientific issues’ for review. Each issue was set out as a positive allegation as follows:

(a) the risk assessment failed to identify and further examine all potential hazards associated with a change in the HIS;

(b) the likelihood that meat for slaughter weight pigs will contain infectious PRRS was underestimated;

(c) the impact of anticipated changes to the volume of trade in pig meat (as a result of the proposed changes to the IHS) was not incorporated into the import risk assessment’

(d) the likelihood that importing PRRS infected pig meat will result in successful infection of a New Zealand pig was underestimated;

(e) knowledge of the contact structure and inter-relatedness of the New Zealand commercial and non-commercial pig industries was not adequately incorporated into the exposure or consequence of components of the IRA;

(f) the importance and likelihood of aerosol and ‘area’ spread of PRRS

between herds has been underestimated;

(g) A quantitative approach to modelling the risk of PRRS exposure and consequence was developed during the IRA/IHS process. Despite it being published in a peer reviewed journal, neither the modelling approach nor the results from the model were accepted by MAF as having enough merit to warrant any modification to the IRA or IHS;

(h) changes to the volume and distribution of the waste stream arising from modifications to the IHS were not properly taken into account in the IRA or IHS; and

(i) poor compliance with current New Zealand waste food feeding regulations has been documented but not fully considered in formulating the provisional IHS.

[30] NZPIB signalled that it wanted the operative (as opposed to provisional) IHSs retained. In the alternative, NZPIB wanted MAF to either commence work on a new IRA with more formal identification of hazards and points of dispute, or modify the provisional IHS to give tighter control over raw pig meat importation and waste stream management.

[31] The Director-General acceded to NZPIB’s request for an inquiry and duly

issued detailed terms of reference for the IRP as required by the 2008 Gazette notice.

He requested the Panel to consider whether MAF had sufficient regard to the scientific evidence in nine areas as follows:

(a) the identification and analysis of potential hazards associated with the importation of pig meat and pig meat products;

(b) the likelihood that meat from slaughter-weight pigs will contain infectious PRRS virus;

(c) the impact of changes to volumes of trade in pig meat as a result of the proposed changes in the IHSs;

(d) the impact of changes to the volume and distribution of the waste stream as a result of the proposed changes in the IHSs;

(e) the likelihood that PRRS infected imported meat will be fed to

New Zealand pigs and cause infection;

(f) the structure and inter-relatedness of the New Zealand commercial and non-commercial pig industries, and subsequent exposure and spread risks;

(g) importance and likelihood of aerosol and “area” spread of PRRS virus

between herds;

(h) quantitative modelling of the risk PRRS virus exposure and consequence, using the model developed during the IRA/IHS process;

(i) each of the above issues sits within the context of the overall assessment of risk. The Panel should consider whether MAF’s overall treatment of the issues was reasonably open on all the evidence.

[32] Apart from the last general issue, all areas arose from the issues raised by NZPIB in the submission accompanying its request for review. NZPIB’s Issue (i) relating to compliance with waste feeding regulations was not pursued in the Terms

of Reference. The general issue requiring overall assessment of the evidence was inserted by MAF.

[33] Three areas were out of scope. The Panel was not to:

(a) determine the ultimate question: what is an appropriate level of protection;

(b) comment on New Zealand’s compliance with the SPS Agreement except to evaluate whether the science used in developing the IHS is SPS consistent;

(c) comment on or recommend changes to the Biosecurity Act or related legislation.

THE IRP REPORT

[34] The IRP was chaired by the late Helen Aikman QC and comprised, in addition, Professor Katharina Stärk (Royal Veterinary College, UK), Professor John Wilesmith (retired from the Veterinary Laboratories Agency, Weybridge and Defra, UK), and Professor James McKean (Iowa State University, USA). The Panel sat between November 2009 and March 2010 – 13 meetings by teleconference and once by videoconference, otherwise by email. NZPIB and MAF both submitted to the IRP.

[35] The IRP issued its report on 31 March 2010.

[36] The executive summary provided (inter alia) as follows:

The Panel concluded that in most but not all areas MAF fully considered the science available and that it applied the recognised international standards for risk assessments. In some areas, however, there have been developments in the science and the availability of relevant data and information since MAF undertook the import risk assessment (IRA) on which the standards are based.

...

It concluded that further work may be useful to ensure a revised IRA takes into account the relevant and current science, and associated demographic and epidemiological data. In making the recommendations the Panel was mindful of the need for pragmatism and appropriate proportionality. Additional work required is not extensive. The Panel estimated that it would not take more than 6 to 9 months to complete, depending on the resources available. It suggests that, to the extent possible, MAF and the New Zealand Pork Industry Board (NZPIB) together discuss these recommendations to determine what further steps should be adopted.

[37] The Panel concluded that further work was required on MAF’s risk modelling

and that this work was best carried out in co-operation with local industry.

Regarding the relative merits of qualitative or quantitative modelling, the Panel found both approaches had advantages and disadvantages. MAF’s risk assessment was largely qualitative. The Panel does not agree with NZPIB’s criticisms of the model used, or the premise that only peer-reviewed papers should be considered. It concluded that both qualitative and quantitative approaches should be used and integrated. It suggests that a joint technical working group be established by MAF and NZPIB including experts with relevant local expertise. Any resulting information should be integrated into a revised IRA.

The Panel noted that the results obtained from such a working group, and the other information the Panel suggests that MAF obtain, might not necessarily change the conclusions of the IRA, but would ensure that it is based on more robust scientific and epidemiological data.

In the final term of reference which considers MAF’s overall treatment of the issues and openness to the evidence, the Panel considered a number of concerns about the process involved. In particular it noted the very protracted nature of the process since the identification of PRRSv in 2001, including the process leading to the IRA in 2006 and then to the publication of the provisional IHSs in April 2009.

[38] The Panel reflected on the need to accept and manage uncertainty.

It should be recognised that in some areas, scientific proof is not possible, but relevant data can nevertheless be obtained from various sources, not confined to scientific journals. In some areas MAF made assumptions based on relatively little data. Where this is unavoidable, MAF needs to state this clearly, along with a level of sensitivity for the key elements. There is a need to improve clarity and transparency by defining terms such as ‘high”, “moderate”, “low” and “negligible” in the process of conducting the risk assessment.

[39] The Panel took the view that a change in approach was required.

In conclusion, the Panel had some concerns regarding the overall process involved in obtaining and reviewing the relevant scientific evidence, and suggests that a more collegial and broad-based approach might have both

expedited the process and improved the outcome. It therefore recommends that the risk assessment be revised, taking into account these recommendations.

[40] The IRP provided discrete findings and recommendations in respect of each of the nine terms of reference. I briefly summarise each below:

Question A: Scope of standards

[41] The Panel refused to go beyond the issue of the PRRS threat to a consideration of other pathogens. A wider analysis “was not feasible within the available time” and, in any event, the documents provided to the Panel were limited to PRRS related matters.

[42] Recommendations were nonetheless made suggesting greater transparency around hazard identification and trigger events going forward. Better communication was also recommended around risks generally.

Question B: Virus levels in imported meat

[43] A key finding in this section was that it is impossible to estimate a minimum infectious dose of PRRS because there were too many variable contributing factors. It was further considered that a “safe” minimum meat scrap size cannot be determined because transmission can depend on virus concentration in each individual scrap or the volume of infected scraps consumed. The Panel also doubted MAF estimates of the fall off in infectivity as a result of the time between slaughter and consumption or as a result of freezing and thawing. The Panel’s view was that MAF estimates were overstated. The IRP recommended that:

(a) consideration should be given to the composition of pork scraps that may find their way into the waste stream;

(b) better knowledge be required of the various forms of meat treatment between slaughter and disposal including:

(i) source countries; and

(ii) the shelf life of imported chilled meat

(c) concentration reduction as a result of storage, freezing and thawing required verification.

[44] The Panel also made special mention of the fact that a study estimating likelihood of the presence of PRRS in pig meat after slaughter (the Lelystad trial mentioned earlier) had significant shortcomings. Since this was a key value in risk assessment, it was recommended that this uncertainty be expressed and sensitivity around this value be included in any subsequent risk assessment.

Question C: Impact of changes in volume of trade

[45] The IRP disclaimed any expertise in the dynamics of international trade in meat products but accepted that a “change in volume and sources of imports, during a prescribed period, is ... directly correlated with risks.” The Panel also found that there was a lack of supporting evidence for scenarios and assumptions on proportion of chilled and frozen meat likely to be imported. The Panel doubted MAF’s assumptions in this respect.

[46] Recommendations were left to the next section.

Question D: Waste stream volume

[47] The Panel noted there was an issue in this question with regulatory non- compliance – that is raw pork being fed to pigs in breach of the 2005 ban on such feeding17 – but noted that this issue was out of scope for their inquiry.18

[48] Recommendations were that MAF should consider collecting information on potential changes in volume and source of imported pork, and the importance of

17 _{Biosecurity (Meat and Food Waste for Pigs) Regulations 2005, reg 5.}

18 As I noted earlier non-compliance issues were raised by NZPIB but left out of scope in the terms of reference.

waste disposal pathways in ongoing monitoring under the IHS. Despite considering the issue to be out of scope, the Panel also recommended that further advice be sought on the effect of non-compliance with waste feeding regulations on the IRA assessment and international obligations.

Question E: Likelihood of infection

[49] The Panel questioned MAF estimates of average raw meat waste at

3.8 per cent suggesting that a UK study assessing waste at 7 per cent was an accurate estimate for that country. The Panel considered Australian estimates of 5 per cent may well be more applicable in New Zealand.

[50] In the area of commercial waste the Panel found that the IRA focused on scrap generation during preparation in restaurants and supermarkets but neglected the question of disposal due to freezer or cold store failure, or that of stock exceeding its sell-by date. The Panel concluded “disposal avenues and frequency should be derived to assess this risk consideration.”

[51] In the area of disposal pathways, the Panel said as follows:

In the suggested IHS, the restriction to a maximum size of 3 kilogram joints of uncooked/uncured pork was clearly considered to be a significant and critical risk reduction measure. The Panel’s interpretation of the IRA and the accompanying responses by MAF to comments from outside organisations and individuals is that it was assumed that this limitation would result in the generation of minimal uncooked waste from such imported cuts. This assumption was used to limit the risk presented by uncooked food waste that contains pig meat.

The Panel is concerned that there is no reliable estimate presented of the amount of uncooked pork/pig meat which is likely to be discarded from the various sources in New Zealand. (my emphasis)

[52] The Panel made five recommendations, the three more important of which were as follows:

(a) evidence to support or refute the hypothesis that 3kg cuts would significantly reduce raw pork entry into the waste stream should be found;

(b) disposal frequency pathways and weights due to freezer failure and exceeded sell by dates should be estimated;

(c) an effort should be made to determine raw pork waste proportion (in

New Zealand).

Question F: Knowledge of New Zealand industry and consequent spread risk

[53] Under this heading the Panel noted that there was a lack of information around the size, structure, movement and networks of the current pig population, and about the spectrum of compliance with biosecurity requirements in different farm formats. The Panel concluded that it was necessary to obtain current pig population data and ensure its ongoing availability. It was also felt necessary to obtain data in relation to the structure of the pig industry. In particular little was known of the non- commercial part of the industry even though it comprised 90-95% of the industry by holdings. Another crucial variable was the amount of movement between commercial and non-commercial holdings. The implication appears to be that in the absence of information in these three areas, a proper risk assessment could not be completed. Recommendations were as follows:

(i) The Panel recommends that MAF, working with NZPIB, obtain the current population data in order to provide unbiased data on pig population size, location and movements and ensure its availability.

(ii) Definitions should be agreed for terms such as “backyard”, “commercial”, “para-commercial” and “non-commercial” to delineate the population.

