Greggs Plc v Zurich Insurance Plc

This Judgment relates to the expedited trial of certain Preliminary Issues arising in the Claimant’s (‘Greggs’’) claim against the Defendant (‘Zurich’) for an indemnity under an insurance policy with Number LB903381 (‘the Policy’), on the ‘Marsh Resilience’ Form, which insured Greggs, amongst other things, against ‘Business Interruption – Specified Causes’. Greggs claims in respect of business interruption loss and related losses / costs caused by Covid-19.

The hearing in this case has been coordinated with the hearing of early or preliminary issues in two other cases, Stonegate Pub Co Ltd v MS Amlin Corporate Member Ltd and Others (CL-2021-000161) and Various Eateries Trading Ltd v Allianz Insurance Plc (CL-2021-000396). Originally the hearing in this case was due to take place second, but as a result of one of the counsel in this case contracting Covid-19, it was heard third. The issues overlap with those in the other two cases, and this was the reason why the hearings were coordinated. The three actions remain separate, however, and there are also some significant differences in the arguments being advanced. This Judgment will accordingly deal with the agreed Preliminary Issues in the Greggs action, but it will, as has always been envisaged by the parties and the court, make reference to the Judgments in the other actions.

Greggs is a UK food-on-the-go retailer, offering a number of products including sandwiches, savouries, confectionary, coffee, breakfast and evening menu items on both a retail and wholesale basis. Greggs employs some 25,000 people. It pleads that within the Insured’s Business as defined in the Policy were 1778 stores in England, of which 312 were franchise stores; 157 stores in Wales, of which 12 were franchise stores; 279 stores in Scotland, of which 9 were franchise stores; and 21 stores in Northern Ireland, of which 5 were franchise stores.

Greggs claims that each of these stores suffered, to a greater or lesser extent, interruption or interference as a result, to put the matter broadly, of Covid-19 and its consequences, and has made a claim on Zurich estimated to be in excess of £150 million.

Greggs announced on 23 March 2020 that it would be closing all its shops in order to protect its customers and staff. It proceeded to close its shops from about 25 March 2020 until late May 2020, suffering business interruption loss. There was then a phased re-opening from May 2020, but Greggs contends that further restrictions imposed by government in one or more of the four nations and/or the effect of the disease in the UK led to further business interruption losses. It claims that it can recover these losses under the Policy.

Zurich contends that all of Greggs’ business interruption loss arises from, is attributable to, or is in connection with a ‘single occurrence’ and is to be aggregated as one Single Business Interruption Loss (‘SBIL’) subject to a limit of liability of £2.5 million, and that Zurich has discharged its obligation to indemnify Greggs by paying a sum of £2.5 million to Greggs on or around 19 January 2021.

The Policy under which Greggs was insured consisted of a Policy wording and a Schedule. The relevant provisions of the wording appear at Annexe [2] to this Judgment. The relevant provisions of the Schedule to the Policy appear at Annexe [1] to this Judgment.

The Period of Insurance of the Policy was 1 January 2020 to 31 December 2020. This may be compared with the Period of Insurance in Stonegate v MS Amlin which ended on 30 April 2020, and that in Various Eateries v Allianz which ended in September 2020.

The Preliminary Issues, which the parties agreed, and which have been ordered to be tried, are set out in Annexe [3] to this Judgment. In broad terms, those issues can be divided as follows:

It is agreed between the parties that all questions of the existence, timing, extent, continuity, causation and quantum of any loss(es) Greggs may have suffered will not be decided as part of these Preliminary Issues, but will fall to be determined (if necessary) at a later stage.

For the purposes of these Preliminary Issues, the parties have agreed various Statements of Agreed and/or Assumed Facts. They have agreed a Statement of Agreed Facts as to the nature and origins of Covid-19 and its spread and transmission; another as to the chronology of the Government Response to the pandemic; another as to the UK emergency framework and scientific advisory structure; and a Statement of Agreed and Assumed Facts as to Government Support received by Greggs.

The parties also served expert reports from three experts, relating to virology and epidemiology. For Greggs, from Professor Joel O. Wertheim, an Associate Professor of Infectious Diseases and Global Public Health at the University of California San Diego, who gave evidence in relation to virology and epidemiology. For Zurich, reports were served from Professor Julian Hiscox, who holds a Chair in Infection and Global Health at the University of Liverpool, on virology; and from Professor Matthew Baylis, an expert in infectious disease epidemiology.

This expert evidence was adduced in relation to the Aggregation Issue. There was a great deal of common ground between the experts in their reports. Ultimately the parties agreed between themselves that it was not necessary for the experts to be called or cross-examined.

The issues which arise are to a large degree questions of construction of the Policy. I have considered the applicable principles in my Judgment in Stonegate v MS Amlin, and I adopt what is there said.

As formulated, these Issues are lengthy. But there was, in fact, limited difference between the parties as to them.

I will commence by considering the ‘Covered Events’, and the remaining issues as to what the relevant Covered Events were and how many, and then proceed to consider the contentious issue of what constituted the ‘triggers’ of cover.

There was, as I understood it, no dispute that for the purposes of what was called the ‘Disease Peril’, there was a Covered Event under Clause 2.3(viii) each time there was the occurrence of a case of Covid-19 in the Vicinity (which it was accepted by Zurich was nationwide) during the Period of Insurance.

