CL-2023-000064 - [2025] EWHC 1870 (Comm)

Once this stage was reached, there seemed in the end to be rather less dividing the parties than appeared at first sight. It is useful to recap the key concessions made by each:

I should add that neither party suggested that the details of the regulatory arrangements mattered: e.g. how regulations happened to be imposed, revised or renewed. It was agreed that we are interested in the substance of the restrictions imposed. It was agreed (for example) that the announcement of the first lockdown by the Prime Minister and Welsh First Minister on 23 March 2020 was not a separate measure from the legal enactment of that instruction in the regulations promulgated on 26 March 2020.

There were perhaps three remaining areas of disagreement in this context. First, there was an issue as to whether the instruction by the Prime Minister to stay at home on 16 March 2020 was a relevant measure or action. Second, the parties disagreed as to whether a new set of regulations which reduced the overall level of restrictions could give rise to a new loss. Third, the U/Ws argue that any change to the level of restrictions must be material in order to trigger a loss, and that many of the supposed increases in restrictions did not pass that test.

To put some meat on those bones by reference to one example used by the Cs: Newcastle Racecourse (or High Gosforth Park Racecourse) is operated by High Gosforth Park Ltd, one of the Cs. The Cs describe the stages through which the restrictions passed in 2020. I can adopt that description because the factual content is not controversial (even if the analysis is):

The U/Ws’ basis for rejecting the instruction by the Prime Minister of 16 March 2020 as a trigger event was explained orally by Mr Scorey as follows:

I do not agree. Limb (b) of the denial of access cover requires only an “action” by a relevant authority (and there is no dispute that the Prime Minster qualifies as such an authority). The nature of an “action” is not described or limited, save only by reference to its effect: it must prevent or hinder use of the Premises. The paragraph does not require an order or a prohibition. It seems to me obvious that the Prime Minister’s instruction hindered the use of the Newcastle Racecourse (or at least would have done if there had been any races scheduled). It is likely to have meant that fewer people went there (indeed, that could be described as its purpose), which amounts to hindrance in the use of those premises: see FCA v Arch in the Supreme Court at [153].

To engage directly with the example used by Mr Scorey, I take the view that paragraph (b) of the denial of access extension would be engaged if, as a result of some local danger, the police were setting up a cordon and discouraging the public from crossing it on safety grounds, even if they were not formally prohibiting anyone from doing so. Giving advice to the public in that way still amounts to an “action” by the police. It may not “prevent” use of the Premises, but it would (or at least could) hinder that use.

I therefore take the view that the instruction by the Prime Minister of 16 March was a trigger event. If it in fact caused any interruption or interference to the business carried on at (say) Newcastle Racecourse, there would need to be a separate loss calculation in respect thereof.

Turning to the second controversial area, I am not persuaded that an instruction or regulatory change which reduces the extent of the restriction on the use of the premises would qualify as a trigger event. In my judgment, there needs to be prevention or hindering of use as compared with what was possible immediately before.

I accept that it is possible to read it as a requirement only that the “action” prevent or hinder use in an absolute sense; i.e. that it suffices if it hinders any use which hypothetically might have been possible, even if that use had not in practice been available immediately before the “action” is taken. But that does not seem to me the natural reading of a provision of this kind. I would read the words of the denial of access extension as being concerned with actual interruption or interference and actual prevention or hindering, which can only sensibly be identified by asking “what would have happened, but for this action?”.

If that were not so, I struggled to follow the logic for the Cs’ concession that a new instruction or order which does not alter the extent of the restrictions on use was not a trigger event. If any action which might be said (ignoring the existing regime) to prevent or hinder use of the premises amounted to a trigger event and required a new loss calculation, I could not see why an action would not also have that effect if the regime was unchanged. If you must ignore the existing restrictions, then surely it makes no difference whether the existing restrictions remain unchanged, or whether the effect is that those restrictions are reduced?

My conclusion is buttressed by consideration of how I would understand the loss calculation to be carried out in scenarios involving more than one trigger event. I dealt at paragraphs 163 - 166 above with the scenario where the fire affects a restaurant at the racecourse and then another fire damages one of the stands. I have suggested that, in such a case, there would be a fresh loss calculation which takes into account the impact of the first fire and that the loss calculation for the original damage would continue (subject to the maximum indemnity period) until the first fire was no longer affecting the results.

My expectation was that the analysis would be broadly similar in the context of non-damage denial of access. For example, if an order by the police prevented access to the restaurant, and then, while that order remained in place, the police also prevented access to the adjacent stand, it seems to me that the loss calculation for the first order would take into account all of the consequences of the first order. It would not artificially stop at the point when the second order was given and require the second calculation to be performed as if the first order had never been given. The calculation in respect of the increase to the restrictions (i.e. preventing access to the adjacent stand) would have to adjust the standard gross revenue to reflect the fact that the first order already meant that the restaurant was out of use, else there would be double-recovery.