(iii) MAF should consider use of data collected by Pearson (2008) to formally update the risk assessment and/or simulation model. After conducting a sensitivity analysis, some targeted surveys may have to be considered to address relevant data gaps.

Question G: Importance and likelihood of aerosol and “area” spread

[54] The Panel reflected on the causes of PRRS spread within New Zealand and the understood mechanisms by which this spread could be reduced. The Panel was of the view that although simple biosecurity measures could be very effective (such as limiting inter-herd movement of animals, using basic human hygiene procedures

(changing clothes and hand washing etc) when moving between herds and cooking any pork fed to animals, uneven compliance with these procedures meant, in reality, they were unreliable as safeguards. The Panel considered that there was insufficient information about airborne (or aerosol) spread of the virus in New Zealand’s low pig density conditions, but acknowledged that it may be an important cause of PRRS spread even where other on-farm biosecurity measures are properly employed. Recommendations were made to derive better data around commercial herds in order to assess the relevance of aerosol spread compared to other transmission pathways.

Question H: Relevance of quantitative models

[55] The Panel considered that there was a role for both qualitative and quantitative (that is computer modelling based) approaches to risk assessment. The Panel acknowledged that modelling work had been carried out by Neumann and Morris from Massey University in 2007. Both experts gave evidence for NZPIB in this case. According to their model (generally called the Neumann model), the proposed new IHSs would produce an average of four PRRS infections in New Zealand each year with a 95% likelihood of multi-farm outbreak. The Panel shared some of MAF’s concerns in relation to the Neumann model, agreeing that some risk parameters that had been fed into it had been shown by subsequent research to be unrealistically high. The Panel also disagreed with NZPIB’s “major argument” that only peer reviewed papers should be used in modelling. The Panel took the view that the Neumann model was a work in progress that needed updating. Finally, the Panel expressed significant concern that MAF and the NZPIB had become unproductively polarised around the whole modelling question. Recommendations were as follows:

(i) the Panel recommends that the quantitative models be updated with the currently available information by a joint technical working group, including epidemiologists experienced in quantitative risk assessment named and agreed by MAF and stakeholders, including NZPIB. The results of the models should be presented in a report and should then be integrated in a revised version of the IRA. The Panel notes that the conclusions of the original IRA may remain unchanged.

(ii) if significant uncertain and influential model inputs are identified, the collection of additional data should be considered.

Question I: Overall assessment of risk

[56] Under this heading the IRP offered general reflections about the process that had been gone through and ways in which it could be improved from a science perspective.

[57] The IRP made findings:

(a) taking issue with terminology used for risk levels arguing that specific values should be attributed to the terms high, moderate and low, as well as to “negligible” and “non-negligible”, to make them more useful and transparent. The concern was that it was impossible for the reader to calibrate the likelihood of any particular risk outcome when confronted with these words;

(b) that there were areas in which MAF had reached conclusions that lacked supporting evidence:

There are a number of aspects in the risk analysis for which no evidence was available. One example is the amount of uncooked pork waste that is generated from the various sources and the relative amounts that enter the range of available disposal pathways in New Zealand. The Panel’s experience with risk analyses is that it is unusual not to make some effort to fill gaps in the necessary information base. Data gaps create uncertainty, especially when there is no assessment of the sensitivity of the analysis to the assumptions made on estimates. Estimating the sensitivity of likelihoods utilised and their level of certainty is important to provide a framework for a formal review and update, should more substantiated evidence become available. (my emphasis)

(c) that MAF had failed to reconsider areas in the IRA the subject of “valid and constructive comments from Professor Hurd and other peer reviewers.” The Panel believed:

There is scope for a more open-minded approach to these expert reviews which should be sought in a more structured way.

(d) casting doubt on MAF’s assessment of inter-herd transmission:

Related to this issue, the Panel noted inconsistency in MAF’s practice of accepting and acknowledging validity of evidence. MAF stated that evidence was accepted if research findings had been published in a peer reviewed journal. For some aspects, such as area spread of infection and windborne/aerosol transmission, MAF stated that it would only acknowledge the importance of these factors if supported by “scientific proof”. These particular epidemiological aspects are not open to what might be termed conventional scientific proof. Such transmission has, however, been documented during natural epidemics and the weight of the evidence is that both means of transmission occur. The pathways should be compared with other methods of between-herd transmission given the specific epidemiological circumstances in New Zealand.

More generally the Panel commented as follows:

The Panel stresses the need for MAF to express and utilise levels of uncertainty in conducting the IRA and in drawing conclusions. Data and information, with a known basis and source, provide logical resources to enable a consensus to be reached between all parties. MAF made assumptions for which there was considerable uncertainty and for which no or limited effort was reported to substantiate or refute in the course of conducting the IRA. The Panel suggests that MAF review this inconsistency in its consideration of evidence and in the acquisition of data and information to fill gaps in the knowledge base essential for the risk assessment.

[58] Recommendations were of a more generalised nature:

(i) A more collegial and structured methodology in executing risk analyses involving stakeholders and experts would aid achievement of a consensus. This methodology should encompass the process of assessing any new evidence with a bearing on the original IRA.

(ii) The possibility of combining resources with other national authorities to conduct IRAs to ease the workload on MAF staff and obtain benefits from a larger group of experts is worthy of consideration.

(iii) When conducting qualitative risk analyses there is a need to produce definitions or guidance for the terms used. The binary classifications such as “negligible” and “non-negligible” should be used with care and clarity.

(iv) The level of uncertainty associated with likelihood and risk estimates should be clearly stated together with a level of sensitivity for the key elements on the conclusions of the IRA. When information and/or data are weak or absent on aspects, consideration should be given to instigating specific investigations to resolve these data limitations. Alternatively, expert opinion should be obtained.

(v) The procedure for peer review of IRAs should be re-examined with a view to producing a more structured and specific process. Identification of specific issues should precede the selection of peer

reviewers. All comments received should be given due consideration and action.

(vi) There is scope for re-evaluating the means by which evidence is acquired, considered and accepted. All assumptions and likelihoods should have an identifiable basis and source, ideally, but not exclusively a scientific, peer-reviewed publication.

[59] NZPIB felt vindicated. There were too many gaps in the CTO’s science and, in some areas at least, officials had not gone about their work with a sufficiently open mind. The Director-General had now to decide how he would deal with the report’s wide ranging findings and recommendations.

CONSULTATION CONTINUES

[60] On 30 April 2010 a month after the report was issued, MAF held a stakeholders’ meeting to obtain feedback on the IRP findings, key recommendations and how MAF should respond to them. Mr Matthew Stone, Manager, Animal Import and Export Standards, convened the meeting. NZPIB was present along with other industry players. The EU, France, Spain, Canada, the US and Mexico all had trade or embassy officials present.

[61] At the meeting, NZPIB argued that the IHS should not now be simply backfilled in light of the IRP report. Rather, it was argued, the process should begin again with a view to issuing a new IRA. This would not be burdensome, NZPIB argued, because 95% of the base work had already been done. On the question of import volumes Mr Stone referred to Sweden: a country that had overcome PRRS importation and become disease-free. He also noted that volume, source and product type for pork imports are influenced by a wide range of market factors that can change quickly, making predictions about these variables inherently difficult and uncertain. Mr Stone implied that MAF needed to be convinced that any other prediction being advanced was better than the admittedly uncertain one in the IRA analysis.

[62] On waste stream, MAF accepted that a major report prepared in collaboration between MAF and the pork industry – the 2008 Pearson Report referred to extensively in the IRP report – had been completed too late for its findings to be

taken into account in the 2006 IRA. Mr Stone indicated that it could have a major effect on onshore waste streams analysis. It also related to New Zealand pig production and distribution.

[63] On the question of quantitative modelling there was a significant discussion, with general acceptance of the need for a multi stakeholder group to progress that issue. NZPIB agreed that this was necessary and committed to participating in the collaboration.

[64] On 2 July, Steve Stuart, Acting Deputy Director-General, wrote to Sam McIvor, Chief Executive of NZPIB, indicating that a decision was imminent on a path forward. A teleconference was held between Ministry officials and three former members of the IRP – Katharina Stärk, John Wilesmith, and James McKean – on

14 July 2010. According to Mr Stone, MAF wanted to check that the evolving work programme was consistent with the expectations of IRP members. Mr Stone described the feedback as ‘very encouraging’, although John Wilesmith who later swore an affidavit for the applicant, considered Mr Stone’s optimism was unfounded.

[65] Professor Stärk was invited in the teleconference to review the new quantitative risk model that MAF had been developing based on the Neumann 2007 version put to the IRP by NZPIB. She agreed to do so. It was acknowledged that there was a potential conflict of interest here, but the efficiencies achieved outweighed that in the view of officials. Professor Stärk’s work was completed on

5 September 2010.19

DIRECTOR-GENERAL’S FIRST DECISION

[66] Meanwhile, on 1 September 2010, the Director-General released his decision on the IRP report. This, it transpired, was not a final determination of the issues in dispute in terms of s 22A(3). It signalled further processes not specifically provided

_{for in s 22A.}

19 Since the Acting Deputy Director-General wrote to the CEO of NZPIB on 2 July to indicate that a decision was imminent on the path forward, it seems likely that MAF knew in July what work stream one was going to be.

[67] The covering letter from the Acting Deputy Director-General Biosecurity explained that a further programme of work would be undertaken before the Director-General reached a final decision on the shape of the IHSs.

[68] Two key work streams were identified. In the first, officials would upgrade the Neumann model taking into account the issues raised by the Panel. As can be seen by the fact that Professor Stärk had already been engaged to review MAF’s work in this area, this work was well advanced before the 1 September decision.

[69] The second work stream involved utilising experts in a co-operative manner to assess both the new quantitative risk model and inputs to it. An Expert Working Group (EWG) would be established to advise MAF on the validity of quantitative risk assessment conclusions and the likelihood of PRRS incursion into New Zealand. Stakeholders (including NZPIB) were invited to nominate their own experts to participate in the group by 15 September. The work would proceed from there.

[70] In support of this decision on the way forward, the CTO, the official ultimately responsible for the 2006 IRA and the proposed IHSs, provided a 24-page advice paper to the Director-General. The Director-General obviously relied upon that advice heavily. It formed the basis for his decision. The CTO’s recommendations were as follows:

1. I recommend that you use materiality as the primary criterion in determining whether to implement panel recommendations prior to making a decision on issuing the pork import health standards.

2. I recommend that the Panel’s process related recommendations be passed on to the Rules and Requirements workstream of the Border Implementation Programme for consideration. This area of work is already considering the processes by which the risks associated with importation are managed.

3. I recommend that MAF carries out quantitative modelling on the risk of a PRRS incursion. This modelling will use any recently published material that is relevant. It will incorporate dose-response data highlighted by the panel.

Consequential Recommendations

3a I recommend that once a revised model has been reviewed within

MAF, its design is peer-reviewed by external specialists in order to

ensure technical robustness. Subject to availability, MAF would seek the involvement of Independent Review Panel members.

3b Additionally, you may consider it is appropriate for stakeholders to comment on model design. If so, I recommend that MAF allows a fixed period for key stakeholders to comment on design of the quantitative model. The suggested period is four weeks, with stakeholders advised at the outset that no extensions to this time period can be granted. If implemented, this would commence as soon as the external expert review is completed.

4. I recommend that you direct MAF to take one of the following two courses:

4a MAF uses contracted external experts, including former Panel members where possible, to review input variables for quantitative modelling; or

4b MAF forms a technical working group made up of experts nominated by key stakeholders to review input variables for quantitative modelling.

