Search
  • Geoff Mulgan

COVID-19 inquiries: how should they be run?




Boris Johnson has announced a UK inquiry into COVID-19 to start in 2022, a parallel one is being planned in Scotland, and many more will emerge all over the world. But how should such inquiries be designed and run? What kind of inquiry can do most to mitigate or address the harms caused by the pandemic?

We’re beginning to look at this question at IPPO (the International Public Policy Observatory), including a global scan with our partners, INGSA and the Blavatnik School of Government, on how inquiries are being developed around the world, plus engagement with governments and parliaments across the UK.

It’s highly likely that the most traditional models of inquiries will be adopted – just because that’s what people at the top are used to, or because they look politically expedient. But we think it would be good to look at the options and to encourage some creativity.

The pandemic has prompted extraordinary innovation; there is no reason why inquiries should be devoid of any. Moreover, the pandemic affected every sector of life – and was far more ‘systemic’ than the kinds of issue or event addressed by typical inquiries in the past. That should be reflected in how lessons are learned.

So here are some initial thoughts on what the defaults look like, why they are likely to be inadequate, and what some alternatives might be. This article proposes the idea of a ‘whole of society’ inquiry model which has a distributed rather than centralised structure, which focuses on learning more than blame, and which can connect the thousands of organisations that have had to make so many difficult decisions throughout the crisis, and also the lived experiences of public and frontline staff. We hope that it will prompt responses, including better ideas about what kinds of inquiry will serve us best.

A brief history of UK inquiries There is a long tradition of inquiries after big public disasters or crises. In the UK, these have included: the Chilcot report into the Iraq War, the Hutton inquiry prompted by the death of Dr David Kelly, the Grenfell Tower inquiry, Lord Leveson’s inquiry on abuses by the media, and the Bloody Sunday inquiry. There are also many other types of inquiry: parliamentary, royal commissions, audits that try to get to the truth, as well as internal civil service inquiries. And there are more localised inquiries – such as the various ones conducted on individual hospitals after crises such as the Mid-Staffordshire hospital deaths. There are also many parliamentary inquiries – one estimate suggests that there have already been around 60 COVID-related inquiries carried out by different select committees.

Our political system is clearly fond of inquiries, and typically the first calls made by the media and opposition parties when something goes wrong is that there should be one. Some of these are governed by UK Government legislation passed in 2005, which tried to standardise their formats. A key aim was to ensure that inquiries would be more trusted – yet inquiries continue to be denounced as ‘inside jobs’ and ‘establishment stitch-ups’, and this will be a key challenge for any inquiry into COVID-19. The problem isn’t helped by the tendency of government to appoint friends to run its own inquiries – which then usually conclude that no serious mistakes were made.

The standard inquiry model

The UK’s more formal and serious inquiries generally aim to find out:

  • what happened;

  • why it happened;

  • who is to blame; and

  • how to prevent it happening again.

They take their models from the law and courtrooms, with witnesses, cross-examinations and written judgments. They happen in a physical place, usually in central London. As in a courtroom, they aim to establish the key facts, determine guilt, and then have an additional aim of recommending new rules or laws to prevent mistakes being repeated. They are usually created by the executive but have a degree of independence, and are usually advisory. Their results can then reframe how an issue is seen; can lead to new rules, laws and institutions being introduced; and can prompt new procedures – for example, for professions such as the police or social workers.

The default for any COVID inquiries will be very similar. There will be fierce political arguments about the precise terms of reference. But as in the past, a leading establishment figure with a legal or governmental background will be put in charge. Written and oral evidence will be taken. A series of reports will be produced at some point in the future. Even within this standard model, there are many design choices to be made. How fast or slow should the inquiry be? – with obvious trade-offs between their speed and depth, and cost (the Bloody Sunday inquiry took 12 years and reportedly cost more than £200m). How much should the inquiry formalise blame, even recommending prosecutions? (i.e. will anyone lose their job, go to prison?) Should it have powers to award damages?

We can expect many key groups to be nervous about the prospect of such an inquiry, and to wish to shape it to protect themselves: those politicians who made a series of misjudgments, particularly in the autumn of 2020; scientists whose advice at various points, particularly early on, may have been erroneous; officials whose machineries for crisis management were shown to be very uneven at best. Many people will be working hard to establish narratives and explanations to protect their reputations. In general, governments will probably want pandemic inquiries that stretch as far as possible into the future. In contrast, opposition parties will probably want short, sharp inquiries with conclusions ahead of the next election. However, neither may actually serve the public interest very well.

