Day: March 28, 2023

Summary/TLDR

• Sir Peter Gluckman and Anne Bardsley have called for an apolitical NZ national risk assessment.
• We support this call but note several weaknesses of existing national risk assessments both in New Zealand and globally.
• The new assessment must be systematic, include global catastrophic risks, engage stakeholders, cooperate with other countries, and appropriately represent risk.
• For more on global catastrophe and NZ resilience, you can read about our ongoing Aotearoa NZ Catastrophe Resilience Project (NZCat).

Risk and New Zealand

Risk analysis will be topical in New Zealand in 2023 argues a recent opinion by former NZ Chief Science Adviser Sir Peter Gluckman and Dr Anne Bardsley: Risk listening: rethinking how we understand and manage risk

Ongoing disruptions from the Covid-19 pandemic, and more recently extreme weather in the form of cyclone Gabrielle have highlighted New Zealand’s infrastructure gap and cut across multiple government agencies and ministries.

Gluckman and Bardsley note that there will be ongoing acute events, increasing climate impacts, and harms from a degrading digital and information environment. Programmes such as the UN’s sustainable development goals (SDGs), the Paris Climate Accords, and the Sendai Framework for Disaster Risk Reduction are all intended to work in concert to reduce risk and optimise development by 2030. However, progress remains variable, and the systemic disruption of the Covid-19 pandemic has impacted progress.

The opinion also notes that current shortcomings of the risk assessment process include:

• Insufficient transdisciplinary science and knowledge brokerage between science and policy.
• A low priority given to developing redundancy and resilience (reactive rather than proactive policy).
• The problem of short-termism in decision making.
• There is insufficient ‘risk listening’, ie decision makers are not open to taking on board the risk assessments

Gluckman and Bardsley’s article proposes that NZ undertake an ‘extended’ and ‘independent’ national risk assessment. The resulting product must be ‘public facing as well as policy facing’. We support this call.

However, we note that national risk assessment processes of the past have exhibited significant weaknesses. These will need to be explicitly overcome to ensure that these assessments can act as a repository for risk information and a foundation for action.

Problems with existing national risk assessments

Some very recent publications that have raised concerns about national risk assessment processes include the following:

• A February 2023 report by Kevin Kohler of ETH Zurich on cross-border risks that describes discrepancies and shortcomings in their assessment across nine European national risk assessments.
• An Feb 2023 academic paper describing two key weaknesses of existing national risk assessment processes, namely exclusion of almost all the major risks, and a lack of stakeholder engagement.
• A March 2023 report commissioned by the UN Office for Disaster Risk Reduction that considers two major global risks often excluded by national risk assessments and notes that national risk assessments have failed to adequately account for them.

Important problems of national risk assessments can be summarised as follows:  

National risk assessments, even those with the best intentions, omit critical risks. This is evidenced in the omission of volcanic eruption as a risk in the UK National Risk Register prior to the immensely disruptive eruption of Eyjafjallajökull in Iceland in 2010, and also the assessment prior to the Covid-19 pandemic that a non-influenza emerging disease outbreak might kill ‘up to 100 people’ in the UK. Any assessment of national risks needs to be comprehensive. Including diverse known risks allows for comparison of risks and prioritisation of resources. Decision makers who exclude risks from the assessment are exposing countries to increased risk and are circumventing a democratic discussion of risk prioritisation.

National risk assessments often omit the entire class of global catastrophic risks (GCRs). Assessments thereby plausibly omit analysis of most of the actual risk. GCRs include anything that would impact the entire world, create severe systemic harm and/or kill a significant proportion of the global population. These risks include: engineered pandemics, risks from artificial intelligence, nuclear war/nuclear winter, major volcanic eruptions, rapid severe climate change, severe solar storms, asteroid/comet impacts, etc.

National risk assessments seldom engage the appropriate spectrum of stakeholders. This is evidenced, for example, in the Swiss national risk assessment (2020) in which members of the public accounted for 0% of stakeholders engaged in risk workshopping processes, and the NZ National Risk Register which was developed but then kept confidential. Decision makers need to analyse and address risks that an informed public is concerned about, and risk information must be freely accessible in order for all stakeholders to be able to prepare to mitigate the risks. Indeed, the UN has noted that lack of access to risk information is a critical weakness of present disaster risk reduction activities.

National risk assessments have produced highly variable assessments of cross-border risks. For example, the probability of impact from a major volcanic event in Europe was 1:4 to 1:20 per annum in the UK assessment, but 1:70,000 in the Swiss assessment. Neither assessment included the 1:625 likelihood of an even larger eruption elsewhere in the world that could cause massive disruption to global trade. This implies that working with other countries to share analyses and align expert findings would be valuable.

National risk assessment can become politicised or focused on recent salient events to the exclusion of major likely harms. It is well known that risk assessment processes can end up being manipulated for political gains, or to help consolidate the status quo or protect existing budgets. Resources often end up allocated to studying comparatively lesser risks in forensic detail, rather than addressing the low-hanging fruit for big inevitable ones. An example is plausibly the time and effort spent analysing domestic terrorism versus preparation for a catastrophic pandemic.

National risk assessments, and consequence-probability risk matrices, fail to appropriately represent the salience of risks. For example, categories are often used to represent the likelihood and consequences of hazards in a risk matrix. However, it is evident both that the categories chosen in many national risk assessments fail to adequately discriminate amongst risks, especially at the more serious end, and the salience of very severe risks such as global pandemics is not readily apparent. These shortcomings have been described by ourselves, and others.

Moving forward with NZ risk assessment

No one knows what the next catastrophe will look like, so preparation and resilience measures that cut across risks, and address systemic weaknesses (in trade, energy, transport, infrastructure, health, and plans to supply essential goods, etc) should be prioritised. There are many domains where investment in prevention offsets much greater impact costs.

Indeed, the biggest threats of undermining the SDGs globally, and of harming New Zealand specifically are global catastrophic risks. Yet, arguably, even with only short term (1 year) thinking, some GCRs appear to be the priority for mitigation measures as we have argued elsewhere.

We support Gluckman & Bardsley’s call for a comprehensive, public, apolitical, NZ national risk assessment. But the approach will need to address the key shortcomings we identified above.

Recommendations

• A systematic national risk assessment should be a legislated requirement, and it should include global catastrophic risks. The recent US Global Catastrophic Risk Management Act (2022) could be used as a template.
• The assessment needs to be cross-government and cross-sectoral and be undertaken by an anticipatory and apolitical entity. We have described how a well-resourced Parliamentary Commissioner could play this role (but other structures are possible, such as a Chief Risk Officer).
• The national risk assessment needs to engage experts and stakeholders widely. This could help overcome the problem of poor foundational assumptions, for example around such things as scenario choice, discount rate, time-horizon, and decision rule. We have explained these issues in depth elsewhere.
• Any national risk assessment needs to be connected to a capabilities analysis that assesses not just the consequences of a risk in expectation, but also the marginal benefit of various actions (eg CBA) and the value obtained from acting (eg equity or Treaty issues).
• Options and trade-offs need to be put to stakeholders for a national discussion. We need more than a list of ‘bad things’, we need an agreed robust strategy to reduce risk across time.