In relation to what was called the ‘Prevention of Access Peril’, there was no dispute that there was a Covered Event under Clause 2.3(xii) each time actions or advice of a governmental authority in the Vicinity of an Insured Location prevented or hindered the use of or access to an Insured Location during the Period of Insurance. There remained issues as to whether: (i) there was a Covered Event for each Insured Location affected by government actions or advice, or whether government actions or advice affecting multiple Insured Locations constituted one Covered Event; (ii) there was a Covered Event in each case where government actions or advice continued, renewed or replaced actions or advice which already hindered use of or access to an Insured Location during the Period of Insurance; (iii) there was only a single Covered Event where government announcements or measures were part of a ‘coordinated governmental response’; and (iv) each of the announcements and measures pleaded by Greggs was capable of having caused a prevention or hindrance. I will address these issues in paragraphs [22-27] below.

In relation to what was called the ‘Enforced Closure Peril’, there was no dispute that there was cover under Clause 2.3(viii) taken with Definition 69(v) when there was the Enforced Closure of all or part of an Insured Location by a governmental authority or agency. I did not understand Zurich to pursue a case that there could only be an Enforced Closure if the whole of an Insured Location was prevented from opening (which appears to be raised by Issue 3.1), and had such a case been made I would have rejected it. It would have been inconsistent with the decision of the Divisional Court in the FCA Test Case at para. [303]. Consistently with this, I also understood there to be no dispute that there would be a Covered Event if Greggs was required by law or government action to close the indoor and outdoor seating areas at an Insured Location even though it could continue to carry on take-away service (Issue 3.2).

In relation to Issue 3.3, Greggs accepted in advance of the hearing that when it had decided to close all its shops completely in circumstances where government measures in force required it only to close the eat-in section of its shops and permitted Greggs’ shops to remain open for take-away business, that was not an Enforced Closure in relation to the parts of Greggs’ shops that were permitted to remain open for such take-away business. My understanding was that that concession must also be relevant in relation to Issue 3.4.1, in that if premises were used solely for take-away services prior to 23 March 2020, then the measures introduced in March 2020 did not enforce the closure of that business. For its part, Zurich accepted that there was a partial Enforced Closure in respect of stores that had been converted to provide a take-away service only after Covid-19 had entered the Vicinity, and thus that Issue 3.4.2 no longer arose.

There remained issues as to whether: (i) there was a Covered Event in respect of each Insured Location where there was an Enforced Closure, or of all Insured Locations together (which is contained within Issue 3.5); (ii) there was a new Covered Event each time a government announcement or measure that resulted in an Enforced Closure was continued, renewed or replaced by another announcement or measure (which is contained within Issue 3.6); (iii) there was only one Covered Event where government announcements or measures were part of ‘a coordinated governmental response’ (Issue 3.7); and (iv) each of the matters relied upon by Greggs constituted an Enforced Closure. I will deal with these issues in paragraphs [28-32] below.

I have identified the issues remaining in paragraph [18] above.

As to the first of those issues, I have expressed in paragraphs [71-72] of my Judgment in Stonegate v MS Amlin the reasons why I consider that there will be the same number of Covered Events as there are relevant advices / actions, and not that number multiplied by the number of Insured Locations affected.

As to the second, in keeping with what I have said in paragraph [73] of my Judgment in Stonegate v MS Amlin, I consider that the number of Covered Events will be the number of occasions on which there were materially different restrictions imposed or advised by government or a relevant agency which prevented or hindered the use or access to ‘Insured Locations’ and that steps taken or advice given which merely repeated or renewed an existing prevention or hindrance of access would form part of one set of ‘actions or advice’.

As to the third, I see no warrant in the language of the Policy for saying that materially different restrictions imposed or advised by government or a relevant agency will constitute only one Covered Event if made in pursuance of a ‘coordinated governmental response’. It is not therefore necessary at this point to consider whether there was a ‘coordinated governmental response’. For reasons which I will give in more detail in the context of aggregation, I reject the case that there was such a response which applied during the whole of the Period of Insurance.

I do, however, accept Zurich’s alternative case on this point, namely that the announcement of a given measure and regulations subsequently giving effect to it will (at least usually) have constituted a single Covered Event, as being ‘actions or advice’ which caused the same prevention or hindrance.

As to the fourth, I was not asked by either party, at this juncture, to go through each of the 120 or so announcements and measures pleaded by Greggs to say whether it constituted a Covered Event. If this remains an issue of significance in the light of this Judgment, there can be further argument on this in due course.

I have identified the issues in paragraph [21] above.

As to the first of those issues, I have considered this in paragraphs [67-68] of my Judgment in Stonegate v MS Amlin. Given the terms of this insuring provision, I consider that the relevant Covered Event is the actual closure of all or part of an Insured Location under relevant compulsion or instruction. The number of Covered Events for this Peril is accordingly the number of Insured Locations so closed (unless such Insured Locations opened and were then closed again in which case there will have been more Covered Events).

As to the second, I have considered a similar issue in paragraph [69] of my Judgment in Stonegate v MS Amlin. I consider that it is not correct to say that there is a separate Covered Event where the closure was enforced by the reiteration, continuation or renewal of regulations which were, materially, to the same effect. The Covered Event is the Enforced Closure and there will be one Covered Event unless the Location (or the part of it which was closed) opens and is then closed again.

As to the third, and as with the equivalent argument in relation to the Prevention of Access Peril, I see no warrant in the Policy for considering that, if there were several Enforced Closures which were separate in accordance with what I have said in the previous paragraph, they should nevertheless be regarded as one Covered Event if made in pursuance of a ‘coordinated governmental response’. That is a concept of somewhat uncertain scope and of difficult application, and is not referred to in the Policy.

As to the fourth, I consider that an identification of what exactly were the Enforced Closures which constituted separate Covered Events is a matter which should be resolved subsequently, if and insofar as it is necessary, in the light of this Judgment.

