News Corp UK & Ireland Ltd v Commissioners for His Majesty’s Revenue and Customs

Introduction

Under the Value Added Tax Act 1994 (the “VAT Act”), Value Added Tax (“VAT”) is not charged on newspapers. In the language of section 30 of the VAT Act, newspapers are “zero-rated”. The central question in this case is, in the specified period of 30 August 2010 – 4 December 2016, whether that zero-rating extended beyond print newspapers to what we shall refer to as “digital editions” of newspapers (ie editions for e-readers, tablets, smartphones and websites).

At a high level of generality, the answer to that question requires the application of principles of legislative interpretation and EU law that may be said to be pulling in different directions. On the one hand, there is the “always speaking” principle (of domestic statutory interpretation) which may be thought to support a wide interpretation. On the other hand, there is the EU law requirement to interpret exemptions from VAT strictly and to give effect to a “standstill” provision.

It is common ground that the decision in relation to the period in question will also be applicable to the period up to 1 May 2020. However, as from 1 May 2020, there can be no dispute because by reason of the VAT (Extension of Zero-Rating to Electronically Supplied Books etc) (Coronavirus) Order 2020, (SI No 2020/459) (the “2020 Order”), zero-rating has been “extended” to newspapers “when supplied electronically” (unless wholly or predominantly devoted to advertising or consisting wholly or predominantly of audio or video content). In other words, it is clear that, unless falling within the exception, digital editions of newspapers are zero-rated as from 1 May 2020.

Jonathan Peacock KC, counsel for News Corp UK & Ireland Ltd (“News Corp”), described this case as providing a classic example of where an “always speaking” interpretation is appropriate. This is because the facts concern a technological development (from printed newspapers to digital editions of newspapers) where the underlying purpose behind the VAT zero-rating carries through to the new development. This is an attractive submission not least because of its rational simplicity. But there are also powerful arguments to the contrary.

News Corp’s central claim was rejected by the First Tier Tribunal (“FTT”): [2018] UKFTT 129 TC, accepted on appeal by the Upper Tribunal (“UT”): [2019] UKUT 404 (TCC), but rejected again by the Court of Appeal: [2021] EWCA Civ 91. News Corp now appeals to the Supreme Court.

In answering the central question, the parties were in agreement that one should initially put to one side the relevance, if any, of the EU principle of fiscal neutrality. This principle was essentially relied on by Mr Peacock as a fall-back position in the event of the court rejecting his central submissions. In line with this, we shall leave consideration of the principle of fiscal neutrality to the end of this judgment.

It is common ground between the parties that the withdrawal of the UK from the European Union (“EU”) has no impact at all on the issues in this case. While the UK was part of the EU, VAT was governed by EU Directives and those Directives were implemented in the UK by domestic statutes, in particular by the VAT Act. By reason of the European Union (Withdrawal) Act 2018 and the European Union (Withdrawal Agreement) Act 2020, the relevant EU law and EU derived domestic legislation is “retained EU law” after the implementation completion day (31 December 2020) but, in any event, the period with which this case is concerned expired before the implementation completion date.

The factual background

The newspapers in question are The Times, The Sunday Times, The Sun and The Sun on Sunday. In the relevant period, the relevant digital editions of The Times and Sunday Times were available in an e-reader edition, a tablet edition, a smartphone edition and a website edition; and The Sun and The Sun on Sunday were available on The Sun classic app and, as introduced in 2013, The Sun+ package.

The e-reader edition was an exact facsimile of the newsprint edition and could be downloaded on a daily basis onto a tablet computer or viewed on a personal computer. The tablet edition could be downloaded on a daily basis onto a tablet computer and then read off-line. The smartphone edition could be downloaded on a daily basis onto a smartphone and then read off-line. The website edition could be viewed at www.thetimes.co.uk on any internet browser. The Sun classic app consisted of a digital replica of each print edition of The Sun each day available for use on tablets but allowed customers to skip between sections using the contents page. The Sun+ package was a bundle including entitlement to view The Sun website which was behind a paywall at the time, and access to The Sun+ goals app showing Premier League football goals and allowing subscribers to take advantage of certain perks such as free cinema tickets.

The Sun+ was not successful and was withdrawn in November 2015 at the same time as the paywall to the website was removed. During the period that The Sun+ was available The Sun website edition closely mirrored the content and structure of the print edition.

The FTT found - and by the time this case reached the Court of Appeal these findings were not challenged by HMRC - that the digital editions of The Times, The Sunday Times and The Sun on Sunday had "similar characteristics to those of the newsprint editions" (at para 150). The digital editions were essentially periodic-based publications (at para 152) and the “content of the digital and newsprint editions was ... fundamentally the same or very similar" (para 153). The "digital editions were essentially, when the evidence was viewed in the round, the same as or very similar to the newsprint editions" (para 155); and readers considered them "to be fundamentally the same as the newsprint editions" (para 156).