5. I recommend that if you wish to provide stakeholder nominated experts an opportunity to comment on both the structure (Recommendation 3b) and input values (Recommendation 4b) of the model, you direct MAF to form an expert working group to carry out both of these processes simultaneously.

6. I recommend that sensitivity analysis is used to guide MAF and stakeholder-nominated experts in examining input variables for quantitative modelling.

7. I recommend you note that MAF does not propose to carry out studies to measure trimming and disposal of fresh pork by New Zealand consumers, but will model a range of values and test assumptions through the expert working group if such a group is formed.

8. I recommend you note that MAF has begun further investigating the effect of meat treatments such as pH and age curing on PRRS virus survival, and the potential for supply chain failures leading to disposal of meat, in line with panel recommendations B(iv) and E(ii).

9. I recommend you note that the purpose of further work is to assess whether risk management measures specified in the provisional import health standards are effective.

10. I recommend you note that output of the model, and predicted frequency of outbreaks, will largely inform my future recommendations to you on the effectiveness of measures to manage the risk of PRRS introduction.

[71] The Director-General declined Recommendation 4a preferring 4b but otherwise accepted all recommendations.

[72] In this decision matrix the Director-General isolated quantitative modelling as the primary gap in MAF’s risk assessment knowledge, accepting the Panel’s view that qualitative assessments alone were insufficient. More broadly, the Director- General signalled a desire to concentrate only on the Panel’s recommendations having the potential to affect the overall assessment materially, with more wide ranging and less focussed recommendations being directed to other generic MAF work streams for further consideration. MAF would not, however, be following the IRPs recommendation to carry out new research into how New Zealand retailers, restauranteurs and consumers trim and dispose of fresh pork waste. There would then be no work to test the 3kg hypothesis. Nor would the Director-General consider himself bound to follow the IRP’s findings and recommendations in relation to area spread.

MAF upgrades quantitative model

[73] Work stream one saw MAF working on a revised version of the Neumann

2007 model. This revision was designed to address some of the criticisms made by the IRP report. Work on the model was completed and written up in a formal report. The model focused on New Zealand’s high number of backyard piggeries, the general consensus being this was the primary risk area. According to the executive summary of that report:

The focus of this model is the number of pork meals consumed on premises that have backyard pigs, and in particular, the proportion of those meals that are derived from imported pork from infected countries. The model calculates the volume of raw meat scrap generated prior to cooking, and uses the available scientific information to estimate the likelihood that such a volume of scrap will be infectious. It is assumed that all kitchen scraps are fed to an individual pig on properties that have these animals, so the result of the model is that a primary infection in pigs is the natural outcome of generating infectious scrap.

[74] Remembering that the Neumann 2007 model predicted a mean of four infected premises per year, the MAF revised version produced strikingly different results. The report continued:

Running 10,000 simulations of our revised version of the Neumann et al

(2007) model returns a mean result of one introduction of PRRSv into a

backyard herd in New Zealand every 4,638 years. Ninety-eight percent of model iterations returned results within the range of 192 to 37,572 years.

[75] In accordance with the CTO’s recommendation 5, the modelling work was then handed on to the EWG convened by MAF and involving industry players. They would now assess both the design of the model and its inputs.

Quantitative modelling and the EWG

[76] The EWG commenced deliberations on its quantitative modelling risk brief in September 2010. It included a broad spectrum of experts nominated by MAF and stakeholders. The group was chaired by Matthew Stone: the official who had been responsible for drafting the CTO’s advice paper to the Director-General and who had also chaired the post-report stakeholders’ consultation meeting. The EWG included Professor John Wilesmith (a former IRP member), Dr Nigel Perkins, Dr David Lawton, Mr Steve Glass, together with Drs Lis Albane, Pascale Aubry, Klaus Depmer, Raymond Rowland, and Eric Neumann (the co-author of the Neumann model and an NZPIB nominee to the Working Group). Drs Francisco Zagmutt and Huybert Groenendaal also formally joined the group as active members after its first meeting. They had been contracted before the EWG was formed to provide software support to the first work stream, MAF’s in-house revision of the Neumann model. When the EWG was first set up they were also asked to provide software support to that group, but were subsequently invited by MAF to become full members as MAF nominees.

[77] The Group’s approach was to divide its analysis of the model into what was considered to be its four key parameters. It was to analyse each, one by one, in four steps as follows:

Step 1: Annual proportion of PRRS infected pig meat entering New Zealand. Step 2: Number of premises that own pigs and practice meat swill feeding. Step 3: Volume of swill generated by New Zealand households.

Step 4: Likelihood that swill will result in PRRS infection.

[78] Views amongst the EWG on the utility of the revised model they were to assess were predictably polarised. Some experts considered that the model should be completely reconstructed because it did not address the most important exposure pathway of imported pork for pigs in New Zealand, being food waste generated from non-household sources – in particular the food service industry. Others felt that the model had shortcomings but that the conclusions it produced were valid and credible. Those of this view took confidence from the participation of expert veterinary epidemiological modellers during the model’s development.

[79] As to Step 1 – the proportion of PRRS infected meat entering New Zealand – some experts felt that the assumptions for changes in import volume were far too low. On the other side it was felt that the parameters overestimated risk.

[80] In respect of Step 2 – the extent of meat swill feeding in New Zealand – one side felt that, in the particular conditions of New Zealand, between 77 and 95 tonnes of cured or fresh pork from PRRS infected countries would enter the waste stream and become available to pigs for feeding. The other side of the debate was prepared to rely on the legal prohibition on raw meat feeding into pig swill, arguing that industry non-compliance was not a valid basis to impose trade barriers.

[81] In respect of Step 3 – household swill generation – there was also a difference in views with one side arguing that food service and household raw meat scraps were higher than that assumed by the MAF model, and the other side accepting as justifiable the argument that small pre-packaged raw meat quantities would generate less waste.

[82] As to Step 4 – likelihood of PRRS infection – similar divisions arose. No agreement could be reached in respect of the likely infectiousness of meat at the point of slaughter, the effect of bleeding at slaughter, or the rate of viral decay between slaughter and consumption. Nor could agreement be reached on an acceptable range of infectiousness from tainted to non-tainted meat scraps because of

disagreements in relation to infectiousness of meat versus meat juice, effect of pig age at slaughter and raw meat scrap quantity assumptions.

[83] Other disagreements arose particularly around the significance of the period from 1995 to 2001 when there were no controls on raw meat importation and no infection had been detected. One side argued this fact proved that the conservative MAF model was appropriate. The other side considered that since exposure in New Zealand was only for six years, no specific conclusion could be reached if the risk of importation was considered by appropriate modelling to be greater than one in six years.

[84] Other issues arose around the experiences of Switzerland and Norway, where significant tonnages of raw pork from PRRS infected countries had been imported without evidence of infection.

[85] Toward the end of the EWG’s deliberations, Dr Neumann presented a restructured model representing both household and retail food sector waste to para- commercial and non-commercial herds. Food sector waste streams had not been modelled before. This was potentially significant because it brought whole cuts into the modelled waste stream where they were generated by freezer failure or sell by dates being exceeded. There was not time for it to be subjected to EWG scrutiny, but its results were markedly different again from those of the MAF model. It predicted a median of 9.8 PRRS introductions every year with mean introductions being higher still at 16.

[86] The EWG reported to MAF on 7 November. Mr Stone wrote a summary of the EWG’s work but because there was no consensus around the issues addressed, each of the experts wrote his or her own separate report. Four experts, Drs Alban, Aubry, Depner and Roland, concurred in a single report.

“Best model”

[87] Once the EWG reported, MAF instructed Drs Zagmutt and Groenendaal to review and modify Neumann’s latest model (that put to the EWG) in order to create

what MAF officials felt was a “best model” overall. Zagmutt and Groenendaal were

given five further tasks to that end. They were:

(a) review all variables and ensure they were capable of reflecting uncertainty around mean pork consumption;

(b) investigate concerns around double counting of effects on viral decay from slaughter to waste;

(c) audit the new components provided by Neumann in his second model tabled at the EWG;

(d) investigate the best approach to modelling Viraemia; and

(e) investigate the best approach to low-end extrapolation of the data on the likelihood of infection from differing infectious doses.

[88] Zagmutt and Groenendaal produced their report on 14 December 2010 including what would be the final model adopted by MAF. Their work produced a radically lower risk assessment to that provided by the Neumann models. It also concluded that officials had somewhat understated the risk in their amended version of Neumann’s first model that had been fed into the EWG. Zagmutt and Groenendaal concluded that there would be a mean of 0.038 PRRS primary introductions into New Zealand per year under the provisional IHSs. In their view, the new model predicted one PRRS outbreak on average every 1,227 years. This was considered by the authors to be a very conservative estimate.

[89] As far as I can tell, the radically different result from that derived by Neumann is due to the following key changes Zagmutt and Groenendaal made to inputs into the Neumann EWG model:

(a) the assumption in the Neumann EWG model that all scraps fed by one producer are eaten by one pig was amended to share all scraps equally to each pig in the herd;

(b) minimum necessary doses to cause an outbreak were increased so that they exceeded the lowest doses reported in the literature where those reported doses did not cause infection;

(c) Neumann relied in his modelling on estimates of food sector waste volumes provided by EWG member Mr Steve Glass. Zagmutt and Groenendaal rejected these estimates as unreliable because they were mainly based on Mr Glass’ own expert view without supporting data. Zagmutt and Groenendaal preferred MAF sourced inputs wherever they conflicted with Glass because, in their view, these inputs were referenced and more reliable;

(d) Zagmutt and Groenendaal considered Neumann had not allocated a sufficient reduction in Viraemia due to time during which meat scraps are stored at ambient temperature.

[90] MAF treated this work as resolving the scientific debate.

DIRECTOR-GENERAL’S FINAL DECISIONS

[91] On 10 April 2011, five months after the EWG report and a year after the IRP report, new Director-General, Wayne McNee, issued his decision under s 22A(3) of the Act. In text running to 34 pages the Director-General made introductory comments and then addressed each recommendation of the IRP.

[92] He pointed out that MAF had adopted the Zagmutt and Groenendaal model as “fit for purpose”, rejecting the two iterations of the Neumann model. Reliance on this model was obviously a key aspect of the Director-General’s response to the IRP report.

[93] The Director-General isolated the “matter in dispute” as required by s 22A(3)

as follows:

The matter in dispute can be described at two levels.

Overall, the matter in dispute is whether MAF has taken appropriate account of the available science in determining that the provisional import health standards provide for effective management of biosecurity risk, considering the legal obligations of section 22(5) of the Biosecurity Act 1993.

In detail, the matters in dispute are effectively summarised as each of the individual matters in the terms of reference MAF established for the Panel. These are referred to in the terms of reference under the following headings:

1. Scope of the standards

2. Virus levels in imported meat

3. Impact of changes in volume of trade

4. Likelihood of infection

5. Knowledge of New Zealand industry, and consequent spread risk

6. Relevance of quantitative models

7. Overall assessment of risk20

[94] The Director-General made a determination in respect of the overall matter in dispute before turning to address the seven particular matters. On the overall matter, the Director-General concluded:

The final import health standards, and MAF’s process to develop them (including the additional work undertaken in response to the Panel), has taken appropriate account of the available science, and will provide for effective management of biosecurity risks, considering the legal obligations under s 22(5) of the Biosecurity Act.21

[95] The Director-General did not directly address any of the detailed findings of the IRP. It was implicit in many of the Director-General’s responses that he felt MAF’s subsequent modelling work had overtaken the Panel’s report by plugging the gaps in the science that the Panel had identified as the IRA’s primary deficiencies.