The other aspect of Issues 1-3, as they are drafted, concerns what ‘triggers’ a right to indemnity. It is Zurich’s case in this action, at least in relation to the Disease Peril, that the Policy is ‘triggered’ by interruption of or interference to Gregg’s business as whole, resulting from a Covered Event, and it is not ‘triggered’ by the underlying Covered Event itself. There was only one loss because there was only one period of interruption or interference, or, if there were distinct periods of interference, then there were the number of losses that there were discrete periods of interference. Zurich acknowledged that the number of such periods of interference, and the number of losses were not matters to be determined as part of these Preliminary Issues.

Zurich accepted, as I understood it, that for the purposes of Issue 2.1 the distinction between a Covered Event and what it called a ‘trigger’ was not material because there would only be a Prevention of Access if there was a prevention or hindrance of access or use. Further Zurich accepted that a discrete prevention or hindrance would in principle give rise to discrete interruption or interference. I also understood Zurich’s position to be the same, mutatis mutandis, in relation to the Enforced Closure Peril; certainly, I did not see why it should be different.

Accordingly, the main significance of this part of Zurich’s case related to the Disease Peril. Its case was that there were not as many ‘triggers’ as there were cases of the disease in the Vicinity during the Period of Insurance; there was instead only one ‘trigger’ (or at most a handful of ‘triggers’), because there was only one (or a handful) of interruption(s) or interference(s) caused by the covered cases of the disease. On the basis of this argument Zurich contended that there was only one loss (or a handful of distinct losses), as a result of one interruption giving rise to one reduction in Turnover, and therefore only one Limit was applicable, namely £2.5 million.

Greggs noted that in the Stonegate v MS Amlin action, Stonegate Insurers, who included Zurich, had contended that issues as to ‘triggers’ had been included at the insistence of the insured, and that Stonegate Insurers had contended that they did not matter; but here it was Zurich which was seeking to rely on such arguments to pre-empt the issue of aggregation. Greggs contended that the case was based on a ‘manipulation’ of the concept of ‘triggers’, was entirely artificial and had no basis in the Policy terms. It was Greggs’ case that counting ‘triggers’ did not matter at all.

The starting point is to observe that the use of the term ‘trigger’, as I have said in my Judgment in Stonegate v MS Amlin, is apt to confuse. It is, indeed, used in a different sense by Zurich from that in which it was used by Stonegate. Zurich’s contention here is that the Policy will be ‘triggered’ only if there is a right to indemnity under it. Zurich’s argument is that, even giving full effect to what was decided by the Supreme Court in the FCA Test Case at paragraph 215, namely that business interruption is a description of the type of loss or damage covered by the Policy, there nevertheless needs to have been such loss or damage for there to be a claim under an indemnity policy, and so there will be no claim under the Policy unless there is such business interruption or interference. That much of the argument I accept.

I do not, however, accept the later stages in the argument. In the first place, considering the Disease Peril, I do not consider that it is correct to say that there will have been only one interruption or interference to the Insured’s Business as a result of the very many Covered Events. While each case of Covid-19 in the Vicinity is to be regarded, in accordance with the reasoning of the Supreme Court in the FCA Test Case, as an equally efficient cause of the governmental and public response, that is a recognition of the fact that it was because there were very many cases that the government was caused to act. I consider that each of those cases must be regarded as causing its own interruption or interference, although of course, because they acted in the same way, the resulting periods of interruption or interference may have been coterminous. I therefore do not consider that there was a single or ‘unitary’ interruption or interference, and thus a single ‘trigger’ in Zurich’s terms, even in relation to periods in which all shops were shut, when this was the result of multiple Covered Events.

Even if that is wrong, and while I am not called on or in a position to say how many periods of interruption or interference there actually were, it appears highly doubtful that, on any view, there can be said to have been only one period of interruption or interference. It seems certain that the degree of interruption or interference with Greggs’ business changed over time between the first cases of Covid-19 and the end of the Indemnity Period; and that, as Greggs put it, ‘the position as to interruption and interference was variegated both temporally and (at most times) also geographically.’

Nor do I accept that Zurich’s argument is supported by the way in which BIL is defined. To recap, BIL is defined, in part, as Reduction in Turnover. Reduction in Turnover is calculated by taking the Turnover of the Insured’s Business and comparing it with the Standard Turnover of the Insured’s Business as a whole. (The Insured’s Business is, for present purposes, Greggs’ Business, as Greggs Foundation is not relevant, and the exception in relation to business conducted in departments was also not said to be relevant.) This way of defining BIL does not, however, indicate that there must be considered to be a single loss for the purposes of the Limits provisions. Such an approach would be inconsistent with the fact that there is a distinction drawn between BIL and SBIL. The existence of the latter term and concept indicates that there can be a plurality of losses which fall to be aggregated as one SBIL. Indeed, the logic of Zurich’s argument in this respect would appear to be that there was one loss (and one SBIL Limit) even if there were two wholly distinct Covered Events at different shops, because the disruption to each will be reflected in the overall Reduction in Turnover of the Insured’s Business. I regard that as clearly incorrect.