Mr Kramer agreed with that analysis in respect of BI premised upon physical damage, but suggested that it was different for non-damage BI, or at least different when the two orders had a common cause (e.g. COVID-19). He submitted that this was a consequence of the decision of the Supreme Court in FCA v Arch (e.g. at [284]) that “the trends or circumstances for which adjustments should be made do not include trends or circumstances arising out of the same underlying or originating cause as the insured peril, namely the Covid-19 pandemic”. I do not agree. That aspect of the decision can be seen to be a product of the need to construe clauses forming part of the machinery of quantification consistently with the insuring clause, and hence “they should be construed so as not to take away the cover provided by the insuring clauses” (see [262]). That is not a concern in the scenario with which we are presently concerned. There is no doubt that the effect on the turnover caused by the first order will be included in a loss calculation; the only question is whether it is included in the first or the second.

Indeed, if one were to take the words of the Supreme Court entirely literally in the context of a series of increasing restrictions imposed by the police, all supposedly arising out of the same “originating cause”, the result would be double or treble recovery, as multiple loss calculations are performed for these overlapping time periods, with each ignoring all of the restrictions already imposed by the previous orders. Mr Kramer suggested that the answer was that the indemnity period in respect of each order came to an end when the next was put in place, to prevent double-recovery. That seemed to me an unprincipled solution, unless the new order can always be said to end the impact of the first order. If all it did was add some further restrictions, that argument would not work. If necessary, I would suggest that COVID-19 should not be treated simplistically as a single “originating cause” over a prolonged period. Rather, the underlying cause of each relevant instruction was the extent of the danger as at the time of that specific instruction.

In the end, I repeat that the issue as to whether there is a new trigger event and the need for a new loss calculation seems to me to depend on whether limb (b) of the denial of access extension tests whether the use of or access to the Premises has been prevented or hindered by comparison with (a) the actual situation immediately before that action is taken, or (b) what might be described as the hypothetical optimal situation for those premises. My reading is that it is generally the former, rather than the latter, for two reasons:

I would illustrate this last by reference to the simplest type of denial of access: imagine the police gave an instruction which prevented access to the bar and restaurant at a hotel, and then, a few weeks later, permitted access to the bar (but not the restaurant). It seems to me that a reasonable observer would say that access to the premises had been hindered by the first instruction, and then the extent of that hindrance was reduced by the second. That reduced (but continuing) hindrance would still fall to be taken into account as part of the loss calculation triggered by the first instruction. But to characterise what is plainly the police allowing access to the bar, as them imposing a (new) restriction on the use of the restaurant, is not consistent with common sense. In much the same way as for “occurrences”, as discussed by Butcher J in Greggs, I suggest that the informed policyholder would not approach our denial of access extension in that artificial way. To the extent that Mr Kramer’s point was that the way in which regulations etc. in the context of COVID-19 were promulgated was more complicated than the police giving one instruction and then another, that risks disappearing into the legalities of revocations and/or amendments, when we all agree that the reasonable policyholder would only be concerned with the substantive effect.

For these reasons, I conclude that that there would only be a new risk trigger, and a fresh loss calculation, if the action of the authority imposed an increased restriction.

The U/Ws add that the increase must be material. In principle, I agree. If the change was minimal and without substance, there could not really be said to be a change. However, it seemed to me that the examples that the U/Ws gave did not fit with this description. To repeat the one identified above (see paragraph 206.4), it was suggested that racecourses moving from tier 3 to tier 4 “only” resulted in the owners of horses being prevented from attending races. Mr Scorey invited me to find that this was not a material change from the perspective of the owner of a racecourse. That struck me as an optimistic submission. After all, as Mr Kramer pointed out, the change must have been expected to prevent the attendance of sufficient people for it to be considered a worthwhile step for the Government to take. That being so, I would expect it to have a more than negligible impact on the Cs’ gross earnings.

The short point is that “materiality” is really only featuring here to avoid being excessively technical about whether there has been a new “action”. There is nothing in the description of the risk which refers to materiality. As such, I would suggest it need only be a relatively low hurdle. To my mind, if the denial of access extension is triggered as a result of a further restriction which at least has the potential to affect the Cs’ gross earnings, the next stage is for a calculation to be carried out, not to prejudge what that loss calculation might reveal.

I should add that, in the end, this felt like an empty debate. If U/Ws are right that, for example, the move of some courses from tier 3 to tier 4 in December 2020 made no difference to the Cs’ bottom line, then that new loss calculation will yield a figure of zero, and the limits end up being irrelevant.

For completeness, Mr Scorey had a specific point about the St Leger “pilot”, where it had been agreed that a race did not happen as was intended. He said this was not an increase in the restrictions, because it had only ever been allowed to take place as a “pilot” for reduced restrictions. His submission about this did not seem to me to be covered by the Agreed Facts and hence I do not see how I could make any findings in that regard. On the face of it, a race was cancelled as a result of a tightening of restrictions. That seems to me, all other things being equal, to represent a material increase in the restrictions imposed on the use of Doncaster racecourse. But if that factual proposition is being challenged for some reason which is not covered in the Agreed Facts, the evidence about it must wait for another day.