[96] On the scope of standards (TOR Question A), the Director-General found that there were no matters requiring a decision in respect of the IHSs and passed the recommendations on to the Border Change Programme for consideration in that

_{wider context.}

20 Note that heading 3 incorporated Questions (c) and (d) of the original terms of reference, while Question 5 incorporated Questions (f) and (g). That explains why there are only seven headings rather than nine in the final determination.

21 A further detailed amendment to the IHS was proposed but it is of no relevance here.

[97] On the question of virus levels in imported meat (TOR Question B), the Director-General concluded generally that the new Zagmutt and Groenendaal model properly addressed the IRP’s recommendations in relation to diagnostic tests and infectious dose. He found that it also properly addressed uncertainty and the impact of different parameters for risk.

[98] On the question of impact of changes in volume of trade (TOR Questions C and D), the Director-General again relied on the Zagmutt and Groenendaal model. He found that it “appropriately used historical data22 and provided logical explanations for input values and distributions on the likely volume and source of imported pork.” He also found that the model “considered the relative importance of waste disposal pathways for imported pork and estimates of compliance with waste

feeding requirements on those pathways.”

[99] In respect of the likelihood of infection (Question E), the Director-General once again found that the model robustly assessed risk. He found that adjustments made to the model by the EWG process included the introduction of separate waste streams for backyard pigs, exposed via household waste, on the one hand, and para- commercial and non-commercial pig herds, on the other, exposed via food service and retail sectors.

[100] In addition, the Director-General was satisfied that the 3kg limit was a readily verifiable requirement which was least trade restrictive in the context of risk management. He also rejected the risk of freezer failure or inventory management resulting in significant quantities of raw pork finding its way directly into the waste stream. This risk was considered to be negligible. Further, non-compliance with pig swill laws was not a significant contributor because, he said, MAF has developed and implemented a programme of education, inspections and prosecutions designed to increase compliance.

[101] In respect of concerns expressed by the IRP as to MAF’s knowledge of the

structure of the New Zealand industry and consequent PRRS spread risk

_{(Question F), the Director-General rejected the need for further data gathering in}

22 Presumably the 1995-2001 open gate policy.

relation to pig population size, location and movement. The Director-General concluded that concerns expressed by Neumann and others in 2007 in respect of animal movements from low health status para-commercial herds to high health status commercial herds were “significantly overstated”. Since new modelling from Zagmutt and Groenendaal limited the likelihood of outbreak to one every 1,227 years, further data gathering in this area was disproportionately expensive given the extent of risk.

[102] For the same reasons, the Director-General considered aerosol and ‘area’ spread of PRRS (Question G) was an insignificant risk and rejected any suggestion of further study in that area.

[103] In respect of the relevance of quantitative models (Question H), the Director- General considered that the further work of the EWG and further modelling by Zagmutt and Groenendaal met all recommendations in this area. In particular, stakeholder groups were extensively involved, new research material was incorporated into assessments and the resulting updated model provided a conservative estimate of risk that is consistent with the 2006 IRA.

[104] In respect of overall assessment of risk (Question I), once again the new modelling was considered to be a complete answer to any concerns expressed by stakeholder groups.

[105] On 13 April 2011, the Director-General approved new IHSs for pig meat. Except in one respect (not relevant to these proceedings) the IHSs were substantially the same as those initially published in draft in 2009.

ARGUMENTS

[106] NZPIB mounts its attack on three fronts:

(a) the Director-General has failed to comply with the express requirements of ss 22 and 22A;

(b) the Director-General’s decision followed an improper process; and

(c) the Director-General and/or MAF more generally, failed to consult on the Zagmutt and Groenendaal model.

[107] On the first ground – misapplication of the express terms of ss 22 and 22A –

NZPIB argued:

(a) the Director-General’s decision was not made “as soon as reasonably practicable” after the IRP report as required by s 22A(3). Rather it was made one year after the report and following a completely new and unmandated post-report process;

(b) the Director-General did not “determine the issue in dispute” following receipt of the IRP report, as required by s 22A(3) but determined his own reformulated issues;

(c) the Director-General relied on the updated Zagmutt and Groenendaal risk analysis that was produced by an entirely different process, rather than the IRP report as required by s 22A;

(d) the Director-General failed to consider the findings of the IRP as required by s 22A(3), referring only to the Panel’s recommendations;

(e) the Director-General failed to determine the issues contained in Questions (a) and (g) as required by s 22A(3), relying on wrong advice that such determinations were no longer necessary.

[108] The procedural impropriety claim has two aspects. The first argument is that by creating the EWG and updating the Neumann model as a new basis for risk assessment, the Director-General adopted a procedure that was not mandated by s 22 or s 22A. The second argument is a challenge to the impartiality of key officials. The argument is that it was their biased belief in their own pre-determined views, in preference to arguments being put forward by NZPIB among others, that led to the

Director-General constructing a new process to paper over the problems identified in the IRP.

[109] The consultation argument is that the Zagmutt and Groenendaal model was so different from Neumann’s model, and such a game changer, that either the Director-General or MAF more generally should have consulted on it, before using it as the basis for departing from the IRP’s findings and recommendations.

[110] MAF replies that the IRP is an advisory process to assist the Director-General in making regulatory decisions with a significant science content. It is argued that the Panel is not a dispute resolution process and the Director-General is not bound by its result. MAF argues that NZPIB’s construction of the statute is designed to frustrate rather than advance the relevant statutory purpose.

[111] Before I address these arguments, it is important to remind myself what my role is in this application. This is driven as much by the way in which NZPIB has pitched its case, as it is by the strict limitations imposed upon me by the ordinary principles of judicial review. NZPIB’s case is focused entirely on MAF’s decision- making process in the broadest sense. The issues are confined to whether MAF properly understood the tasks given it by ss 22 and 22A and whether it complied with the ordinary public law requirements of fair process. It will be a relief to scientists everywhere to know that I am not called upon to decide whether (or indeed, which of) the CTO, the IRP, the EWG or the Director-General, got their science right. That judgment will be made by history in due course.

[112] I turn now to address each argument is more detail.

THE DIRECTOR-GENERAL FAILED TO COMPLY WITH SS 22 AND 22A

[113] As indicated there are five aspects to this argument, and I will address each in turn.

“As soon as reasonably practicable”

Arguments

[114] Section 22A(3) requires the Director-General to receive the IRP’s report and to determine the issue in dispute “as soon as is reasonably practicable”, once he has taken into account the report’s findings and recommendations. He must also give reasons.

[115] NZPIB argues that the Director-General did not comply with the “as soon as reasonably “practicable” requirement. It was argued that this wording left no room for the Director-General to appoint an ad hoc expert working group and, after that, undertake the new modelling completed by Zagmutt and Groenendaal. NZPIB argued all that was open to the Director-General was to vote the IHS up or down, and if, in light of the IRP report in this case, the response was to vote it down, then the s

22 process would have to start again.

[116] The Crown on the other hand argued that the IRP process was not a statutory dispute resolution procedure with a winner and loser in the way of orthodox litigation. Rather, it was a scientific review process and so needed to be sufficiently flexible to cope with the subject matter. The Crown pointed to the fact that the IRP itself indicated that further work should be done. This would take six to nine months the Panel suggested in the executive summary. The Crown pointed out that NZPIB witnesses Dr Wilesmith and Prof Morris agreed that this additional work needed to be done and agreed also with the process.

Analysis

[117] It is important to remember that the purpose of this part of the Act is to develop rules and policies and to facilitate the making of decisions that advance the effective management of risk. I agree with the Ministry that ss 22 and 22A were intended by Parliament to interlink in a flexible rather than rigid way. That approach is made necessary by the complex and (at least sometimes) open-ended nature of both the subject matter of the IRP’s inquiry and the analysis the Director-General

must ultimately deploy in promulgating an IHS. As always, “as soon as reasonably practicable” is to be read in its context and circumstances. One of those circumstances is that both the IRP and the Director-General are addressing the “imponderable”23 of future risk of PRRS incursion. It is perhaps not surprising therefore that eminent New Zealand and international experts in various aspects of the science of this particular risk calculation, seem to disagree sharply in their

assessment. That is another circumstance. These elements are in turn exemplified in the IRP’s report and recommendations. The report reads rather more like a comprehensive scientific peer review than a special statutory inquiry. It is wide ranging and discursive, posing more questions than it answers. Given the dominance of eminent scientists on the Panel, this is perhaps not surprising – even if the Ministry found the discursive result a little inconvenient. The important point is that the Panel itself knew and expected that the gaps it identified in the evidence in support of the IHS would take some time to plug – six to nine months of additional research and analysis. The Panel incidentally was keeping its powder dry as to whether that analysis would make any difference to the result. It said so expressly at the outset.

[118] “As soon as reasonably practicable”, could not, in light of the subject matter

of the inquiry and the result produced, be taken to mean immediately.

[119] In another context, a similar phrase was construed by the Court of Appeal in Malone v R.24 There at issue was s 45 of the Evidence Act requiring a formal procedure for identification evidence to be undertaken “as soon as practicable” after an offence is alleged to have been committed. Randerson J for the Court said:25

The expression “as soon as practicable” means the formal procedure must be carried out as soon as it is feasible to do so in the circumstances. This involves issues of practicality ...

[120] In the present case, the IRP’s brief was to determine whether the CTO had paid “sufficient regard to the scientific evidence” in relation to the matters of

23 Director of Animal and Plant Quarantine v Australian Pork Limited & Ors [2005] FCAFC 206; (2005) 224 ALR 103 at [61].

24 _{Malone v R [2010] NZCA 59.}

25 At [19].

significant concern that had been raised by NZPIB. The Director-General’s brief

_{was to:}

receive the IRP’s report;

take into account its findings and recommendations;

determine the issue as soon as is reasonably practicable; and provide reasons.

[121] No specific timeframe is given in s 22A(3). I am sure that was an intentional omission.26 In the present case the Director-General could not be in a position to make a determination until after the six to nine month period of gap filling work had been completed as recommended by the IRP. Equally self evidently, the imprecise and open ended nature of the calculations the Director-General is required to make under s 22A suggest that this kind of delay is contemplated by the phrase “as soon as

reasonably practicable”. That phrase must be able to accommodate a science-based inquiry about future risk that sometimes, as here, will produce more questions than answers.

[122] There was, and remains, of course a dispute between NZPIB and MAF over the basic science underpinning the IRA and the IHS. Within that dispute the Panel’s mandate was only to determine whether there had been sufficient regard to the scientific evidence. It did not itself determine the matter, that was for the Director- General. Thus it is the Director-General who must ultimately resolve the dispute. But that does not make the s 22A process a dispute resolution process in the way argued for by NZPIB. It was not a process apt to produce a winner and a loser – at

_{least not every time. In this case, there were too many shades of possibility to}

26 Unlike Whangamata Marina Society v Attorney General [2007] 1 NZLR 252 (HC) in which the requirement in s 119 of the Resource Management Act that the Minister of Conservation provide his or her decision on the report and recommendations of the Environment Court within 20 working days. This constrained timeframe was seen by Fogarty J as an indicator of the permitted width of the Minister’s inquiry.

produce, under the Director-General’s hand, what might be described as an orthodox

judgment.

Issues wrongly formulated into a single overall question

Introduction

[123] In the final advice paper to the Director-General the CTO identified two levels of dispute.

Overall, the matter in dispute is whether MAF has taken appropriate account of the available science in determining that the provisional import health standards provide for effective management of biosecurity risk, considering the legal obligations of s 22(5) of the Biosecurity Act 1993.

[124] The advice paper then identified the detailed issues raised by NZPIB in its original submission to the IRP. Although Issue 7 (overall assessment of risk) was MAF generated, as I have said.