More generally, I accept Greggs’ submission that the correct approach to whether its claim in respect of its Covid-19 losses is limited is in the following four stages: (i) If Greggs has incurred BIL (including Reduction in Turnover, as defined), (ii) and if that BIL has been proximately caused by interruption or interference with the Business as a result of one or more Covered Events under Clause 2.3, (iii) and if that loss has not been proximately caused by an excluded peril, (iv) then Zurich is obliged to indemnify Greggs for that loss, save insofar as one or more of the Limits of Liability or Sub-Limits place a limit on Zurich’s liability. I accept further that, prior to stage (iv) the Policy terms do not divide the Insured’s loss into a countable number of ‘losses’ and place limits on the Insured’s recovery by reference to that number of ‘losses’. Any limit derives from the application of the Limits of Liability, including the limit any one SBIL, which incorporates an aggregation provision.

The issue of aggregation is Item 4 in the List of Preliminary Issues. As in the Stonegate v MS Amlin and Various Eateries v Allianz actions, the issue is whether the alleged BIL can be said to arise from, be attributable to, or to be in connection with a ‘single occurrence’ so as to be aggregated as a SBIL in accordance with Definition 105 of the Policy. Issue 4 sets out the cases on each side as to what constituted a ‘single occurrence’.

It is necessary to record, at the outset, that at the hearing Greggs made clear that its primary case was that the relevant ‘single occurrences’ were the various government announcements and measures (ie the case referred to in Issue 4.1.3), and that its cases that the relevant ‘single occurrences’ were each occurrence of Covid-19 discovered at an Insured Location or occurring within the Vicinity of an Insured Location, or all occurrences of Covid-19 within the Vicinity of each Insured Location or discoveries of Covid-19 at each Insured Location affecting that Insured Location (ie the cases referred to in Issues 4.1.1 and 4.1.2), were alternative cases. On these Greggs said that it adopted what Stonegate had submitted in the Stonegate v MS Amlin action.

Equally, it should be recorded that Zurich made no arguments in favour of the ‘single occurrence’ being any one occurrence of Covid-19 in the Vicinity of an Insured Location, but not each and all of them (ie the case referred to in Issue 4.1.12).

As I have already said, the parties both agreed a number of facts, and served expert evidence, to form the basis on which the court should decide the issues in relation to aggregation.

As to the expert evidence, there was not only a great deal of common ground between the experts, but a very considerable degree of agreement with the evidence which was either agreed or which I have accepted in the Stonegate v MS Amlin and Various Eateries v Allianz actions.

In brief, the virology experts were in agreement that:

There were certain disagreements between the virology experts. In particular, Professor Hiscox expressed the view that a single infection of a single cell by the MRCA triggered the pandemic. Professor Wertheim did not agree with that statement, and was of the view that it is unlikely that the single MRCA infection was the trigger of the pandemic because it is not clear that it was in a human cell, and a single infected human would not be sufficient to give rise to a pandemic. Further, Professor Hiscox considered that there was an inevitability of the outbreak occurring, even if draconian measures had been taken. Professor Wertheim does not believe that the pandemic was inevitable in early January 2020 because various countries were able to stanch the spread of the earliest transmission clusters.

The epidemiology experts were in agreement that:

There were certain disagreements between the epidemiology experts. These included that:

As the experts were not called to give evidence, it would not have been possible for me fairly to resolve the differences between them. In the event, and for reasons which will be apparent from what follows, I did not consider that those differences were material to my decision in this case.

The arguments in relation to aggregation in this case, and in particular as to the proper approach to aggregation clauses, and the assistance to be gained from authority in relation to the words used in the definition of SBIL, overlapped very substantially with those presented in Stonegate v MS Amlin and in Various Eateries v Allianz. I considered the arguments adduced in this case (on which the Stonegate and the Various Eateries parties had the opportunity of commenting) before concluding my views in relation to those presented in the Stonegate v MS Amlin case itself. My reasoning in my Judgment in that case, in particular at paragraphs [78-90] and [101-116] is applicable in this case.

I should, however, specifically refer to one aspect of Mr Hancock KC’s helpful arguments on this aspect of the case which did not have a precise counterpart in the submissions in the earlier two cases. This was his analysis of various other usages in the Policy of the phrase ‘in connection with’. He referred in particular to Insuring Clauses 1.3 and 2.5, and in Definitions 32 and 34(iii). These provisions were not directly analogous to Definition 105, in that they use different language and have a different contractual context. They revealed the versatility of the linking words ‘in connection with’. I would, however, accept Mr Hancock’s submission that they all envisage a ‘straightforward and direct’ connection, while recognising the limits of the assistance this gives when considering the definition of SBIL. These usages do not, to my mind, suggest that the interpretation which I have attributed to the words as they appear in the SBIL definition, and which I have expressed in my Judgment in Stonegate v MS Amlin, is incorrect.

Against that background, I turn to consider the various candidates for ‘single occurrences’. I will consider first Zurich’s different cases, and then the cases made by Greggs.

As I have already set out, Zurich’s cases can be grouped under four headings. The first such heading is the emergence of SARS-CoV-2 and its sub-lineages. Under this heading are Zurich’s cases that Greggs’ BIL arose out of, is attributable to or in connection with: (i) the first coming into existence of the genomic sequence now called SARS-CoV-2; or (ii) the MRCA of all SARS-CoV-2 genomes; or (iii) the MRCA of all SARS-CoV-2 lineage B genomes; or (iv) the MRCA of all SARS-CoV-2 lineage B.1 genomes.

For the reasons I have given in relation to the ‘virology’ options in Stonegate v MS Amlin at paragraphs [158-163] of the Judgment in that case and the ‘virology arguments’ in Various Eateries v Allianz at paragraph [92] of the Judgment in that case, I would not regard any of these as being ‘single occurrences’ for the purposes of the SBIL definition. In summary:

I would accept that, if they, or any of them, count as a ‘single occurrence’ that it or they had a sufficient causal connection to Greggs’ losses, in the sense that those losses would not have occurred, at least in the way they did, but for such ‘occurrence’. But, as in the two previous cases, I am firmly of the view that, to the extent any of them counts as a ‘single occurrence’ it is too remote – geographically, temporally and causally – from Greggs’ losses to be a relevant ‘single occurrence’. This point is strongest in the present case, in that the aggregation for which Zurich contends extends to losses deriving from Covered Events all the way up to the end of December 2020.