Although the digital editions did provide some additional features to those in the print editions (such as videos, interactive puzzles, links to podcasts, access to Scottish and Irish sections and the ability to store articles in a “my articles” section and to share articles via social media) (para 105), these additional features were "relatively lightly used" (paras 65 and 105) and were “a relatively minor aspect of those digital editions" (para 155). Some of the digital editions could update their content, had search functions, and provided access to archived material and links to additional content (see, eg, paras 38, 68, 72, and 107). From the point of view of the subscriber , it was more "the content than the medium of its delivery to which most value was attached, although subscribers also valued the additional convenience of the digital platform" (para 156).

As found by the FTT (para 17) the social policy behind the decision to zero-rate newspapers was the promotion of literacy, the dissemination of knowledge and democratic accountability by having informed public debate.

EU and domestic legislation

Between 1940 and 1973 the UK levied an indirect tax on the wholesale price of goods called Purchase Tax. Certain items were exempt from Purchase Tax, including newspapers. Following the UK’s accession to the European Economic Community (“EEC”) on 1 January 1973, Purchase Tax was replaced by VAT with effect from 1 April 1973.

VAT in general applies to the supplies of all goods and services at the standard rate, though the EU has specified that certain supplies shall be exempt from VAT (“EU mandated exemptions”). The EU has also under specified conditions permitted member states to continue to exempt certain supplies that are taxable under EU law, to maintain zero rates for supplies that are taxable under EU law, and to apply a reduced rate.

A member state's ability to apply a reduced or zero rate to certain supplies originated in Directive 67/228 (“the Second Council Directive”) which stated in the last indent of article 17 that member states were permitted, on what was described as a transitional basis, to "provide for reduced rates or even exemptions with refund [ie zero rates]”. Such measures were only authorised where they were “taken for clearly defined social reasons and for the benefit of the final consumer” (“the cumulative conditions”). These conditions were imposed in light of the general aim of harmonisation of VAT and the fact that, as stated in the fifth recital, “the introduction of zero rates gives rise to difficulties, so that it is highly desirable to limit strictly the number of exemptions”.

In anticipation of joining the EEC, the UK took advantage of the authorisation conferred by article 17 of the Second Directive to preserve the tax-free treatment of newspapers by enacting section 12 of the Finance Act 1972 (“the 1972 Act”). It did so in the knowledge that the authorisation was intended to be on a transitional basis but, in circumstances where the cumulative conditions were met, it wished to preserve the status quo in relation to the taxation of newspapers. Section 12 of the 1972 Act accordingly provided for the zero-rating of supplies listed in Group 3 of Schedule 4 to the Act. The supplies listed as Item 2 of Group 3 were “Newspapers, journals and periodicals” (“Item 2”). Zero rates are termed under EU law “exemptions with refunds”.

In May 1977 Directive 77/388 ("the Sixth VAT Directive") was adopted. Article 28(2) of that Directive provided:

“2.

Reduced rates and exemptions with refund of the tax paid at the preceding stage [ie zero rates] which are in force on 31 December 1975, and which satisfy the conditions stated in the last indent of Article 17 of the second Council Directive of 11 April 1967, may be maintained until a date which shall be fixed by the Council...

On the basis of a report from the Commission, the Council shall review the abovementioned reduced rates and exemptions every five years and, acting unanimously on a proposal from the Commission, shall where appropriate, adopt the measures required to ensure the progressive abolition thereof".

The position following the enactment of the Sixth VAT Directive was therefore that member states were permitted to continue to apply certain zero rates that were in force prior to 31 December 1975 provided that the cumulative conditions were met. No provision was made, however, permitting new zero rates to be adopted by member states, even if the cumulative conditions were met. Thus, in effect only zero rates in force prior to 31 December 1975 were permitted. Article 28 of the Sixth VAT Directive was referred to as a “standstill” provision by the Court of Justice of the European Union (“the CJEU”) in Talacre Beach Caravan Sales Ltd v Customs and Excise Commissioners (“Talacre Beach”) Case C-251/05, [2006] STC 1671 at para 22. As the zero-rating of newspapers was still in force on 31 December 1975, the UK could continue to maintain that zero rate after the enactment of the Sixth VAT Directive. Article 28(2) was amended by Directive 92/77 of 19 October 1992 so as to permit zero rates that were “in force on 1 January 1991” to be maintained but only if they “are in accordance with Community law”.