The limitations of hierarchical models So the traditional model of inquiry is a highly centralised, formalised and legalistic approach based on prose. In short, a hierarchical exercise designed for hierarchies – apportioning blame to some of the people in charge. Clearly it’s right to do this in hierarchical systems such as governments; many crucial decisions are taken at the top of big structures. It’s also right to want inquiries to be independent.

This approach satisfies a deep human need for explanation and justice – if something has gone wrong, we want to see who is to blame and to see them shamed or punished (in the ancient past, kings who lost big battles were sometimes killed). Westminster-style parliamentary systems are particularly keen on ‘sacrificial accountability’, while the US has its own style of often quite partisan inquiries.

But there are also serious limitations with this approach. It is essentially a model for the state to ask questions about itself – and in many ways, is less democratic both in practice and in spirit than the courts system, which has juries to represent the public. It generates big incentives to distort or divert. It can cut against the kinds of honesty and self-awareness that are vital for learning. It can become a kind of theatre, encouraging performance rather than understanding.

Moreover, it’s not a particularly good way to change how a complex system works. Just as criminal courts are mainly designed to establish guilt, not promote learning, reconciliation or rehabilitation, so the same is true of court-like public inquiries. The centralised model may sometimes be quite good at handling the technical aspects of an inquiry – were hospitals adequately prepared? what went wrong with procurement? – but they risk failing in another goal: giving people a chance to express their pain and grief, and to get answers from the powerful.

This has been an important aspect of some inquiries – such as Grenfell – and is a vital part of post-conflict inquiries. Without the catharsis that comes from hearing testimonies from experience, it’s very hard to rebuild trust.

Possible alternatives: ‘whole of society’ approaches

There are many different options for inquiries, and this is a good moment to consider them. They range from ‘truth and reconciliation’ inquiries to no-fault compensation processes to the ways industries such as airlines deal with crashes, through to academic analyses of events like the 2007/08 financial crash. They can involve representative or random samples of the public (e.g. citizens’ assemblies and juries) or just experts and officials.

Over the next few weeks, with colleagues around the world, we will be developing a more comprehensive taxonomy of inquiries, and hopefully clarifying their pros and cons. Here are some of the dimensions on which different choices can be made.

1. Single or multiple? A first step is whether to think of this as a single inquiry, or multiple inquiries. Unlike a courtroom judging guilt on a particular action, countries will have to make sense of many different levels of governance and decision-making, from the WHO and Whitehall down to local councils and businesses. And they will involve many different sectors or ‘verticals’, from retail to education. So the units of inquiry will be much more like the cells of a matrix than a single question or process. These could in theory all be run separately – but it would make more sense to link them up.

2. Staged? The traditional inquiry spends a lot of time gathering evidence, before the chair goes away to write a report. A better model might be to organise the inquiry with distinct stages:

  • A stage to gather key facts and data;

  • A stage to generate key hypotheses (e.g. the hypothesis that the UK should have locked down sooner on several occasions; or that it needed greater institutional capacity – for example, to integrate scientific, economic and other factors at the heart of government, or to coordinate national and local governance);

  • A stage to explore these hypotheses with evidence and witnesses;

  • A stage to look at blame or individual responsibility; and

  • A stage to look at structural recommendations.

There are many other options – but the classic inquiry model can be too slow, too disorganised and too much at risk of going around in circles.

3. People or state? Another important question is whether the public could be in the driving seat – in other words, can they be thought of as the primary client or audience for an inquiry, rather than it being an internal exercise for different arms of the state? A more democratic model suggests approaches to evidence that give weight to people’s experiences and harms: what mattered to them, not just to government. It suggests a direct role for members of the public in the process – and could involve a role for them in the judgment process too.

4. What kinds of evidence? The traditional inquiry model depends on witness statements from high-status individual experts, a method long superseded in many fields. An alternative approach seeks out multiple kinds of data and knowledge, and asks what would best help the relevant systems to learn the right lessons. It includes formal evidence synthesis of the kind that IPPO does, along with systematic learning from practitioners and professionals to tap into the kind of tacit knowledge that may not be apparent in formal evidence. This may be particularly relevant in a complex story like that of COVID-19, with many layers (infection and epidemiology; economics; social dynamics; politics) and a lot of uncertainty.