The second heading under which Zurich’s aggregation options are grouped is as to the initial outbreak of Covid-19 in Wuhan in late 2019. Zurich contends that the ‘explosive’ outbreak of Covid-19 in late November or December 2019 constitutes a ‘single occurrence’.

Insofar as this case is that there was a ‘single occurrence’ in the zoonotic spillover(s) which probably occurred in the Huanan Market in Wuhan, I am prepared to accept that they can be regarded as an occurrence. Insofar as the case relates to the wider spread of the disease in Wuhan, I do not consider that this is a ‘single occurrence’, but a number of occurrences, occurring over a period of time, affecting many different individuals, at various places in Wuhan.

I would accept that the spillover(s), and indeed the outbreak in Wuhan, can be said to have had a causal link to Greggs’ losses, in that they would not have occurred, in the way that they did, without them. But I am, again, firmly of the view that each is too remote from those losses to count as a relevant ‘single occurrence’.

The third heading under which Zurich’s aggregation cases are grouped is the emergence of the global and UK pandemic. The case here is that there was a ‘single occurrence’ at the point when the number of cases and/or their spread in the UK or globally had reached such a level that an epidemic or pandemic had occurred or become inevitable or likely, and/or that a governmental response of the kind eventually made by the UK Government was inevitable or likely.

These are very similar to the ‘tipping point’ arguments which I considered in Stonegate v MS Amlin. I do not consider that the matters identified constituted ‘occurrences’. What are now relied on as ‘occurrences’ did not happen at any identifiable time or place. When they are regarded as having happened depends in part on what can be said to be an ‘epidemic’ or a ‘pandemic’, as to which there is no universally-agreed definition; and also on an assessment as to inevitability or likelihood, as to which there would have been (and still is) room for disagreement. Similar considerations apply in relation to governmental response. It is difficult to say when a particular governmental response became inevitable or likely. The fact that something has changed from avertable or unlikely to inevitable or likely is not something which can be said to happen at a particular place or in a particular way. It is rather a description of probabilities than an occurrence.

The fourth heading relevant to Zurich’s aggregation cases is the coordinated government response. It is Zurich’s case that in March 2020 the UK Government, in collaboration with the Scottish, Welsh and Northern Ireland governments, formulated and then began to implement a plan or strategy in response to the Covid-19 pandemic. This plan or strategy was then implemented in the succeeding months, and all governmental measures adopted by the four governments, through to 31 December 2020, were pursuant to that plan or strategy, and constituted a ‘single occurrence’.

Zurich put forward alternative cases to the effect that ‘elements of the co-ordinated government response’ for which it contended could be sub-divided. These alternative cases were that, if there was not one occurrence, by reason of one single coordinated response, then there were two (namely an initial one and another in the autumn and winter 2020), or four (being the response of each of the four nations, throughout the Period of Insurance), or five (being the initial coordinated response of the four nations and then one for each from about 16 October 2020 onwards), or eight (being two for each of the four nations, one in the initial period and the other in the period from about 16 October 2020 onwards).

In support of its primary case as to a single coordinated response Zurich relied, inter alia, on: a ‘battle plan’ announced by the DHSC on 1 March 2020, which included the statement that ‘All 4 parts of the UK will work together against the potential of a Covid-19 epidemic in the UK under the government’s new approach to pandemic preparation’; an Action Plan produced on 3 March 2020 by the UK Government, together with the Devolved Governments, which described its ‘fundamental objectives’ as the deployment of ‘phased actions to Contain, Delay, and Mitigate any outbreak’, noting that ‘the different phases, types and scale of actions depends upon how the course of the outbreak unfolds over time’; the Prime Minister’s Chairmanship of meetings of COBR in March 2020; the adoption in Scotland, Wales and Northern Ireland of measures similar to those announced for England by the Prime Minister on 16, 20 and 23 March 2020; on various statements by ministers in each of the governments that responses were being coordinated between the four administrations; the establishment of the Joint Biosecurity Centre in May-June 2020; and the adoption of similar measures in the latter part of the year.

I am unable to accept that there was a ‘single occurrence’ constituted by a coordinated government response from March to December 2020. What happened over that period of time was a series of increasingly divergent responses to different conditions as they presented themselves. It is true that in the early onset phase, in March and into April 2020, the four governments coordinated their policies, and the UK Government used intergovernmental structures (and in particular COBR and cross-UK Ministerial Implementation Groups (‘MIGs’)) to achieve this. The result was that the governments moved together in imposing the first lockdown. But from May 2020 the degree of coordination decreased, and the approach of the four nations increasingly diverged.

Thus, on 10 May 2020, the Prime Minister outlined how England would begin to ease restrictions, including by adopting a ‘Stay Alert’ rather than a ‘Stay at Home’ message, and by encouraging people who could not work from home to go to work. The devolved administrations did not follow that approach, and there were some complaints that there had been a lack of consultation. In Scotland, the First Minister informed Scottish residents that they should continue to ‘stay at home’. In Wales, the Deputy Minister for Economy and Transport tweeted that the position announced by the Prime Minister ‘is not the position in Wales … We don’t agree with the change of message’. The message also remained to stay at home in Northern Ireland.

