BGC Services Holdings LLP v The Commissioners for HMRC

On 8 March 2024, Officer Andrew Swaby decided to issue BGC Services Holdings Ltd (“BGC”) with determinations under Regulation 80 of the Income Tax (Pay As You Earn) Regulations 2003 (“Reg 80” and “the Determinations” respectively. The Determinations totalled £96,037,893 for the tax years 2017-18 through to 2019-20. Officer Swaby’s decisions were authorised by Officer James Miller.

The text of each Determination read as follows:

As can be seen from the above, the text of the Determinations did not include any reasons for their issuance, and there was also nothing in the covering correspondence which explained why HMRC had decided to make them.

BGC appealed the Determinations to the Tribunal. Its grounds of appeal read:

HMRC applied to the Tribunal for BGC to provide further and better particulars of its grounds of appeal. I refused that application in a case management decision issued on 5 March 2025 (“the First Decision”).

HMRC issued its Statement of Case (“SoC”) on 26 March 2025, but that too did not give reasons for the Determinations. On 2 April 2025, BGC applied for HMRC properly to particularise its SoC. On 17 April 2025 I allowed that application by a second case management decision, and I issued Directions to HMRC (together, “the Second Decision”).

On 30 April 2025, HMRC applied for permission to appeal the First and Second Decisions. I refused permission by a decision published under reference [2025] UKFTT 00700 (TC). HMRC did not renew their application to the Upper Tribunal (“UT”).

On 14 July 2025, HMRC filed and served a “further amended Statement of Case” which included reasoning about the application of Conditions A, B and C of the SMR, and other relevant matters.

A case management hearing was then listed with the aim of resolving any remaining issues, in particular as to the scope of HMRC’s pleaded case. The parties also had permission to make other related applications.

The parties filed and served skeleton arguments addressing the various points to be decided at the case management hearing. However, all but one of the matters in dispute were resolved between the parties shortly before the hearing. It is not necessary to set out those agreed issues in this decision, as they are relevant only to the directions subsequently issued to the parties.

The remaining issue was BGC’s application for a preliminary issues hearing. This related to the wording and application of Reg 80, which is headed “Determination of unpaid tax and appeal against determination”, and so far as relevant to this decision, reads (emphasis added):

BGC’s original position, as set out in its application dated 14 November 2025 and in its skeleton argument, was that the Tribunal should direct a preliminary hearing to decide whether HMRC had met the two requirements set out in bold above, namely that (a) at the time the Determinisations were issued, it appeared to HMRC that there may be tax payable, and (b) the Determinations were issued to HMRC’s best judgement.

However, two working days before the hearing, BGC notified the Tribunal and HMRC that its application for a preliminary issue hearing would be limited to the first of those two points, and so no longer included best judgement.

I had the benefit of very helpful written and oral submissions from Mr Grodzinski KC and Ms Belgrano for BGC, and from Mr Carey, Mr Way, Ms Arya and Mr Clement for HMRC. I was particularly grateful for the skilful amendments to their written cases which were required by the changed scope of BGC’s application.

I set out below the positions taken by the parties on Reg 80(1).

BGC’s position was that the phrase “if it appears to HMRC that there may be tax payable”had to be read in the context of the PAYE Regulations as a whole, the purpose of which is to collect tax to which a person was already liable. In Mr Grodzinski’s submission, the “may” in Reg 80(1) referred to the collection of that tax from a particular employer (and not from an employee or someone else). That reading was, he said, supported by Reg 80(3A), which prevented HMRC issuing a Reg 80 direction if they had already directed that an employee was to pay the tax in question under Reg 72F.

Mr Grodzinski went on to say that the necessary threshold would therefore not have been met if Officer Swaby “did not know one way or the other” whether tax was due; if he had only “a suspicion that the relevant legislation may apply pending further investigation”, or if he had thought “there was some extremely low possibility” that the SMRs applied to BGC, and had issued the Determinations on a “protective” basis because he was worried about time limits. In his submission, the position was very similar to that in Go City Limited v HMRC [2024] UKFTT 00745 (TC) (“Go City”), where the Tribunal panel was myself and Ms Shillaker. We found that two VAT assessments were invalid because the HMRC officer in question “had not formed a view” as to whether the related VAT returns were “correct or incorrect”, but had issued assessments under VATA s 73 to stop the periods going out of time, see [163] and [165] of that judgment.