The following list of knowledge types – all likely to be relevant to pandemic inquiries – shows the variety of potentially relevant evidence:

  • Statistical knowledge (e.g. of unemployment rises in the crisis);

  • Policy knowledge (e.g. on what works in stimulus packages);

  • Scientific knowledge (e.g. antibody testing);

  • Professional knowledge (e.g. on treatment options);

  • Public opinion (e.g. quantitative poll data and qualitative data);

  • Practitioner views and insights (e.g. police experience in handling breaches of the new rules);

  • Political knowledge (e.g. on when parliament might revolt);

  • Economic knowledge (e.g. on which sectors are likely to contract most);

  • ‘Classic’ intelligence (e.g. on how global organised crime might be exploiting the crisis).

5. The right questions? Any inquiry is defined by the questions that are thought to be important. Some are basic questions of competence and process (what right and wrong decisions were made about lockdowns?). But the crisis has also thrown up many more fundamental questions that would benefit from systematic attention – for example:

What has the crisis shown us about loneliness and isolation, and how these might be better addressed in future? How can we create better machineries for coordination between the UK government, devolved administrations and local government? How should science advice be organised in future, particularly to make more use of social sciences that were fairly marginal within SAGE? How can we help the residential care system improve its own use of knowledge and data – which is currently far behind the health system? Or, given that we now know that the ‘stay home’ message fuelled a spike in domestic violence, what might we do differently in a future pandemic?

It soon becomes obvious that answering these very varied questions requires very different styles of inquiry. 6. Requisite skills? A single chair? Who is best-placed to preside over an inquiry? The answer depends, to some extent, on its method, questions and purpose. While the traditional answer is to depend on a single figure, usually from a legal background, it is not obvious that a judge, however eminent, would have the right skills to understand deeper structural background issues that may have contributed to the UK’s problems during the pandemic.

Judges are not well-placed to reach conclusions on the questions listed above, though they have the virtue of being independent – and being seen to be independent. At a minimum, a small group with a wide range of expertise spanning science, administrative systems and technology could be necessary; for example, a three- to five-person commission of inquiry with a judge as chair.

7. Written reports? The classic public inquiry ends with a report – usually a long one. In terms of output, it’s not obvious that a large written report is a good way to help a system learn, whether in terms of sense-making or impacts. Indeed, there is vast evidence to suggest that an inquiry is unlikely to be effective as a way to get lessons learned unless these can be easily translated into rules and laws. Instead, findings need to be tailored to multiple audiences, then organised in ways that help with their digestion, much of which is likely to happen through conversation and events rather than prose.

8. Timing? Some inquiries take years or even decades. Whether an inquiry takes a traditional or novel form, there are good reasons for beginning it after the crisis has passed but while it is still fresh in the memory – a maximum 12-18 months after the crisis has abated – with conclusions reached quickly enough to be relevant to action.

9. Commissioned by whom? Any inquiry needs to be set up and commissioned by an institution. That could be the executive, or it could be parliament. Most inquiries – whether statutory or non-statutory – are set up by a government minister who also appoints the chair, which clearly brings with it major problems.

10. A courtroom or a platform? Any inquiry in the 2020s should, as a matter of course, use multiple (digital) platforms to link its key elements – gathering evidence and experiences, doing semantic analysis of key patterns, and supporting deliberation. Unfortunately, this is not a competence of the judicial system, nor an approach it is culturally at ease with. But it would be strange to operate in an essentially analogue way after a crisis that has shown just how much other fields can be transformed by digital tools.

Similarly, this would be an ideal opportunity for the BBC to demonstrate its public service remit, not simply through individual programmes but through a more systematic production of materials – online, TV and radio – to support the public in understanding the pandemic.

How a ‘whole of society’ inquiry model could work Alternative inquiry models could be more distributed and decentralised, creating spaces for many institutions to think about and absorb the key lessons. We could think of this as the COVID-19 Inquiries Programme: a ‘whole of society’ approach that would run more like a matrix of linked inquiries. These might cover:

  • The Prime Minister and the ‘centre’;

  • Other UK Government departments;

  • Schools;

  • Hospitals;

  • Local authorities;

  • Central risk management capability in government;

  • Public health;

  • Social security;

  • Parliament;

  • Police;

  • Science advice;

  • Universities.

The aim might be to ensure evidence-based deliberation on the key issues and lessons for each institution. So, for example, the lessons for schools are likely to be very different from those for the police or Whitehall, and they are most likely to be accepted by teachers if people they respect – including other teachers – are closely involved. The design challenge is then to ensure there are proportionate inquiries within each sector or field, but with shared questions, facts and evidence.