Arguments

[125] NZPIB argued that the overall issue was never posed by NZPIB. This was an unwarranted reformulation by MAF. Its effect was to diminish to sub-issues the real issues that NZPIB wanted put to the Panel. These were the “significant concerns” raised by “a person consulted”, in terms of s 22A(1).

[126] NZPIB argued further that paragraph 16 of the terms of reference excluded the Panel from considering this issue anyway. The terms of reference provided:

It is not the role of the panel to:

Determine what is an appropriate level of protection against which the import health standards should be assessed; or

comment on New Zealand’s compliance with the SPS Agreement, beyond evaluating whether the use of scientific evidence in developing the import health standards meets its international obligations as a signatory to the SPS Agreement.

[127] NZPIB also argued under this heading that MAF had made a fundamental error in approaching the issue by splitting the scientific evidence around the quantification of risk from the allegedly public policy question of effective management of it. NZPIB argued that MAF was wrong to approach its task on the basis that risk quantification is a s 22A issue but effective management is a s 22 issue. In fact risk and effective management are substantially overlapping, NZPIB argued, and ironically, this is reflected in MAF’s reformulation of the overall question.

[128] NZPIB argues that the effect of posing the overall question was to reduce the threshold MAF was required to meet under s 22A and to cause it to apply the less rigorous standard under s 22(5) including, as the latter section does, non science- based mitigators such as international obligations.27 In other words by introducing s 22(5), MAF wrongly brought into the s 22A analysis the classic trade-off between the science of risk, a truly s 22A matter, and the political imperatives of free trade for

a nation reliant on exports for its survival. Section 22A, NZPIB said, is not about trade, it is about science.

[129] The Crown argued that the overall question posed in the advice paper and answered by the Director-General was in fact contained in paragraph 15 of the terms of reference:

The panel is appointed to consider whether MAF, in developing the provisional import health standards, had sufficient regard to the scientific evidence about which the Board has raised a significant concern. The panel should assess whether MAF’s treatment of the issues, given all the evidence, was reasonably open to it, and whether there is a reasonable chain of logic linking the science to the provisional import health standards.

Analysis

[130] I accept that s 22A is designed to provide a process of scientific inquiry into issues raised by a stake holder. It is, as NZPIB says, those issues that must be the

_{focus, not any reformulation or recasting by MAF. IRP inquiries must be about}

27 See for example s 22(5)(c).

science and they must be about the scientific issues raised by the relevant stakeholder as a matter of significant concern to it. Section 22A(1) is explicit about that. There should not, in my view have been an attempt to step back and recast the debate. The statement of an overall issue both in the terms of reference and in the advice paper over-stepped the disciplined line required to be maintained by s 22A.

[131] Having said that, I do not see that any harm was done by what, in the end, should be seen as a technical misstep. The Director-General did in fact address each of the issues raised by NZPIB. A review of the advice papers, appendices, consultation documents and so forth amply demonstrates that the science around NZPIB’s issues was carefully assessed by the Panel, the CTO and officials and the Director-General himself. Thus the statement of an overall issue was a slip but, in the end, it had no material effect on the quality of analysis applied to the issues NZPIB did raise.

[132] Similarly, the inclusion of a reference to s 22(5) in the s 22A decision was unhelpful but not fatal on a first run at these processes. Parliament’s insertion of s 22A left the decision paths for the CTO and Director-General even more complex than they already were and the drafting is somewhat opaque. It is clear from the workings that led to the final s 22A decision that the result was completely science focussed. It would be wrong in assessing the impact of the s 22(5) reference to ignore that wider and more important context.

Wrong detailed questions and answers

Arguments

[133] NZPIB argues that the answers ultimately given by the Director-General do not in fact address the questions originally set out in the terms of reference. NZPIB identifies Issue (i) as an illustration. The questions posed were as follows:

Taking collectively all of your findings, are there any deficiencies in the regard MAF has given to the science that should cause MAF to reconsider

_{the provisional import health standards?}

If there are any deficiencies, what steps does MAF need to take to give

_{sufficient regard to the science?}

[134] NZPIB says it did not ask these questions as matters of significant concern under s 22A(1). Under Issue (i) the Director-General’s decision refers to the up-

_{dated Zagmutt and Groenendaal model as effectively answering all questions. Thus:}

The development of the up-dated quantitative risk assessment has drawn on identified sources of data where these are available. Where new evidence was provided during the EWG process, this has been considered in the final quantitative model developed by the expert consultants.

...

The up-dated quantitative risk assessment provides a conservative estimate of the risk of introducing PRRS virus into New Zealand through the importation of pig meat and gathering of further data is considered unnecessary.

...

The up-dated quantitative risk assessment and subsequent EWG process has facilitated input from affected stake holder groups in the spirit of collegiality suggested by the panel. However it was not possible to achieve consensus due to the earlier establishment of polarised opinions.

_...

[135] NZPIB argues further that this decision of the Director-General does not answer the initial question but rather avoids it.

[136] The Crown argued that when the Director-General’s decision is taken as a whole and in context, it is clear enough. There were no relevant deficiencies in the science MAF employed in its IHS.

Analysis

[137] Here, NZPIB’s attack appears to be that MAF’s over reliance on the Zagmutt and Groenendaal model meant it never answered the question in Issue (i) or indeed in any of the other questions.

[138] I do not think that proposition is correct. What actually happened was that the Panel pointed out a number of deficiencies in MAF’s science and MAF spent a year remedying that (as suggested by the Panel) before handing the matter (together with the new work) on to the Director-General for final determination. By that time the new work had pushed the debate forward. It is to be remembered also that the Zagmutt and Groenendaal model was not entirely new work. It had built on the original Neumann model that had been before the IRP, and then on Neumann’s subsequent iteration provided right at the end of the EWG’s work.

[139] In MAF’s and the Director-General’s view, the new modelling made irrelevant the initial deficiencies pointed out by the IRP. The Panel contemplated this as a possible outcome. Recommendation H(ii) provided, it will be remembered:

if significant uncertain and influential model inputs are identified, the collection of additional data should be considered.

[140] This can only be a reference to the additional data identified under other issue headings: trimming quantities, waste streams, area spread and so forth. The flipside of this is that if inputs in these areas were not influential – that is if the model’s assessment of risk was not sensitive to variable inputs in these areas – then there would be no need to collect additional data in those areas.

[141] That is why it was felt that was a sufficient answer on the science under each issue heading, even where the IRP had concluded there were deficiencies. On the Director-General’s view of the utility of the model, it made good sense to hold it up as a complete answer in most areas. Whether he and Drs Zagmutt and Groenendaal are right in that is not a question for me. All I can conclude is that there was no difficulty with the process the Director-General adopted in this respect or the particular answers he gave to the specific issues he was required to consider.

Director-General considered the wrong risk analysis

Arguments

[142] NZPIB argued that the Director-General erroneously relied on the up-dated Zagmutt and Groenendaal quantitative risk analysis, arrived at following a completely new process when he should have relied on the IRP report alone. NZPIB argued that s 22A(3) required the Director-General to make his decision based on the IRP’s findings and recommendations together with the material that fed into the original IRA and IHS. By relying on evidence not before the IRP, the Director- General failed to comply with the requirements of s 22A.

[143] The Crown says that in fact the Director-General did what the IRP recommended.28 It formed the EWG to review inputs to the quantitative risk analysis in the hope that this would assist the various sides to build consensus.

Analysis

[144] Generally speaking the problems identified by the IRP were information gaps in MAF’s analysis rather than wrong interpretations of settled scientific facts. The focus of the IRP’s recommendations was therefore to fill the gaps identified. The Panel’s recommendations in Issue H on quantitative risk models are an example of this. While it may have been possible in a different kind of debate under s 22A for the Director-General to simply resolve the debate on the papers as it were, either by following an IRP recommendation or rejecting it, that was not possible given the IRP’s report in this case. The Panel said in most areas (but especially in modelling where input variables could have a multiplier effect on risk quantification) the Director-General needed to do more work. It is hard to see how he could have done anything but undertake new research and engage in new processes if he was to properly address the findings and recommendations of the IRP. By convening the EWG and, when that process failed to find consensus, undertaking his own

_{modelling work, the Director-General was doing exactly what the IRP contemplated.}

28 See recommendation H(i).

[145] There is a downstream question about whether, having undertaken that work, the Director-General should have recommenced consultation under s 22(6) or whether he was under a more general implied obligation to go back to the parties. I will return to that matter below.

Failure to consider the IRP’s findings and over-focus on its recommendations

Arguments

[146] NZPIB argues that the Director-General was able to avoid the IRP’s inconvenient findings by ignoring them and focusing on the recommendations that were less sharp. In doing so, the Director-General does not give reasons for accepting or rejecting the IRP’s findings. This does not comply with the requirements of s 22A(3), NZPIB argues.

[147] NZPIB relied on the Court of Appeal decision in Air Nelson Limited v Minister of Transport29 in which, in a fee setting exercise, the Minister was not told of Air Nelson’s opposition to fee hikes nor the grounds for that opposition. The misleading impression left for the Minister was that Air Nelson accepted the increase when it had not. In this case, NZPIB argued that the CTO did not advise the Director-General about the IRP’s criticisms of the IRA. These criticisms were contained in the IRP’s findings but the CTO’s advice paper focused only on their recommendations. NZPIB argued that this failure undermined the Director-

General’s decision just as the MOT’s failure to properly advise its Minister had done

in the Air Nelson case.

[148] The Crown accepts that the CTO’s advice did not point out the various criticisms contained in the IRP’s findings. However, the Crown argues that, unlike the Minister of Transport in the Air Nelson case, the Director-General had the actual IRP report before him and he confirmed in his affidavit that he had read it. It is irrelevant, the Crown argues, that the IRP’s findings were not generally referred to in

_{the CTO’s advice.}

29 Air Nelson Limited v Minister of Transport [2008] NZCA 26; [2008] NZAR 139 (CA).

Analysis

[149] In his affidavit, the Director-General said:

Alongside the decision [advice] paper I had 20 annexes totalling approximately 900 pages, including the full Panel Report and the expert working group report. I read all of the material before me carefully and in many cases more than once. This includes the text of the Panel Report. I found the way in which the material in the decision papers was set out clear and useful, given the voluminous material which I had received and read in support of that paper.

[150] While I am not generally supportive of statutory deciders having another bite at the cherry, as it were, by further explaining in later (and often self serving) affidavits, further reasons for their decisions, it is in order to take account of the foregoing passage from the Director-General’s affidavit. It confirms what the record already shows: that extensive material was placed before the Director-General for his consideration and that this included both the findings and the recommendations of the IRP.

[151] The Air Nelson decision has no application to this case. The Director- General did take into account the negative findings of the IRP.

Failure to confront Issue A (non-PRRS hazards) and Issue G (area spread)

Background and arguments

[152] Issue A related to the identification and analysis of all potential hazards associated with the importation of pig meat and pig meat products. It necessarily required an assessment beyond just the PRRS virus. MAF considered that this issue could not practically be addressed in the context of this dispute (although it had been expressly referred to in the terms of reference). The advice paper provided that no decision was required on this issue as these matters were already covered in MAF’s separate border change programme.

[153] NZPIB argued that this amounted to a refusal to determine the issue after taking into account the findings and recommendations of the IRP and giving reasons

– the process required by s 22A(3).

[154] On the question of non PRRS hazards, the Crown argues that no adverse finding was made. Rather, the Panel said that it did not wish to look beyond the PRRS issue because there was no time; because it was only given PRRS information; and because PRRS was the pressing issue in light of New Zealand’s SPS obligations and the operation of the provisional standard.

[155] The Crown points out that the IRP made an observation about the way in which MAF goes about identifying hazards generally because this does impact on PRRS as well as other hazards. Two recommendations are made, essentially suggesting that MAF be more transparent over hazard identification.