BGC’s skeleton said:

Although HMRC had not as yet disclosed any evidence relating to the making of the Determinations, Mr Grodzinski said it was likely that the threshold had not been met in this case, because:

HMRC’s position was that the wording of Reg 80(1) was clear on its face. The Determinations were valid as long as “it appeared” to HMRC that tax “may” be due. Mr Carey described the threshold as “extremely low”, and said it was therefore “highly unlikely” that HMRC would not succeed in showing that it had been met.

Mr Carey submitted that Go City was irrelevant because it related to entirely different VAT provisions; he instead relied on CooperVision Lens Care Ltd v HMRC [2026] UKFTT 00324 (TC) (“CooperVision”) (Judge Morgan and Mr Robertson) where at [265] the FTT had held:

It was common ground that when deciding whether to allow BGC’s application, I should follow the guidance in Wrottesley v HMRC [2015] UKUT 637. The UT first said at [14]:

Having considered various case law authorities, the UT then set out at [28] the following “key principles” to be applied when deciding whether or not to direct a preliminary hearing:

In Jelly Vine Productions Ltd v HMRC [2024] UKFTT 562 (TC) (“Jelly Vine”), Judge Brown KC considered an application to direct a preliminary hearing following HMRC’s issuance of determinations for three years. She held at [56] that, in addition to the factors set out in Wrottesley:

I respectfully agree with Judge Brown that the merits are a potentially relevant factor.

The next section of this judgment considers each of the above points, so far as relevant, albeit not in the order set out in Wrottesley.

It was common ground that the Reg 80(1) issue was a mixed question of fact and law, because findings of fact would be required about whether Officers Swaby and Miller had formed the view that tax may be payable.

Mr Grodzinski’s position was that this did not prevent the Reg 80(1) issue being decided at a preliminary hearing, because only very limited evidence was required, namely a few documents and the statements from the two Officers.

Mr Carey disagreed, saying the Tribunal would also have to make findings about the Officers’ view as to the application of Conditions A, B and/or C, which would involve “multifaceted, multifactorial assessments”.

I agree with Mr Grodzinski: BGC are seeking to determine whether the Officers had formed a view that tax may be payable, not why they thought tax was payable. Although the UT warned that “the tribunal should be particularly cautious on matters of mixed fact and law”, here the evidence and the consequential factual findings would be clearly delineated and confined to a narrow time period. The need to find some facts therefore does not preclude allowing BGC’s application for a preliminary hearing.

It was also common ground that if a preliminary hearing were to be directed, it could take place in July this year over a two day period. In contrast, if the appeal proceeded to a hearing on all the issues in dispute, that hearing would not take place until the autumn of 2027 at the earliest; it would last between 10 to 12 days, and would require months of expensive preparatory work.

A further relevant factor is that the parties cannot currently take steps to progress the SMR parts of their cases, because finalising their positions depends on the outcome of the Supreme Court judgment in HMRC v BlueCrest Capital Management (UK) LLP (“BlueCrest”). That case was heard in January this year, and the judgment may well not be published before July, when the preliminary hearing can take place. Thus, no time would be lost, and significant costs would be saved. These factors therefore strongly favour allowing BGC’s application.

Mr Grodzinski emphasised that whether or not “it appeared to HMRC that there may be tax payable” was a “knock-out point”. If the answer to that question was “no”, the Determinations would be invalid and the appeal decided in BGC’s favour. He noted that in Jelly Vine, Judge Brown had recorded:

Mr Carey did not dispute that the position would be the same in this case. I therefore agree with Mr Grodzinski that the issue would be “a knock-out point”. That too is a factor in favour of allowing the application.

It was common ground that the evidence about Officers Swaby and Miller’s actions at or around the time of issuing the Determinations would be relevant not only to the Reg 80(1) point but also to Reg 80(2), namely BGC’s challenge that the Determinations had not been made to HMRC’s best judgement. Although the preliminary hearing would not address that issue, if BGC did not succeed, the Tribunal would have to consider the self-same evidence (together with other material) at the subsequent hearing, in order to decide whether best judgement had been exercised.