Each of the four administrations published its own plans for easing lockdown restrictions, which differed to varying degrees in strategy and approach. On 11 May 2020 the UK Government set out a recovery strategy, which included three phases for easing lockdown restrictions in England; the Northern Ireland Executive set out five steps on 12 May 2020; the Scottish Government set out a route map through four phases on 21 May 2020; and the Welsh Government set out a traffic light system on 15 May 2020. This difference of approaches was commented upon in the House of Lords Select Committee on the Constitution, 3^rd Report of Session 2021-22 (10 June 2021) para. 103.

The position appears to be accurately summarised in that House of Lords Select Committee report (in paras. 100, 104):

During the period May-August 2020, there were substantive differences in the restrictions applicable in the four nations. For example, from 13 May 2020, two people from different households were allowed to meet outdoors in England; on 19 May 2020, up to six people who were not members of the same household were allowed to meet outdoors in Northern Ireland; meetings between two households were not permitted in Scotland until 29 May 2020; and in Wales meetings between two households were permitted only on 1 June 2020.

There remained differences of approach with the increase of Covid-19 cases in the UK in September 2020. England adopted a three (and later four) tier system; Scotland used five protection levels; Wales used four alert levels; and Northern Ireland, designated districts. On 16 October 2020, Northern Ireland imposed what was in effect a four-week lockdown by specifying various nation-wide restrictions during ‘active periods’. Wales instituted a ‘firebreak’ from 23 October 2020 to 9 November 2020, which required individuals to stay at home save for limited purposes. A second lockdown was announced in England on 31 October 2020, which took effect from 5 November 2020. The Scottish Government did not implement a national ‘circuit breaker’ or ‘firebreak’ during this period.

Thereafter, the UK Government reverted to the three-tier system, with modifications, on 2 December 2020. On 19 December 2020, the UK Government curtailed various of its planned relaxations of restrictions for Christmas; and on 20 December 2020 added a fourth tier. On 6 January 2021 the third national lockdown was introduced in England by every area of England being classified as Tier 4. In Scotland, on 19 December 2020 the First Minister announced a tightening of restrictions, including the application of Level 4 measures to all mainland Scotland for three weeks from Boxing Day. On 5 January 2021 a requirement to stay at home was implemented in all Level 4 areas, and further restrictions were implemented on 16 January. In Wales, on 16 December 2020, the First Minister announced that all of Wales would move to Alert Level 4. On 21 December 2020 the relevant regulations were amended so that there a relaxation for Christmas, but only for 25 December. Wales did not significantly alter its restrictions in January 2021. In Northern Ireland, it was not until midnight on 25 December 2020 that businesses selling food, drink and liquor for consumption on the premises were required to cease doing such business and non-essential retail businesses were required to close.

Given these facts, I do not consider that it can be said that there was a coordinated response, throughout the period, from the four administrations. Equally, it is an over-simplification to say that the responses over that period were ‘for a single purpose’, namely to prevent the spread of Covid-19. There were a number of purposes which dictated what measures would be adopted when, including the protection of public health, of the economy, and of education. It is apparent that somewhat different decisions – which had a political dimension – were made by different administrations.

Insofar as Zurich relied on the Action Plan of 3 March 2020, I do not consider that this established that there was a Coordinated Response. That plan was at a high level. It did not include concrete measures; and it was in effect superseded. It had not contemplated that businesses across the UK would be required to close entirely. I agree with Greggs that the more drastic measures introduced nationally over the period 16-23 March 2020 cannot sensibly be said to be an implementation of that ‘Action Plan’. Furthermore, I have been shown nothing which persuades me that, after March 2020, the Action Plan was significant in determining the measures taken in the four nations at any given time.

Zurich had a further case, to the effect that it would be sufficient for there to be a ‘single occurrence’ if each regulation made by the four nations ‘was made to deal with the Covid-19 pandemic in the UK’. I do not accept that this very general objective means that all the regulations over 9 months were one ‘occurrence’. They differed between themselves in effect, scope and duration; they were introduced over an extended period of time; and in different places. Taking the perspective of a reasonable policyholder, they were not something which happened at a particular time, in a particular place and in a particular way; and were not one happening.

As to Zurich’s alternative cases based on a ‘sub-division’ of the ‘coordinated government response’, which I have set out in paragraph [64] above, I did not consider that it had established any of them. While I accept that, in March 2020, there was a unitary response, after about June 2020 the responses varied, both amongst the nations, and over time. I do not consider that it has been shown that the various actions and measures adopted in any of the nations taken in the period after about June 2020 can be said to be a single occurrence constituted of a ‘coordinated government response’. Those actions and measures were taken or adopted at different times for different reasons. They did not constitute a single ‘occurrence’, nor one occurrence per nation. It is an over-simplification to say that they were adopted for a ‘single purpose’; and in any event, the fact that there may have been a single objective behind them does not make them one occurrence.

I turn to consider Greggs’ primary case. This was that each of the government measures which sought to contain the spread of SARS-CoV-2 and did so in ways which impacted Greggs’ business which were adopted between March 2020 and the end of the Period of Insurance (and – in principle and subject to causation – also those adopted up to the end of the Indemnity Period) was a separate ‘occurrence’.

In this regard, Greggs emphasised that the adoption of each of these measures was something which happened at a particular time, in a particular place and in a particular way, and met the letter and spirit of the ‘unities’ test. Furthermore, as Greggs submitted, each measure triggered loss, and the temporally and geographically differentiated profile of the regulations in force was directly responsible for the particular profile of Greggs’ losses. As Greggs put it in argument: ‘The loss which Greggs suffered on any given day and in any given locality depended essentially on which [type of regulation] was the situation at that time and place.’