[156] Issue G related to the importance and likelihood of “aerosol” and area spread of PRRS virus between herds. The advice paper simply provided that no decision was required since the likelihood of infection in New Zealand was now known to be too low to warrant further study of onward spread.

[157] NZPIB pointed to the following IRP findings on area spread:

7. ... MAF’s reliance on biosecurity to obviate virus movement between production sites following an introduction of PRRSv into New Zealand may be overly optimistic.

8. Since the original risk assessment was completed experimental information has buttressed clinical observations about inter-herd aerosol viral movement ...

...

10. Aerosol transmission may contribute to inter-herd virus spread in the presence of otherwise effective biosecurity measures. Even low rates of inter-herd transmission by this method can be important in extending infection geographically. ...

...

13. Size of operation and geographic density are major considerations when attempting to estimate the impact of aerosol introduction of PRRSv within an area. An accurate profile of the New Zealand pig industry for commercial, para-commercial and non- commercial production types was not available from the AgriBase data in 2006 ...

[158] NZPIB highlighted the following recommendations:

(i) The Panel recommends that data on location, size and quantities of each production class around commercial herds be improved to assess the relevance of aerosol spread relative to other transmission pathways.

[159] And:

(ii) As indicated in section 2.8, the Panel recommends that the more recent evidence on the likely level of biosecurity practices relevant to area spread is considered.

[160] NZPIB argued that the Director-General simply sidestepped these findings and recommendations and therefore, by definition, ignored them in breach of s 22A.

[161] On the question of aerosol and area spread, the Panel’s view was there was little laboratory or epidemiological evidence available in 2006 and that MAF should keep monitoring studies and other evidence as it is generated. The Panel also said that little is known about the structure of New Zealand’s pig industry and that information about this is required to determine relative density and intensity. Density and intensity are important contributors to the efficiency of aerosol transmission. The Crown argued that it is wrong to consider that these were findings. Rather they were in the nature of comments about on-going, larger studies that needed to be undertaken.

[162] The Crown argues that the Director-General considered these comments and then concluded that the revised quantitative risk analysis model suggests that the risk of PRRS getting on to farms is now so remote that it was not worth putting the effort into area/aerosol spread studies or monitoring. The Crown argues that the Director- General was entitled to take this view.

Analysis

[163] Section 22A requires the Director-General to determine the issue in dispute after taking into account the findings and recommendations of the IRP in relation to that issue. The issues as articulated in the terms of reference were:

(1) Issue A: the identification and analysis of potential hazards of pig importation;

(2) Issue G: the importance and likelihood of area and aerosol spread of

PRRS.

[164] On the first issue, the IRP made no findings on overall hazards for the reasons given but, in recommendations, suggested that MAF’s hazard identification work should be more transparent and systematic than it is currently. On the second issue, the IRP felt it had insufficient information to make findings quantifying the risks associated with aerosol and area spread, but accepted that this was a risk area on which further and on-going work should be undertaken. If there is a finding at all, it is that if PRRS does infect a New Zealand herd, further area and aerosol spread to other herds cannot be discounted in New Zealand conditions without further research.

[165] On Issue A, the Director-General took account of the recommendations in concluding that the IRP’s discussion was not primarily PRRS related and that wider issues were in any event being handled in MAF’s Border Change Programme. As I read the report, there were no findings made under this heading. There might be cause to wonder at the efficacy of posing Issue A. Perhaps it would have been wiser to either restrict the Issue A inquiry to PRRS or to leave the issue out altogether. In effect both the IRP and the Director-General took the view that this term of reference was too wide and insufficiently relevant to spend time on it given the demands of the moment. That must surely have been open to the Panel given that it had more time to reflect on the question (and perhaps more expertise) than the officials who drafted the terms of reference. The Panel gave reasons for the stance it took. Those reasons appear logical and the underlying considerations relevant.

[166] The Director-General was entitled to conclude, in effect, that there was no issue to be determined anymore.

[167] On Issue G, the Director-General clearly took account of the IRP’s findings and recommendations, such as they were, in concluding that quantitative modelling of the risk of PRRS entering New Zealand and infecting live pigs suggested strongly that the risk identified by the IRP was too slight to justify further work in those areas. The new modelling had taken the scientific debate beyond the risk of area spread, at least in the Director-General’s view: a view he was entitled to take on the evidence. The important point under this heading is that the Director-General had before him all of this material, including the IRP report itself. He was aware (because he had been briefed by officials in December 2010 and April 2011) of the way the debate had evolved from the Issue G decision in the report to the EWG and then the ‘best model’. There is no merit in the argument that the Director-General ignored the report’s discussion of Issue G.

Conclusion

[168] I conclude that the Director-General complied with the express requirements of ss 22 and 22A in terms of NZPIB’s challenges outlined above.

IMPROPER PROCEDURES

[169] There are two aspects to this challenge. The first has been briefly touched upon already. It is essentially that there is no room in ss 22 and 22A for the EWG and the Zagmutt and Groenendaal modelling work. The second is that key biosecurity New Zealand staff should not have continued to be involved in decision making after the IRP reported.

No room for processes not expressly mandated

Arguments

[170] NZPIB argues that the Director-General was not permitted to establish the EWG or undertake the revised quantitative risk modelling completed by Zagmutt and Groenendaal. This was because the scheme of ss 22 and 22A did not permit him to embark on what amounted to a whole new formal inquiry before determining the issues after the IRP had reported. NZPIB argues that the Director-General must

_{determine the issue in dispute:}

As soon as reasonably practicable;

taking into account the findings and recommendations of the IRP;

_{giving reasons for each determination.}

[171] NZPIB accepted that the Director-General could take advice before making his s 22A decision but says that this is not the same as embarking upon a whole new process.

[172] The Crown argues that the EWG was mandated because the Panel itself recommended it. Beyond that, the Crown argues that it is the responsibility of the Director-General to determine the issue in dispute. There is a subtle but significant difference between that and determining whether to follow the IRP’s recommendations or not. Again, the Crown argued that the IRP inquiry and the Director-General’s discretion had by necessity to be treated flexibly, dealing as it does with unknown and perhaps unknowable risks. In addition, the Director-General must ultimately take account of matters beyond the scientific evidence which was the focus of the IRP. Section 22(5) required him to take account of New Zealand’s SPS obligations, any resourcing issues and time constraints as well as the overall impact of the pork industry on the New Zealand economy and environment. The Crown argues that the Director-General decides what is effective management of this

particular risk in light of the scientific evidence, but not in light only of the scientific evidence.

[173] Both sides call on the decision of Fogarty J in Whangamata Marina Society v Attorney General.30 In that case, the Environment Court was required by s 119 of the Resource Management Act31 to inquire into a proposal to build a marina on the Whangamata estuary. The Environment Court’s role was to make recommendations as to whether consent should be granted, but the final decision belonged to the Minister of Conservation. Fogarty J found that s 119 did not permit the Minister to

undertake his own factual inquiry – he was bound by the Environment Court’s findings of fact. Nor could he go beyond those findings into new areas of inquiry without first going back to the Environment Court for that court to undertake its own inquiry into any new area of interest for the Minister.

[174] NZPIB argues that the situation in Whangamata Marina was essentially similar to the relationship between the IRP inquiry and the Director-General’s determination under s 22A(3). It is to be remembered, NZPIB argues, that s 22A was inserted to provide an independent science assessment before the Director- General resolved the debate. NZPIB argues that because the Director-General is not himself an expert, he is unlikely to be in a position to disagree with IRP findings since the IRP comprised eminent subject matter experts.

[175] The Crown argues that s 119 of the Resource Management Act and ss 22 and

22A of the Biosecurity Act are not comparable.

Analysis

[176] I agree with the Crown. In my view, there are three reasons why the Director-General has more discretion under this legislation than the Minister of Conservation had under s 119 RMA. The first reason is that the respective processes are too different. The Director-General here must, I agree, resolve each debate

_{following an IRP inquiry, but the process is science rather than law driven. The}

30 _{Whangamata Marina Society v Attorney General (supra footnote 26).}

31 This section is now repealed.

terms of reference for the IRP made it clear that there would be no formal hearing and that the Panel could take any advice and obtain any information it saw fit in discharge of its functions.

[177] The relevant provisions of the terms of reference are as follows:

20. The Panel will not conduct any hearings except with the prior approval of the Director-General of MAF, in consultation with the Board.

21. The Panel may carry out its review by meeting together, or by communicating with each other by whatever means promotes the efficient completion of the Panel’s function.

23. The Panel may, if necessary, seek additional relevant information from whomever it sees fit in meeting the terms of reference.

24. Following the provision of documents from MAF and the Board, neither MAF nor the New Board may communicate with the Panel on matters of substance in relation to its review, except as requested by the Panel. If the Panel requests any material from either party, a copy of the material will also be supplied to the other party. MAF may contact the Board on matters of administration, as required.

[178] The second factor is that the subject matter of the inquiry in this case is too different from that facing the Minister of Conservation in the Whangamata case. The terms of reference themselves make the point:

12. Under the SPS Agreement a key question for MAF is whether scientific evidence is sufficient in quantity and quality to complete a justifiable risk analysis. Almost all risk analyses are conducted in situations where the scientific evidence is incomplete and a balance must be sought between trying to acquire complete knowledge and making predictions with a reasonable level of confidence.

13. There is rarely a single interpretation of any scientific issue.

Differences in interpretation are normal and a vital part of testing how robust is the evidence supporting any particular

conclusion. The key issue is to consider the weight of evidence

and whether there is a coherent relationship between the evidence and the conclusions drawn from it.

14. Science can never prove a complete absence of risk, but it is the basis for an assessment of risk and supporting measures to manage risk. Decisions on import health standards must be made on the basis of the best information available at the time. When issuing import health standards, or declining to issue them, the Director-General must be satisfied that there is sufficient information to support the assessment of risk and the measures identified to effectively manage that risk.

[179] The subject matter is incapable of producing a precise resolution. Judgments in this area are inherently value based. In this way, the present case is very similar to that confronting the Federal Court of Australia in Director of Animal and Plant Quarantine v Australian Pork Limited & Ors.32 In that case, Australian biosecurity authorities were dealing with the risks associated with a porcine disease known by the acronym PMWS. The legislation governing decision-making was somewhat different from that in New Zealand but the issues were very similar indeed. In describing the nature of the judgments that needed to be made, the court said:33

The legislation does not suggest that quarantine decisions are to be made on an assumption that every scientific fact is known about every conceivable disease or pest that might be introduced into Australia, or that such decisions are to be delayed until all such facts are discovered and accepted. On the contrary, quarantine decisions have to be made in the existing state of knowledge. Imponderables have to be weighed and value judgments made. No specific criteria are laid down, other than the condition to be established must limit the level of quarantine risk to one which is “acceptably low” – which necessarily assumes there will be some risk.

[180] Finally it is my view that the IRP report was by its nature completely different from the adjudicatory report produced by the Environment Court in the Whangamata case. Here the IRP’s findings and recommendations reflect its science bias and its flexible process. The findings and recommendations are often open- ended and discursive, erring on the side of maximising knowledge collection and factual inputs in accordance with science culture – all despite paragraphs 12-14 of the terms of reference. This is very different from the strictly legal and adversarial

_{process that is the hallmark of Environment Court inquiries.}

32 _{Director of Animal and Plant Quarantine v Australian Pork Limited & Ors (supra footnote 23).}

33 At [61].

[181] The Crown is correct that the report itself asked for the very process NZPIB now complains of. Indeed NZPIB supported the new process in the consultation round and sent its own expert, Dr Neumann, to participate.