Mr Grodzinski sought to allay my concern over this overlap in two ways. First, he said that any findings of fact made at the preliminary hearing about the views held by the Officers would be binding on the Tribunal if it subsequently had to determine the best judgement issue. He cited from Chapter 43 of Philson on Evidence under the heading “Res judicata estoppels”. This says at 43:23 that the “central principle” is as follows:

Unfortunately, that citation was not included in the Bundle, but read out as part of Mr Grodzinski’s oral submissions. Mr Carey very fairly did not criticise Mr Grodzinski for this, given that the issues had narrowed shortly before the hearing. However, he said that even if Mr Grodzinski was correct on this point, the Tribunal hearing the remaining issues would nevertheless have to “grapple” with how those pre-determined facts fitted with the rest of the evidence. He added that there could be other concerns: for example, if the Tribunal in the course of the preliminary hearing had found one or both of the Officers to lack credibility, that could affect the weight given to the evidence the Officers would subsequently give about best judgement. Mr Grodzinski responded to this by saying that if an Officer had not been entirely truthful at the preliminary hearing, it would be “perfectly proper” for that Officer’s subsequent evidence to be treated with caution.

Mr Grodzinski also sought to allay my concern about the possibility of overlap by suggesting that, were BGC to lose at the preliminary hearing, it might abandon the best judgement argument, because the threshold for setting aside an HMRC decision on the basis of best judgement was high, for example requiring bad faith by the HMRC Officer or a decision which was “finger in the air stuff, random”.

In order to assess the extent and reliability of any concession on that point, I considered the words used by Mr Grodzinski. He first said “it might even be that we would rethink about whether we'd be pressing for best judgement at all” and later that “whether a best judgement view will really be pursued once we get to a final hearing is probably doubtful – I can't say for sure but that seems very unlikely”. BGC therefore made no definitive concession that it would not take the best judgement point at a subsequent hearing.

I therefore considered for myself how likely it was that they would do so. In Van Boeckel v C&E Commrs [1981] STC 290, Woolf J set out the approach a tribunal should take when deciding a best judgement challenge in a VAT appeal. He said:

I decided I could not rule out the possibility that (were BGC to lose on the preliminary issue) it would maintain its best judgement challenge in the subsequent hearing. I came to that conclusion because:

In my view therefore, BGC may well seek to argue that the Officers had not “fairly consider[ed] all material placed before them and, on that material, [had not] come to a decision which is reasonable and not arbitrary”.

Were that to happen, I agree with Mr Grodzinski that the Tribunal at that subsequent hearing would be bound by the findings made at the preliminary hearing, for the reasons set out in Phipson. However, I also agree with Mr Carey that the FTT would then have to “grapple” with how to treat those findings. That might be a simple matter, but it could be a more complex exercise, for the following reasons:

Thus, if the best judgement issue remains in contention, which in my view is clearly possible, the preliminary hearing issue “cannot be entirely divorced from the evidence and submissions relevant to the rest of the case”.

The UT in Wrottesley described the fourth factor as follows (my emphasis):

The UT did not say that the Tribunal should have regard to whether there is a “significant risk” that determining the preliminary issue would “prejudice” or “prevent” the Tribunal from arriving at a just result at the subsequent hearing of the remainder of the case. Instead, it set a much lower bar.

That is consistent with the fact that the judge deciding whether to direct a preliminary issues hearing has limited information. Specifically, in this case I have no evidence about either the Reg 80(1) issue or the Reg 80(2) issue, and of course I do not know what will happen at trial.

Taking all the above into account, the risk of overlap with the best judgement issue is a factor pointing away from allowing BGC’s application.

The likelihood of BGC succeeding is also, as Judge Brown said, a relevant factor. She focused on cases where “the argument had good prospects of success and was therefore more likely to deliver the knockout point asserted”. The opposite is also true: it is plainly not in the interests of justice to direct a preliminary hearing on an issue which has a very low chance of success, because the most likely outcome is that the parties are put to the time and expense of a wholly unnecessary hearing when the point could have been dealt with more efficiently at the same time as the other issues in dispute.

BGC’s case in a nutshell is that Reg 80(1) has to be understood in the context of the rest of the PAYE regulations; their purpose is to collect tax to which a person is liable, and HMRC can only use the power given by Reg 80(1) if they have first come to a view that the there is a liability to tax which may be payable by the person in question. I agree with BGC that this point is arguable, and HMRC did not submit to the contrary. However, it is difficult to estimate the chances of success, and I therefore find only that this “merits” point does not weigh in the balance against BGC.

One of the factors listed in Wrottesley is “the possibility of a separate appeal on the preliminary issue”.:

Mr Carey was clear that if HMRC were to lose the Reg 80(1) issue at a preliminary hearing, they were extremely likely to appeal, because other taxpayers may seek to rely on an adverse judgment.