The UK zero rate for newspapers was preserved in the Value Added Tax Act 1983 and in the VAT Act, which came into force on 1 September 1994 and is the governing Act for the claims in this case. Group 3 of Schedule 8 of the VAT Act provides:

"Group 3—Books, etc

Item No

1 Books, booklets, brochures, pamphlets and leaflets.

2 Newspapers, journals and periodicals.

3 Children's picture books and painting books.

4 Music (printed, duplicated or manuscript).

5 Maps, charts and topographical plans.

6 Covers, cases and other articles supplied with items 1 to 5 and not separately accounted for.

Notes

(1)

Items 1 to 6—

(a)

do not include plans or drawings for industrial, architectural, engineering, commercial or similar purposes; but

(b)

include the supply of the services described in paragraph 1(1) of Schedule 4 in respect of goods comprised in the items.

(2)

Items 1 to 6 do not include goods in circumstances where—

(a)

the supply of the goods is connected with a supply of services, and

(b)

those connected supplies are made by different suppliers.

(3)

For the purposes of Note (2) a supply of goods is connected with a supply of services if, had those two supplies been made by a single supplier—

(a)

they would have been treated as a single supply of services, and

(b)

that single supply would have been a taxable supply (other than a zero-rated supply) or an exempt supply.”

Items 1 to 6 have been stated in the same terms since the 1972 Act. That Act included a Note in materially the same terms as Note (1)(a). Note (1)(b) was added in the Value Added Tax Act 1983. The other Notes were added in the VAT Act.

The Sixth VAT Directive was repealed in 2006 and replaced by Directive 2006/112 (which has been referred to in the courts below, and in the submissions to this court, as the “Principal VAT Directive”: we shall continue to refer to it in this way). Article 110 of the Principal VAT Directive (which reflected the amended terms of article 28(2) of the Sixth VAT Directive) provided as follows in relation to zero rates:

"Member States which, at 1 January 1991, were granting exemptions with deductibility of the VAT paid at the preceding stage [ie zero rates] or applying reduced rates lower than the minimum laid down in Article 99 may continue to grant those exemptions or apply those reduced rates.

The exemptions and reduced rates referred to in the first paragraph must be in accordance with Community law and must have been adopted for clearly defined social reasons and for the benefit of the final consumer"

Article 110 required that zero-rating was still being granted on 1 January 1991 for it to be continued but it is common ground for the purpose of this appeal that the relevant “standstill” date for the UK remained 31 December 1975 (ie as under the Sixth VAT Directive). This is because a category of supply which was not zero-rated on 31 December 1975 could not later become zero-rated.

Article 98 of the Principal Directive provided that reduced rates “shall not apply to electronically supplied services” but that reduced rates could be applied to goods and services set out in Annex III. That Annex included: "(6) supply, including on loan by libraries, of books on all physical means of support (including brochures, leaflets and similar printed matter, children's picture, drawing or colouring books, music printed or in manuscript form, maps and hydrographic or similar charts), newspapers and periodicals, other than material wholly or predominantly devoted to advertising". The words "on all physical means of support" were added by Directive 2009/47 effective from 1 June 2009. Article 98 and point 6 of Annex III apply to printed books, books on CDs, CD-ROMs and USB keys but not to electronically supplied books – see Case C- 219/13 K Oy [2015] STC 433 para 34. An argument that electronically supplied books should also fall within this provision, because reading electronically supplied books required physical support (such as a computer), was rejected on the basis that such support was not included in the supply – see European Commission v Luxembourg Case C-502/13, [2015] STC 1714, para 36. Article 7 of Council Implementing Regulation (EU) No 282/2011 of 15 March 2011 laying down implementing measures for the Principal VAT Directive expressly stated that “electronically supplied services” (which were excluded from the reduced VAT rates) included “the digitised content of books and other electronic publications”, “subscription to online newspapers and journals”, and “online news, traffic information and weather reports”.

In 2018 under Directive 2018/1713 member states were allowed to apply a reduced VAT rate to the supply of books, newspapers and periodicals irrespective of whether they are supplied on physical means of support or electronically; and member states then applying zero rates to books, newspapers or periodicals supplied on physical means of support were allowed to apply the same VAT treatment to such books, newspapers or periodicals when supplied electronically. It was pursuant to this Directive that zero-rating was extended to newspapers “when supplied electronically” under the 2020 Order (see para 3 above). The Explanatory Memorandum to the 2020 Order at para 6.2 says the following:

“This instrument amends Group 3 to extend the zero-rating of some of the publications listed in that Group to electronically supplied versions of those publications. The European Union (EU) vires for these changes are contained in Council Directive (EU) 2018/1713 which made changes to the scope of the EU’s optional reduced rate provisions. This Directive came into force in December 2018. Where member States (including the UK during the Transition Period) have a reduced rate (or an equivalent zero rate) for supplies of printed publications, it allows them to extend that reduced or zero rate to supplies of electronic versions of those publications subject to certain exclusions. The UK has decided to exercise this option to extend its zero rate to supplies of some electronic publications.”

