[2025] UKUT 00278 (TCC)

It was common ground between the parties that, as delegated legislation, DRRS §4.5.1 must be interpreted in light of the enabling Act. This reflects the general principle that subordinate legislation should be construed consistently with the purpose and scope of the primary legislation under which it is made.

We also did not understand it to be in dispute that §4.5.1 does indeed function as a deeming provision, as we have described it. In principle, therefore, the provision treats circumstances that would or might not otherwise satisfy the “power to recover” condition as if they did. The scope of the provision in the light of the enabling legislation was however a matter of contention.

The claimants’ overarching submission was that HMRC’s interpretation of §4.5.1 would impermissibly narrow the “no power to recover” condition. It was submitted that this would allow HMRC to deny repayment in circumstances where it quite obviously had no actual right to recover the relevant amount of tax, thereby frustrating the purpose of the Finance Act 2020.

The interaction between delegated rules and primary legislation was considered in R (Sibley) v HMRC [2021] EWHC 3195 (Admin), [2022] STC 336. In that case, a similar argument was raised concerning the scope of HMRC’s powers under delegated legislation. The powers conferred on HMRC by ss. 20–21 of the Finance Act 2020 were discussed by Chamberlain J in his judgment dismissing the taxpayer’s application for permission to apply for judicial review.

Mr Mullan objected that a permission decision could not be cited as authority, referring to the Supreme Court Practice Direction on the Citation of Authorities [2001] 1 WLR 1001. That objection is misconceived. The Practice Direction on the Citation of Authorities does not exclude from citation all types of permission decision. Rather, it sets out (in §§6.1 and 6.2) the following specific categories of judgment which may not be cited unless the judgment clearly indicates that it purports to establish a new principle or to extend the present law:

“Applications attended by one party only

Applications for permission to appeal

Decisions on applications that only decide that the application is arguable

County court cases, unless (a) cited in order to illustrate the conventional measure of damages in a personal injury case; or (b) cited in a county court in order to demonstrate current authority at that level on an issue in respect of which no decision at a higher level of authority is available.”

The Sibley judgment does not fall into any of those categories. It is a decision on an application for permission to apply for judicial review, where the judge concluded that the grounds advanced by the claimant were not reasonably arguable (and refused permission to rely on a further new ground at the hearing). The judgment may therefore properly be relied upon by HMRC. It is clearly relevant to the issues in this case, not least because the claimants’ arguments in the present case bear some similarity to the arguments sought to be advanced in Sibley.

In particular, in Sibley the claimant argued that the DRRS exclusion from repayment of amounts which related to loans which had been repaid by the time of the settlement agreement (not in issue in the present case) frustrated the purpose for which HMRC’s powers to adopt the DRRS were granted, referring in particular to the Morse Report. Chamberlain J rejected that argument, on the grounds that ss. 20–21 of the Finance Act 2020 require HMRC to establish a scheme under which it “may” (not “must”) repay the whole or part of a qualifying amount paid or treated as being paid under a qualifying agreement. He accepted HMRC’s submission that this did not require HMRC to repay every sum which fulfilled the statutory definition of a qualifying amount (§17). He went on to say, at §19:

“In my judgment, the terms of s. 21 are flatly inconsistent with the claimant’s submission that the scope of HMRC’s scheme-making power is limited to dealing with administrative matters such as the form of the application. On the contrary, it is plain from [the] language used that Parliament intended HMRC to have a much broader discretion, both as to the substantive conditions under which repayment would be made and as to the procedural and formal requirements for applications under the Scheme. That discretion must, of course, be exercised according to the usual public law principles. Subject to that, however, Parliament provided that it was HMRC which was to decide which qualifying amounts would be repaid and under what conditions.”

As Chamberlain J explained further at §24, the requirement (disputed by the claimant in that case) that the loan be outstanding introduced a “bright line eligibility condition” which avoided the need for potentially difficult assessments as to the subjective reasons why the settlement was agreed. That was within HMRC’s powers under ss. 20–21, since (§26):

“… Parliament has, in express terms, conferred a power to legislate for, among other things, the conditions for repayment under the Scheme. In doing so, it has signalled with clarity that the legislator (here HMRC) may cut down the range of cases in which repayment is due. In my judgment, the contrary is not reasonably arguable.”

At §§27–31 Chamberlain J went on to reject the further argument that HMRC had acted irrationally by setting a bright line requirement that the loan should be outstanding at the time of the settlement agreement. In his view, even if a bright line rule led to “some distinctions capable of being regarded as anomalies”, it was for HMRC, exercising the discretion conferred on it, to “balance the need to eliminate such anomalies against the virtues of simplicity and administrative workability”.