He invited BGC to say whether they would concede if it lost at the Tribunal, but Mr Grodzinski did not confirm the position one way or the other. He said:

I have thus taken it that whoever wins at the Tribunal, there is a risk of an onward appeal, with the probability rising to a near-certainty were HMRC to lose.

In my judgment, if the losing party were to apply for permission to appeal, it is likely that the application would be allowed. That is because there is no clear case law on the issue. In Go City, the Tribunal considered different legislation: whether it appeared to the HMRC officer in question that the taxpayer’s VAT returns were “incomplete or incorrect”, not whether “there may be tax payable”. In CooperVision, Reg 80(1) was considered together with the Reg 80(2) best judgement issue, rather than as a free-standing point, see [262], [263] and [265] of that decision.

Thus, although a hearing of the preliminary issue can be neatly timetabled to take place while the Supreme Court is finalising its judgment in BlueCrest, that is very unlikely to be the end of the matter. Moreover, an appeal to the UT might be followed by onward appeals to the Court of Appeal and even to the Supreme Court. If, at the end of that process, BGC were to lose, its appeal would have been delayed for at least a year and possibly as long as four years. That is a factor against allowing BGC’s application.

Mr Carey submitted that were I were to direct a preliminary hearing on this issue, this could open the floodgates not only to appellants applying for preliminary hearings on Reg 80(1) and/or Reg 80(2) determinations, but also to recipients of discovery assessments and VAT assessments. He said allowing BGC’s application risked “disaggregating cases that are, properly, matters that could and should be heard together”.

In relation to VAT assessments, my position remains as set out in Go City at [191]:

Mr Carey did not explain how allowing a preliminary hearing on Reg 80(1) could open the floodgates in relation to discovery assessments, which have their own different legislative provisions, and I have therefore not considered this further.

There is already binding case law about preliminary hearings on best judgement, see in particular Pegasus Birds Ltd v HM Customs & Excise [2004] EWCA Civ 1015. It is therefore unlikely that allowing BGC’s application would encourage appellants to apply for a preliminary hearing on Reg 80(2) alone.

I accept however that allowing BGC’s application might encourage other recipients of a determination to apply for a preliminary hearing on Reg 80(1). However, the number of applications is likely to be limited, because in the majority of cases it will be clear from the correspondence that HMRC formed a view before issuing the determinations, so the point will not be arguable on the facts.

In the few cases where that is not the position, an appellant would only make a preliminary hearing applications were BGC to succeed. Then, as Mr Grodzinski said, severing the Reg 80(1) issue and determining it on the facts would save time and money for the parties and the Tribunal, and so would not be a negative factor.

For the above reasons, this factor does not weigh in the balance against BGC.

Pausing there, the factors in favour of allowing BGC’s application are as follows:

The factors against allowing BGC’s application are as follows:

Thus, although there are significant savings in time and cost when the preliminary hearing is considered in isolation, the high risk of onward appeals changes the position, and the evidential overlap with best judgement is a further negative factor. The scales are thus tipped towards refusing BGC’s application. I now turn to the remaining factor.

The UT’s guidance in Wrottesley begins by saying that “the power to deal with matters separately at a preliminary hearing should be exercised with caution and used sparingly”. This echoes the frequent warnings given by other courts: Mummery J (as he then was) said in NUT v St Mary’s School [1995] ICR 317 at p 323 that there were “dangers in taking what looks at first sight to be a short cut but turns out to be productive of more delay and costs than if the dispute had been tried in its entirety”. In Tilling v Whiteman [1980] AC 1 at p 25, Lord Scarman similarly said that preliminary points of law “are too often treacherous short cuts. Their price can be…delay, anxiety, and expense”. In Wentworth Sons Sub-Debt SARL v Lomas [2017] EWHC 3158, Hildyard J said at [29]-[31]:

When that guidance is considered together with my summary at §64-66, I am in no doubt that BGC’s application must be refused.

I have issued directions for the main appeal separately. If BGC were to appeal this decision, those directions will not be suspended but will instead remain in full force and effect unless or until BGC succeeds in overturning this decision.

This document contains full findings of fact and reasons for the decision. Any party dissatisfied with this decision has a right to apply for permission to appeal against it pursuant to Rule 39 of the Tribunal Procedure (First-tier Tribunal) (Tax Chamber) Rules 2009. The application must be received by this Tribunal not later than 56 days after this decision is sent to that party. The parties are referred to "Guidance to accompany a Decision from the First-tier Tribunal (Tax Chamber)" which accompanies and forms part of this decision notice.