Common elements for ‘whole of society’ inquiries The idea of a distributed inquiry is not entirely new. Colombia, for example, attempted something along these lines as part of its peace process. Many health systems use methods such as ‘collaboratives’ to organise accelerated learning. Doubtless there is much to be learned from these and other examples. For the UK in particular, it is vital there are contextually appropriate designs for the four nations as well as individual cities and regions.

As already indicated, a key is to combine sensible inquiries focused on particular sectors (e.g. what did universities do, what worked…) and make connections between them. As IPPO’s work on COVID inequalities has highlighted, the patterns are very complex but involved a huge amount of harm – captured in our ‘inequalities matrix’, below.



So, while the inquiries need to dig deep on multiple fronts and to look more like a matrix than a single question, what might connect all the inquiries would be a commitment to some common elements which would be shared:

  • Facts: In each case, a precondition for learning is establishing the facts, as well as the evidence on what did or didn’t work well. This is a process closer to what evidence intermediary organisations – such as the UK’s What Works Network – do than a judicial process designed for binary judgments (guilty/not guilty). This would be helped by some systematic curation and organisation of the evidence in easily accessible forms, of the kind that IPPO is doing.

  • Experience: Thousands, perhaps millions, of people experienced the pressures and uncertainties of making hugely difficult decisions during the pandemic. Again, we need well-curated platforms that can gather these experiences and spot patterns. Relying on traditional written and oral evidence is a very antiquated and inefficient method for doing this.

  • Engagement: There is a need for conversation at institutional level that combines formal evidence with experience, best done through review meetings that are close to the ground. Here, some of the methods used by health improvement collaboratives or study circles in education are probably most relevant.

  • Interpretation: Another precondition is ensuring interpretation and engagement from key influencers and shapers in each sector. This involves bringing them together into group conversations which are ideally confidential – the precise opposite of the cross-questioning of individuals in a courtroom equivalent.

The draft diagram, below, suggests a way to consider the options over two key dimensions: one concerning how centralised or decentralised the inquiry is; the other how much the goal is blame or learning.



These are not binary choices, however. More decentralised exercises could then feed back into a more traditional and formal central inquiry, helping it to establish those basic issues: what happened; what could have been done differently; how do we ensure mistakes aren’t repeated? Similarly, while such exercises tend to work best when they are separated from judgment and blame, they can feed into more formal processes investigating culpability.

Risk of going MAD?

Inquiries are part of accountability. However, you can have too much as well as too little accountability. Academics speak of ‘MAD’ – multiple accountabilities disorder – when so much inquiring and asking gets in the way of doing the job and achieving other goals, such as expression and catharsis, as well as learning.

The excessive demands of accountability may go upwards – as when higher tiers in hierarchies demand reporting and data without a sense of the opportunity cost. Or it may point outwards, when too much is asked of organisations in terms of public reporting. It’s plausible to predict there may be a bout of enhanced accountability after the crisis. But this may or may not be constructive, since accountability is not the same as learning. So the design of better inquiries may also need to address how we achieve the right balance of accountability and learning.

An initial conclusion The COVID-19 crisis was uniquely wide-ranging and systemic. It needs a comparable model of inquiry. Indeed, what we need is a comparable ecosystem of inquiries, combining some classic top down versions with a much richer bottom up version. The ‘whole of society’ approach recognizes the need to hear from everyone, and to explore the complex dynamics of the pandemic at many levels and in many organisations. The key test of success is then:

  • A clear narrative and explanation of what went right and what went wrong at multiple levels and in multiple fields

  • A clear set of conclusions for institutions ensuring they have learned the key lessons and have embedded lessons in their operations.

At the very least let’s have a discussion. It’s highly likely the most traditional model will be adopted just because it’s what people at the top are used to.

But that would be a huge wasted opportunity and would probably leave behind more distrust and suspicion rather than less.

Sir Geoff Mulgan is Professor of Collective Intelligence, Public Policy and Social Innovation at UCL STEaPP, and IPPO’s Co-investigator. If you would like to contribute to IPPO’s thinking on this topic, please email us at ippo@ucl.ac.uk using the subject line Inquiries


Author’s note My thanks to Professor Matthew Flinders of the University of Sheffield for his thoughts on this topic, to Dr Alastair Stark of the University of Queensland for his work analysing inquiries, and to colleagues at IPPO who will be working on the topic over the next few months, including Nick Bibby, Kristiann Allen, Joanna Chataway, Thomas Hale and Mike Herd.

https://covidandsociety.com/contemplating-covid-crisis-what-inquiry-learn-right-lessons/