[182] As indicated above, the IRP report identified gaps in the information necessary (in the Panel’s view) to quantify and thereby effectively manage the risk of PRRS incursion. Those gaps needed to be filled or set aside as no longer significant. A choice could only be made between these two options after further investigation – the IRP itself did not have the answers. In those circumstances, s 22A(3) could only work to enable the Director-General to determine the issue in dispute if there was enough room in its provisions to permit that extra work to be done. It could have been done by an additional transparent collaborative process or MAF could have taken it in-house. In the event both options were utilised in this case, the first one failing to produce a sufficient consensus.

[183] NZPIB argues that the Director-General should simply have voted the standard down and sent it back for another IRA round, but that is a counsel of legal perfection and inapt for a process which, at its core, combines difficult science and broad policy. In my view, given the subject matter of ss 22 and 22A, the processes they mandate can only result in effective management of biosecurity risks if the statutory players are accorded a reasonable degree of latitude in the way they go about making their assessments. NZPIB’s approach risks, in an area of evolving knowledge, the creation of a permanent loop of argument over science from which there is no escape.

Conclusion

[184] I conclude that the post IRP work undertaken by MAF before the Director- General decision under s 22(1) was permitted.

Biosecurity New Zealand staff should not have been involved

Arguments

[185] NZPIB argues that four key officials should have been excluded from further involvement in these matters after the IRP reported. They were the CTO, Mr Knox, his primary adviser and technical support Mr Stone, Mr Pharo who had been involved in PRRS issues since the 2001 IHS changes and was the lead official on quantitative modelling work, together with Mr Cobb who worked with him. NZPIB argued that the IRP had so roundly criticised these officials that it was inappropriate for them to continue in their roles because they would effectively become appellate judges in their own cause. NZPIB pointed to the fact that the IRP had criticised

“MAF” in this way:34

The panel believes that MAF was too hasty to dismiss solicited comments. MAF did not apparently reconsider aspects of the IRA despite a number of valid and constructive comments from Prof Hurd and other peer reviewers. The panel believes that there is scope for a more open minded approach to these expert views which should be sought in a more structured way.

[186] NZPIB complained that despite these criticisms, Messrs Pharo and Cobb went on to lead the modelling work in the EWG and the revised modelling by Zagmutt and Groenendaal. Similarly, Mr Stone drafted the terms of reference for the EWG and all three of these officials participated in it. Finally, and most importantly, Mr Stone, NZPIB complained, drafted both of the Director-General’s key decisions in August 2010 and April 2011 following the IRP report in close co-ordination with Mr Knox.

[187] NZPIB argues further that the August decision that shifted the focus to quantitative risk modelling set up the means by which officials could avoid a critical report. That is, it was designed to generate competing information to enable the IRP to be effectively by-passed. The work programme that followed after the August

_{decision could only, in NZPIB’s submission, have meant that no recommendations}

34 IRP report at [16].

other than those contained in H would be followed. This, NZPIB argues, amounted to pre-determination.

[188] NZPIB argues that the correct process should have been to treat the IRP as the Director-General’s sole adviser on the science underlying the IHS when making his determination under s 22A(3). It was improper for officials who had been criticised to then draft the Director-General’s decisions, adopted, as they were, word for word. This ran the risk that officials would be seen to be advising the Director- General for the wrong reasons – that is to defend their own original positions.

[189] For the Crown four arguments are made.

[190] First it is argued that the role of the IRP was not to resolve any dispute between MAF officials and NZPIB. In fact, it was the Director-General who determined any dispute not the Panel.

[191] Second, officials had no disqualifying interest in the process. They were interested only in protecting the integrity of the biosecurity system. The Crown said that the report was not in fact so roundly critical of officials as to give them a personal interest in the outcome.

[192] Third, the Crown argues that the recommendations identifying information gaps were about obtaining information for the future – the material itself did not yet exist and so could not be accessed to assist in the decision that was required to be made at that point. The statutory standard of “sufficient regard” to the “scientific evidence” does not cover evidence that can only be collected in the future and on an on-going basis.

[193] The Crown accepts that it was frustrating that the Panel did not itself distinguish between present and future-based recommendations but says that did not undermine the position of officials. It says further that this was the first time s 22A had been used, so some latitude was warranted.

[194] Finally, it is argued that officials throughout the public sector routinely advise statutory deciders including in circumstances where they have provided prior advice preparatory to a public participation process. If officials were not free to do this, the efficiency of government would be undermined.

Analysis

[195] There are relevant leading decisions in New Zealand, Australia and Canada in relation to bias and pre-determination. These decisions relate to officials exercising statutory discretions and to officials whose task it is to advise or assist statutory decision makers. It will be necessary briefly to review these authorities and a trial level decision before turning to apply the relevant principles to this case.

[196] In CREEDNZ v Governor-General35 the plaintiff attacked a declaration by the Governor-General-in-Council under s 3(3) of the National Development Act

1979 allowing the Government to fast track authorisations for what became known as the Bluff Aluminium Smelter. One of the arguments mounted was that the responsible ministers had demonstrated pre-determination of the issue in public statements they had made prior to the order. These statements made it clear that the Government vigorously supported fast tracking. Cooke J took the view that prior expressions of commitment do not disqualify the decision makers. In his view, pre- determination had no application in the particular context of the National Development Act and it was better to approach the matter on an orthodox relevant considerations basis. Richardson J considered that the rules against pre- determination ought still to apply but “with the utmost caution”. In an oft cited phrase, His Honour took the view that the standard to be applied to a pre-

determination test will depend on the context. He said:36

What is fair in a given situation must depend on the circumstances. The application of the rules against bias must be tempered with realism.

[197] Richardson J considered that it was wrong to expect completely open minds or perfect detachment when responsible Ministers were considering the exercise of

35 _{[1981] 1 NZLR 172 (CA).}

36 At [194].

the s 3(3) discretion. Rather what must be established is that their minds are not so closed and so incapable of being changed that they are really just going through the motions in applying the provision. Seen in context, any requirement of aloof detachment would be to depoliticise what was a highly political issue.

[198] In a similar vein (but a very different context) was the decision of Baragwanath J in Friends of Turitea v Palmerston North City Council.37 In that case the Council entered into a contract with Mighty River Power Limited to allow that company to establish a wind farm on Turitea bush reserve. The company would pay the Council $650,000 in three instalments: the first at execution of the agreement, the second when the reserve’s purpose was changed to accommodate the wind farm,

and the third when consents were granted.

[199] The agreement provided that it would operate without prejudice to the Council’s independent exercise of its consenting powers under the Resource Management Act.

[200] Baragwanath J found there was no pre-determination in that case. The requirement to be open-minded, he found, did not mean without pre-disposition in the context of this land owning Council. It meant, he said, being “prepared despite pre-disposition, honestly to consider whether to change its mind”. The agreement’s express exclusion of any commitment to granting resource consents was also an important element in the Judge’s conclusion.

[201] In Baker v Canada (Ministry of Citizenship and Immigration) 38 the appellant, a Jamaican citizen, challenged a decision by immigration officer Caden to deport her despite her having four Canadian born children. The notes of an interview between the appellant and a more junior immigration officer, G Lorenz, were in issue. They were used by officer Caden in making his decision.

_{[202] L’Heureux-Dubé J reviewed these notes and took the following view:}39

37 _{Friends of Turitea v Palmerston North City Council [2008] 2 NZLR 661 (HC).}

38 _{Baker v Canada (Ministry of Citizenship and Immigration) [1999] 2 SCR 817.}

39 At 851.

Whether they were intended in this manner or not, these [notes] give the impression that Officer Lorenz may have been drawing conclusions based not on the evidence before him, but on the fact that Ms Baker was a single mother with several children, and had been diagnosed with a psychiatric illness. His use of capitals to highlight the number of Ms Baker’s children may also suggest to a reader that this was a reason to deny her status. Reading his comments, I do not believe that a reasonable and well informed member of the community would conclude that he had approached this case with the impartiality appropriate to a decision made by an immigration officer. It would appear to a reasonable observer that his own frustration with the “system” interfered with his duty to consider impartially whether the appellant’s admission should be facilitated owing to humanitarian or compassionate considerations. I conclude that the notes of Officer Lorenz demonstrate a reasonable apprehension of bias.

[203] The fact that Officer Lorenz did not actually make the decision did not

change the position in Her Honour’s view:40

In my opinion, the duty to act fairly and therefore in a manner that does not give rise to a reasonable apprehension of bias applies to all immigration officers who play a significant role in the making of decisions, whether they are subordinate reviewing officers, or those who make the final decision. The subordinate officer plays an important part in the process, and if a person with such a central role does not act impartially, the decision itself cannot be said to have been made in an impartial manner. (my emphasis)

[204] In Hot Holdings Pty Limited v Creasy41 the High Court of Australia considered whether two officials who assisted in preparing advice to a minister in granting a mining license had sufficient “interests” in the decision to show bias. The first official, Miasi, owned shares in the applicant mining company. The second official, Philips had an adult son who held shares. All Judges agreed that bias can be fatal to a statutory decision even where the biased official is not the decision maker him or herself. The majority comprising Gleeson CJ, Gaudron, Gummow, Hayne and McHugh JJ (echoing Baker) added the gloss that the biased official must have assisted or advised the statutory decision maker in a material or significant way. In this case it was found on the facts that Miasi’s role was merely to record a decision that had been made by others. This role, it was held, was too peripheral to lead a fair minded and informed member of the public to suspect that the content of the

_{departmental advice recorded in the minute was influenced or affected by Mr Miasi}

40 _{At 849.}

41 Hot Holdings Pty Limited v Creasy [2002] HCA 51; (2002) 210 CLR 438 (HCA).

or his interest in the mining company. The majority also found that the relationship between Mr Philips and his shareholding son was insufficient to found an allegation that Mr Philips – and therefore the Minister – could not have been seen by fair minded and informed members of the public to have acted impartially.42

[205] Taking these decisions together, it is possible to identify a series of basic principles applicable to this case.

[206] First, in the ordinary course of the business of government, it is expected that officials will provide advice, sometimes to a level of great detail, about the way in which ministers or other statutory decision-makers should exercise their discretions.43 Second even though such officials do not exercise that discretion themselves, they must not provide any advice or assistance that is more than peripheral to the statutory delegate, if they have a personal interest in the matter.

[207] Third, although it is generally not permitted to approach the exercise of a statutory discretion with a pre-determined view of the outcome, pre-determination does not necessarily mean having no pre-disposition at all. In some statutory contexts – those where pre-disposition is a structural requirement or a policy or political reality – it will be acceptable if a delegate with a pre-disposed view is “prepared despite pre-disposition, honestly to consider whether to change [his or her]

mind”.44 The principle espoused in CREEDNZ and Turitea will apply not just to the

statutory decision-maker him or herself, but logically also to the officials who advise and assist those decision-makers in some material way.

[208] In the present case there is nothing in the wording of s 22A that suggests that officials should be removed from their usual role as advisors or assistants to the Director-General when he “determines the issue”. The IRP does not replace officials and their unique role in our system of executive government. Its role is to be an

_{independent analyst providing the transparency Parliament sought to insert into the}

42 In an eloquent dissent Kirby J found that in this case the integrity of the public service, “a precious heritage”, would be tainted and undermined even by the limited involvement of Miasi and the perception of the link between Mr Philip’s and his son’s shares.

43 See especially Richardson J in CREEDNZ at 200–201 but also implicitly, Hot Holdings Pty and

_Baker.