I consider that, in general terms, Greggs’ case is correct. The informed observer, in the position of a policyholder such as Greggs, would regard the adoption of governmental measures which significantly affected whether, when and to what extent its shops could open as being occurrences. They met the ‘unities’ test, and they can without any misuse or even imprecision of ordinary language be described as ‘occurrences’. They would be known to the informed observer in the position of the policyholder, who of course had to know of the legal position as to whether it was permissible to open shops, and under what conditions. The informed observer would readily identify them as relevant occurrences, because of their capacity for causing losses.

Issues of causation and quantum are not matters which fall to be considered in this Preliminary Issues hearing. It is not possible to make any final determinations of what losses, if any, are related to which particular measures. But as a matter of generality, it is possible to say that governmental measures had the clear potential for causing loss to Greggs in a very direct manner. I also consider that such measures were not too remote from Greggs’ losses to be relevant as ‘single occurrences’ for the purposes of the SBIL definition.

The real debate in relation to Greggs’ case was whether all the various measures should properly be regarded as ‘single occurrences’, or whether only a handful, or some or many should be so regarded. As I have already said, Greggs pleaded reliance on some 120 announcements and regulations.

Argument was not addressed in relation to each announcement or regulation separately, and I do not consider that it is possible to determine, at this stage, precisely how many ‘single occurrences’ there were. Some findings can however be made, and certain features identified to assist in identifying the relevant occurrences.

Thus, in the first place, I consider that there was a ‘single occurrence’ on 16 March 2020, in the COBR meeting of that date. At that meeting it was decided that there should be a significantly new policy as a result of the new data which had been emerging from about 13 March 2020 in relation to case numbers and NHS critical care capacity. It was decided at that meeting, for the whole of the UK, that there should be: (i) isolation for 14 days for anyone exhibiting, or living with someone exhibiting a high temperature or a new and continuous cough, and if people in those categories had to go out, they should keep a safe distance from others; (ii) the public should stop non-essential contact with others and unnecessary travel, and this included working at home where possible, and avoiding pubs, clubs, theatres and other social venues; (iii) shielding for those with serious health conditions; and (iv) the end of support for mass gatherings by the emergency services. I do not consider that the announcement of these measures on the same or the next day constituted separate occurrences.

Secondly, there were other occurrences in the form of further governmental announcements and measures in the period 16-23 March 2020, which were a progression from the recommendations decided upon on 16 March 2020, to the decision on 20 March 2020 to tell cafes, bars and restaurants to close and not to reopen, to the announcement of the first UK-wide lockdown on 23 March 2020. I regard those measures as being at least ‘connected with’ the decisions at COBR on 16 March 2020, and furthermore that if and insofar as any losses can be said to have arisen from those decisions, then there is at least a strong possibility that they were equally ‘connected with’ the decisions at COBR on 16 March 2020.

Thirdly, Greggs’ decision to close all its shops from 24 March 2020, even though required only to close eat-in sections, and the losses which flowed from that, appear from the material before me to have been ‘connected with’ the occurrence on 16 March 2020. Greggs pleads that this decision was taken to ‘protect the workforce and support the UK Governments’ COVID-19 initiatives’. The Governments’ initiatives clearly included those decided on and announced on 16 March 2020. Furthermore, Greggs’ Update dated 23 March 2020 (which bore the sub-title ‘Protecting our People, Customers and Business’) identified that while sales had grown (on a like for like basis) in the two weeks after 29 February 2020, in the most recent week to 21 March 2020 they had declined by 9.9%, with the rate of decline increasing each day ‘as more and more customers heed the Government advice on social distancing’, and Greggs stated that it envisaged that that decline would continue if it had continued to trade.

Fourthly, I do not consider that an informed observer would have regarded announcements or measures which simply continued existing restrictions or made trivial changes as being separate ‘single occurrences’ for the purposes of the SBIL definition. I do not believe that it conforms to the parties’ intentions to have aggregation by reference to such matters, which effectively continued a status quo rather than marking any significant change to it. Nor would I consider that an informed observer would have regarded changes which simply reduced restrictions as being separate ‘single occurrences’ for the purposes of the definition. They were such as would of their nature be expected to reduce losses not to lead to them and thus would not constitute the type of matter which would sensibly be regarded as a factor unifying different losses.

Fifthly, and as I have already said, from May 2020 the approaches of the four administrations diverged. From that point, the various announcements/regulations applicable in the four nations cannot be regarded as ‘single occurrences’ even when they were to broadly the same effect. Furthermore, there cannot be said to have been a ‘single occurrence’ constituted by all the governmental announcements/measures in each nation for the entire period after May 2020 until the end of the year (and a fortiori, after the end of the year). There were, instead, a number of different occurrences in each nation, by which new restrictions, which were of significance to Greggs’ business, were brought in at different times. To take England by way of example, there were in my view at least separate occurrences for the purposes of the Policy in the following: one in the bringing into force of the three-tiered system on 14 October 2020; one in the announcement on 4 November 2020 and implementation from 5 November 2020 of the second national lockdown; one in the announcement on 19 December 2020 and implementation from 20 December 2020 of the Tier Regulations (by the Health Protection (Coronavirus, Restrictions) (All Tiers and Obligations of Undertakings) (England) (Amendment) Regulations 2020).