Applying the same interpretative approach as in Sibley, it is common ground that it was within HMRC’s powers under ss. 20–21 of the Finance Act 2020 to adopt a deeming provision in the interests of certainty. Such a provision could specify circumstances in which HMRC would be regarded as having a power to recover. Thus, although Mr Goodfellow’s submissions criticised HMRC’s interpretation on the basis that it narrowed the circumstances in which repayment could be made, we agree with HMRC that it was possible (as was the case in Sibley)for conditions in a deeming provision to be drafted more broadly than the underlying statutory power.

However, that does not mean that HMRC’s interpretation of §4.5.1 is correct. HMRC advances its interpretation as one which is consistent with the literal reading of the wording. It is true that on a literal reading the provision deems there to be a “power to recover” as long as a Regulation 80 determination is issued for the year for which the amount may have been payable. But the difficulty is that, when read literally, it also extends beyond limits which even HMRC accepts are there. HMRC accepts, for example, that if the Regulation 80 determination identified specific individuals, the deeming provision in §4.5.1 would not allow it to recover amounts of tax which related to other individuals not specified in the determination. That supports the view that the provision cannot be read entirely literally but must, in line with established principles of construction, be read with the legislative purpose in mind.

HMRC’s interpretation of §4.5.1 would be flatly contrary to the purpose of the DRRS. It would mean, for example, that if HMRC had issued a determination for a small amount, a fraction of the voluntary restitution sum, and the taxpayer had not appealed, such that (as Mr Stone accepted) HMRC would have had no power whatsoever to uplift the determination and claim any larger amount, HMRC would nevertheless be able to retain the disputed amount.

There is nothing in the legislative scheme of the Finance Act 2020to suggest that such an anomaly was intended. Nor is there any coherent reason why that should have been intended as a matter of simplicity or workability. The question of what Regulation 80 determinations have been made at a given point and for what amounts should be capable of being readily established.

The more obviously coherent interpretation is that the determination must be for the amount in question. That is consistent with the fact that the entire provision in §4.5 focuses on the power to recover “the amount”, and §4.5.1 itself also repeatedly refers to “the amount”. In the context of those references, it is difficult to see how the “determination” referred to in §4.5.1 can sensibly mean anything other than a determination in relation to “the amount”.

That interpretation is further supported by §4.5.2, which deems HMRC as having had power to recover an amount of NICs where it did take action or could have taken action to protect or recover “the amount”. If the intention was to draw a bright line for income tax, deeming an amount to be recoverable provided that a determination for any amount had been issued (or could still be issued), it is unclear why the line would be drawn differently for NICs.

It is, however, necessary to consider what is meant by the requirement in §4.5.1 that the determination in question must have been made “in respect of any year for which the amount may have been payable”. Recalling that the provision is a deeming provision, the question also arises as to what function §4.5.1 serves if it is not the one that HMRC advances and which we have rejected.

We consider that the purpose of those words, and therefore also one function of the deeming provision, is to cater for the situation where the arrangements implemented by the taxpayer might involve a number of transactions spanning different tax years, creating uncertainty as to the year in which the liability would have arisen. DRRS §4.5.1 thus enables HMRC to forestall an argument that HMRC had protected the wrong year in such circumstances, since it is sufficient that a Regulation 80 determination had been made for “any year for which the amount may have been payable”. Mr Stone’s oral submissions acknowledge this role for §4.5.1.

A further consequence of §4.5.1 is that, provided that a determination has been issued (in respect of such a year, and on our interpretation for the relevant amount), it is not necessary to go further and consider the effect of any potential appeal proceedings under the mechanisms set out at §24 above. The deeming provision therefore avoids a debate about whether HMRC would, under those mechanisms, have a power to recover the relevant amount in light of the stage which any appeal had reached and the outcome of the appeal process. §4.5.1 is in both senses a bright line rule of the sort envisaged in Sibley, to ensure the administrative workability of the scheme.

For all the reasons above, we reject HMRC’s interpretation of the deeming provision in §4.5.1. In our judgment, while that provision does set out a bright line rule in the respects we have just described, the reference to “a determination” must be interpreted as meaning a determination for the amount in question, namely the amount for which HMRC claims a power to recover at the time of the settlement agreement. We therefore do not consider that HMRC can claim that it had “issued a determination” for the purposes of §4.5.1 where that determination was for an amount lower than the amount subsequently paid under the relevant settlement agreement. Accordingly, it is necessary to consider HMRC’s alternative argument that it was entitled to refuse repayment without reliance on §4.5.1, because on the facts of the claimants’ cases HMRC did indeed have “power to recover” the disputed amounts of income tax.