44 Turitea (supra footnote 37 at [102]).

Biosecurity Act as a replacement for the loss of the HSNO process. But it did not remove the Director-General as the ultimate decider. Nor, in my view, did its existence require the official infrastructure around him to be cut away. There is no policy reason to remove support officials from the equation even if their own decisions are under review – as is the position with the CTO. In fact, given the wide and imprecise nature of the science of PRRS and future risk management, the limited number of experts in MAF capable of dealing with these issues, and the Director- General’s own lack of subject matter expertise in making his s 22A(3) science based decision, the indicators all support advisory officials remaining in place throughout. The nature of the IRP report in this particular case supports this. The report is discursive – perhaps unavoidably so – because it must be science and knowledge focused rather than regulatory and policy focused. It is very obviously a report penned by scientists not officials or lawyers.

[209] Having said that, officials must of course remain alert to circumstances that might give rise to perceptions of bias and pre-determination in the minds of reasonable, fair minded and informed members of the public. NZPIB says that the IRP criticised Knox, Stone, Pharo and Cobb so sharply in its report that to allow them to continue to advise the Director-General – particularly to the extent of drafting his decisions – would create an impression in fair minded people aware of the circumstance that the officials would be interested in covering their backs by overriding the IRP’s findings and recommendations, that those officials could well be biased. As Cooke J in Whangamata and Baragwanath J in Turitea both considered, such a circumstance would also lead to the conclusion that decisions were being made pursuant to irrelevant considerations.

[210] Having reviewed the relevant material in some detail, it is my view that the IRP’s criticisms of officials are not nearly as sharp as NZPIB would have it. The Crown is right that such criticisms as there are, must be seen in context. The IRP’s executive summary provides that broader view:

... in most but not all areas, MAF fully considered the science available ...

[211] It is necessary to retraverse the criticisms and the way they were dealt with in order to get a proper fix on this issue.

[212] The recommendation that was most clearly intended to affect the IRA and

IHS drafting processes was that relating to modelling. The IRP said:

The results of the models should be presented in a report and then be integrated in a revised version of the IRA.

[213] The IRP adds an important rider:

The panel notes that the conclusions of the original IRA may remain unchanged.

[214] And then:

If significant uncertainties and influential model inputs are identified, the collection of additional data should be considered.

[215] Thus additional data could, the IRP felt, be required if the new model is shown to be sensitive to particular input variables. Zagmutt and Groenendaal found, using Neumann’s second model, that most fields in the model were not significantly sensitive to input variables. I will mention the sensitive fields below.

[216] The Panel did, in its overall assessment, criticise MAF’s approach to some

data. The Panel felt:

(a) there were gaps in the data and it was unusual not to make some attempt to fill them;

(b) MAF could be more open minded about expert reviews of its work and in its approach to accepting what data was credible.

[217] But these complaints were not backed up by recommendations in the way that the quantitative modelling area was. It follows that issues relating to sufficiency of data were intended to be filtered through the modelling programme to ensure that any extra work would actually make a difference. IRP members were sensitised to the need to avoid research for its own sake.

[218] The Panel acknowledged that there were other problems associated with waste streams generally and the lack of proof around the three kilogram limit actually succeeding in reducing trimmings to a very small amount. The IRP proposed extra work in those areas but, even so, the Panel did not intend major work. Rather, the Panel accepted that the new work might be of six to nine months duration only. There was no argument about re-starting the process. There was also, the Panel accepted, the possibility that the additional work would not result in any change to the IHS.

[219] It is thus wrong to say that the CTO and other officials were so sharply criticised that their on-going involvement had become “personal” – or at least a fair minded observer properly informed of the facts could think that. The CTO’s advice to the Director-General containing recommendations that became the Director- General’s August decision on the way forward, was frank and reasoned. He said:

While MAF could dispute the assumptions or interpretations of facts behind some individual recommendations, or argue that some recommendations fall outside the panel’s terms of reference or the scope of s 22A of the Act, the general tenor of the report is clear.

... the panel considers that MAF did not have sufficient regard to the available science in some areas under consideration.

[220] The advice went on to suggest that the Director-General adopt a “materiality” standard for assessing recommendations. Just as the Panel had said, only recommendations that would make a difference should be considered for implementation. It was suggested that future-focussed process related recommendations go off to the Rules and Requirements work stream for consideration there. Quantitative modelling should be up-dated, it was recommended, with in-puts to the model reviewed by the EWG including stakeholder experts.

[221] There is no evidence here of defensiveness or avoidance of the report’s criticisms. Rather the indication is that a sorting process was required given that the IRP’s views were science-rich but policy naive. As I have said before, given the science heavy make-up of the Panel, that result could hardly be said to be surprising.

[222] The advisory paper identified eight of the IRP’s recommendations for

additional data gathering. They were set out in a table as follows:

_{Reference Recommendation MAF Comment}

Rec. B(iv) Composition of discarded pork

No NZ data available. Could build range of assumptions into model. Difficult to gather robust data.

Rec. B(v) Range of post-slaughter treatments imported pork has been subjected to

Data available in EFSA report. MAF will consider newly-published data on the effect of pH and age-curing on virus half-life.

Rec. D(i) Potential changes in volume and source of imported pork

Inherently speculative but could build range of scenarios into model

Rec. D(ii) Impact of non-compliance with waste feeding regulations

Some compliance information from the Prime report (2009). Could attempt to gather more by survey

Rec. E (interim injunction)

Impact of pork disposal resulting from freezer failure and stock passing sell-by date

Information only available from industry sources. MAF will interview key industry participants.

Rec. E (iii) Composition of waste fed to pigs

No NZ data available. Likely to be a significant variable. Difficult to gather robust data.

Rec. F(i) Current NZ pig population data

Work by the Prime report
(2009) has provided
detailed information on this point.

Rec. G(i) Data on pigs in proximity to commercial herds

Could be assessed by survey, but not feasible to keep updated. Only relevant if disease established in NZ.

[223] In the vexed area of trimming and waste, the advice given appears to me to be carefully reasoned, suggesting that there is no reliable way of testing the three kilogram hypothesis except that of seeing how it works in practice. It is perhaps

significant that the IRP offered no guidance in this respect. The advice provided as follows:

MAF has investigated gathering information on the trimming practices of domestic pork consumers, and how this would translate into the volume of fresh pork meat in food waste. Feedback from survey experts indicates that it would be time-consuming and expensive to gather data on trimming and disposal of fresh pork, and there would still be significant questions about the accuracy of the information.

Phone surveys, web-based questionnaires and face-to-face interviews all have significant weaknesses in asking people to estimate the percentage or the total weight of meat trimmed from pork cuts during preparation for cooking. Trimming and other practices that generate meat waste are likely to vary widely, and there is a high likelihood of either inaccurate estimation or behaviour being modified under close observation. MAF will therefore model a range of trim percentages based on information previously discussed during the import health standard development process. This could be subject to refinement by an expert working group, if you direct MAF to form such a group.

Conclusion

[224] I would conclude therefore that there is no reason to be found either in the structure of s 22A or in the conduct of these officials to suggest that they should have been removed from the process before the Director-General made his s 22A(3) decision for reasons of bias or pre-determination.

FAILURE TO CONSULT Arguments

[225] NZPIB argues that the revised model completed by Zagmutt and Groenendaal trumped all that had proceeded it including the IRP report itself. It was argued that this new model was so obviously important that the Director-General ought to have consulted on it before finally proceeding to rely on it.

[226] The new model effectively changed the basis of the IRA that had been published in 2006, it was argued, and that change triggered new consultation obligations either under s 22(6) and (7) or as a matter of general principle. It was

argued that by refusing to consult – or even release – the revised model before final decisions were made, MAF intentionally sought to avoid leaving open the possibility of a second s 22A IRP process.

[227] The Crown’s arguments were essentially practical. Dr Neumann had produced his revised model – the model on which Zagmutt and Groenendaal worked

– at the last minute. The EWG was timed to report and dissolve. It was not possible therefore to engage with the EWG on this new development. In any event, the Crown argued, the revised model was very much a product of the work of the EWG which was itself a form of consultation. There was, the Crown argued, no implied obligation to go back to the parties once the EWG had completed its work. A right to consultation does not include a right to be included in the formative stage of the decision-maker’s thinking, the Crown argued.

Analysis

[228] Determining the nature and extent of a statutory decision-makers obligation to consult is primarily an exercise in statutory interpretation. As Cooke J suggested in Daganaysi v Minister of Immigration:45

The requirements of natural justice vary with the power which is exercised and the circumstances.

[229] And as Gallen J pointed out in Greenpeace v Minister of Fisheries46 it is also important to understand the objective of the consultation as expressed or implied in the statute.

[230] In this case the formal consultation obligation is owed by the CTO under s

22(6) and (7). It is to consult on either or both of the IRA and draft IHS before handing it on to the Director-General for decision under s 22(1). If the consultee has raised a significant concern in the context of such consultation, about the scientific

_{evidence upon which the Director-General relies, then an IRP can be appointed to}

45 _{Daganaysi v Minister of Immigration [1980] 2 NZLR 130 (CA) at 140.}

46 Greenpeace v Minister of Fisheries HC Wellington CP 492/93 27 November 1995 at 16.

consider whether there has been sufficient regard to the scientific evidence about that concern in preparing either the IRA or draft IHS.

[231] Both the formal consultation phase and the IRP process were completed in this case. The question is should the CTO have gone back again, or should the Director-General have commenced his own consultation before making his final decision under s 22(1)?

[232] There is some authority for implying into a consultation round a requirement to make a second bite available. In New Zealand Co-op Dairy Company v Commerce Commission47 the plaintiff attacked the Commerce Commission’s decision in part on failure to give a further opportunity for consultation. Wylie J in the High Court sitting with Mr R S Blunt as assessor, said:

We recognise that the procedure which the Commission normally follows, i.e. the circulation of a draft determination inviting responses, and the holding of a conference of the interested parties at which evidence and submissions in response to the draft are received, goes a long way towards meeting the requirements of fairness. But the Commission must always be alert to ensure that the fairness encouraged by that procedure is not undermined by what occurs thereafter. In other words the procedure to that point does not necessarily exhaust the requirements of fairness. It may be necessary to go further by ensuring that any matter of substance which may then emerge and which is likely to be material to the final determination is made known to the applicant or principal participant with an opportunity for reply. In the end it is a matter of degree, weighing the practicalities of the sometimes tight time restraints under which the Commission must operate against the importance of the material involved. ...

Ultimately the Commission must stand back and ask itself whether the use it is intending to make of a particular idea or item of evidence is fair, without first inviting a response from those whom it might affect.

[233] In this case, there was an initial round of consultation, an IRP inquiry and then the EWG. I agree with NZPIB that the EWG cannot be seen as consultation in the strict sense, but it can hardly be set to one side as irrelevant to this question. As the court said in the Co-operative Dairy case, in the end it is a matter of degree weighing the practicalities and (I would add), the history and context of the matter. Here the inputs to the revised model were extensively debated through the IRP and at

_{the EWG. The platform on which Zagmutt and Groenendaal built their work was}

47 New Zealand Co-Op Dairy Company v Commerce Commission [1992] 1 NZLR 601 (HC).

provided by Dr Neumann who was paid by NZPIB to participate in the EWG. It seems that the key change between the Neumann revision and the Zagmutt and Groenendaal revision was that the latter rejected the data supplied by Mr Steve Glass and replaced it with MAF sourced data. This issue, too, was extensively debated in the EWG. The model was not in a true sense new information. It was a new arrangement of information involving a selection of some existing data in preference to other existing data. The substantive content had all been debated. A further consultation round would have served no purpose and in the end the statutory decider had to make a decision. The alternative was to take this around yet another round of consultation and (probably) a second IRP loop after five years of consultative process. That just cannot be consistent with s 22’s objective of effective risk management.

Conclusion

[234] I conclude there was no obligation on MAF to undertake a further round of consultation.

RESULT

[235] There will be judgment for the respondents. They will be entitled to costs for which brief memoranda may be filed if necessary.

Williams J