Sixthly, in cases where restrictions were imposed on limited areas of the country, I would regard these as being separate occurrences. Thus I consider that an informed observer would regard the restrictions applied to some parts of Leicester on 4 July 2020 as being a ‘single occurrence’. Those restrictions, it is true, were extended to other parts of Leicester on 17 July 2020, but I do not think that that should be regarded as a separate occurrence; or, if it should, any losses arising from it would probably be ‘connected with’ the 4 July 2020 occurrence, because they all relate to the particularly high incidence of the disease in Leicester in early July 2020. I consider that a similar position holds in relation to the regulations applied to Blackburn with Darwen, Bradford, Luton, Bolton, and a ‘protected area’ in Greater Manchester, Lancashire and West Yorkshire, which were introduced at various times in the period between July and September 2020, and which are Regulations numbered 8, 9, 11 and 15 in Greggs’ amended Appendix One to its RRAPOC.

Seventhly, I agree with Greggs that, subject to the question of whether there are any losses which can be said to have been caused by Covered Events within the Period of Insurance, which is not a matter before me on this occasion, there could, in principle, be occurrences which are relevant for the purpose of aggregation, in the Indemnity Period.

I hope that, in the light of these points, the parties will be able to agree as to the number of relevant occurrences. If they cannot, and it remains material, I will hear further argument by reference to each of the different announcements / regulations pleaded by Greggs.

One further observation is necessary. Concerns were expressed on behalf of Zurich that a finding that there were multiple relevant occurrences would lead to difficulties in assigning losses to them. I treat this submission with caution. Mr Hancock made the point that a calculation of Greggs’ claimable losses on the basis of its regulation-by-regulation case as to aggregation has already been produced by Marsh, and that showed that this was not an impossible exercise. He further submitted that there is no evidence or material before the court which indicates that the exercise will be impossible, or indeed particularly difficult.

I am prepared to accept that, if an informed observer would readily have recognised that it would be impossible to distinguish losses caused by two or more announcements or regulations, she might not regard there as being two ‘single occurrences’. Subject to that caveat, which I do not consider to be applicable to the occurrences which I have identified, I do not think that I should be influenced, in identifying the number of occurrences by putative but unevidenced practical difficulties of loss-assignment.

In substantial measure I have accepted what was put forward at the hearing as Greggs’ primary case, based on the different governmental announcements / regulations. In the circumstances it is probably not necessary for me to address its alternative case, namely that there was an occurrence with each case of the disease occurring within the Vicinity. This is because, if there can be aggregation by reference to the occurrences I have accepted in accordance with Greggs’ primary case, then it does not matter if the individual cases could also be regarded as occurrences. There would still be the need for aggregation by reference to the announcements / regulations which are ‘single occurrences’.

In any event, I do not accept that the BIL which was sustained by reason of the governmental or public response to Covid-19 cases can sensibly be said to have arisen from, have been attributable to or to have been ‘connected with’ each individual case of Covid-19 considered separately. No single case would have produced that governmental reaction or public response. Any ‘connection’ was rather with all the cases which had occurred or were anticipated at the time of the governmental action or public reaction.

The issue in relation to PRCM Costs is a short one. Greggs’ case is that PRCM Costs are subject to a sub-limit of £75,000 in addition to the Limit of Liability and that for each SBIL there is an additional sub-limit for PRCM Costs. Zurich’s case is that the PCRM Costs Sub-Limit applies in the aggregate, such that there is a limit of £75,000 no matter how many SBILs there may be.

In my judgment Greggs is correct on this point. The Sub-Limit is in addition to the Limit of Liability. The Limit of Liability for BIL is expressed as an amount ‘any one SBIL’. Accordingly, the £75,000 applies in addition to the Limit per SBIL.

Zurich’s argument against this is that the Schedule states that ‘The maximum recoverable under Material Damage Costs & Expenses and Business Interruption Costs & Expenses combined will not exceed GBP 75,000.’ This, it is said, indicates that the £75,000 must be an aggregate limit because it is not confined to BIL. I do not consider that this is the case. What that clause is making clear is that if there is a claim for both Material Damage Costs and Expenses and Business Interruption Costs and Expenses where the material damage and the business interruption arise from, are attributable to or are in connection with the same ‘single occurrence’, then the limit for both combined will be £75,000. This provision is thus analogous to the provision in relation to Retentions in the penultimate paragraph of Item 5 of the Schedule.

In relation to Government Support, essentially the same issues arose in relation to CJRS payments (furlough payments) and BRR as arise in Stonegate v MS Amlin. At the hearing, Mr Hancock in effect confined himself to adopting Stonegate’s arguments. I consider that my reasoning in paragraphs [253-289] of my Judgment in the Stonegate v MS Amlin case in relation to CJRS is applicable to the present case. In this regard I should record that, like Stonegate’s, Greggs’ accounts were prepared using the IFRS Standards and Greggs booked CJRS payments as other income, and so, insofar as my Judgment in Stonegate v MS Amlin refers to the accounting treatment in that case, it is applicable here as well.

In relation to BRR, Greggs accepted that business rates were the removal of a cost and, unlike Stonegate, also accepted that they were usually payable out of Turnover. Given this, only the third of the questions identified in paragraph [255] of my Judgment in Stonegate v MS Amlin was live. As to that, my reasoning in paragraphs [294-296] applies in the present case. On that basis, BRR falls to be taken into account.

I should confirm, as was accepted by Zurich, that these items of Government Support are to be taken into account in assessing Greggs’ loss before the application of the SBIL limit, not that there should be an assessment of how much is payable by reference to the SBIL limit and then a reduction of that by the amount of the CJRS and BRR payments.

I will make orders reflecting the above conclusions. I will receive further submissions as to the precise form that those orders should take.