Summary of the decisions below

FTT

Judge Guy Brannan, sitting as the judge in the FTT, [2018] UKFTT 129 (TC), dismissed the appeals of News Corp against the decisions of HMRC that the digital editions of The Times, The Sunday Times, The Sun and The Sun on Sunday did not fall within Item 2 of the VAT Act and were therefore not zero-rated for VAT purposes. Having made the findings of fact summarised at paras 8-13 above, the judge’s essential reasoning was as follows:

UT

The UT, Zacaroli J and Judge Greg Sinfield, [2019] UKUT 404 (TCC), allowed News Corp’s appeal against the decision of the FTT. The essential reasoning of the UT was as follows:

Court of Appeal

The Court of Appeal allowed the appeal of HMRC against the decision of the UT. The leading judgment was given by Simler LJ with whom Rose LJ and Sir Geoffrey Vos MR agreed: [2021] EWCA Civ 91. Simler LJ’s essential reasoning was as follows:

Domestic statutory interpretation: words, context and purpose; and “always speaking”

It is clear that the modern approach to statutory interpretation in English (and UK) law requires the courts to ascertain the meaning of the words used in a statute in the light of their context and the purpose of the statutory provision: see, eg, Quintavalle, para 8 (per Lord Bingham); Uber BV v Aslam [2021] UKSC 5; [2021] ICR 657, para 70; Rittson-Thomas v Oxfordshire County Council [2021] UKSC 13; [2022] AC 129, para 33; R(O) v Secretary of State for the Home Department [2022] UKSC 3; [2022] 2 WLR 343, paras 28-29.

Within that modern approach, it is also a well-established principle of statutory interpretation that, in general, a provision is always speaking: see, eg, Royal College of Nursing of the United Kingdom v Department of Health and Social Security (“Royal College of Nursing”) [1981] AC 800; R v Ireland [1998] AC 147, 158-159 ; Quintavalle; Owens v Owens [2018] UKSC 41, [2018] AC 899 (approving [2017] EWCA Civ 182, [2017] 4 WLR 74); Test Claimants in the Franked Investment Income Group Litigation v Revenue and Customs Commissioners [2020] UKSC 47, [2022] AC 1. See also Craies on Legislation (ed Daniel Greenberg), 12thed , (2022) ch 21; and Bennion, Bailey and Norbury on Statutory Interpretation 8^th ed, (2020) ch 14.

What is meant by the always speaking principle is that, as a general rule, a statute should be interpreted taking into account changes that have occurred since the statute was enacted. Those changes may include, for example, technological developments, changes in scientific understanding, changes in social attitudes and changes in the law. Very importantly it does not matter that those changes could not have been reasonably contemplated or foreseen at the time that the provision was enacted. Exceptionally, the always speaking principle will not be applied where it is clear, from the words used in the light of their context and purpose, that the provision is tied to an historic or frozen interpretation. A possible example (referred to by Lord Steyn in R v Ireland at p 158) is The Longford (1889) 14 PD 34 where the word “action” in a statute was held not to be apt to cover an Admiralty action in rem: at the time the statute was passed, the Admiralty Court “was not one of His Majesty’s Courts of Law” (p 38).

The great merit of the always speaking principle is that it operates to prevent statutes becoming outdated. It would be unrealistic for Parliament to try to keep most statutes up to date by continually passing amendments to cope with subsequent change.

It is instructive to look in more depth at some of the leading cases on the always speaking principle, several of which were relied on by Mr Peacock.

In Royal College of Nursing the question was whether an abortion carried out under a new technique was lawful under the Abortion Act 1967. That Act legalised “termination [of a pregnancy] by a registered medical practitioner”. Under the new technique the abortion comprised a nurse pumping a fluid into the womb under the supervision of a doctor who would be on call but might not be present. It did not involve surgery or an injection by a doctor. The House of Lords held, by a 3-2 majority, that such an abortion was covered by those words and was therefore lawful. This was so even though in 1967 Parliament could not have envisaged the development of that mode of abortion. Although Lord Wilberforce was one of the dissentients, a passage from his judgment was subsequently approved in Quintavalle. Lord Wilberforce said, at p 822:

“In interpreting an Act of Parliament it is proper, and indeed necessary, to have regard to the state of affairs existing, and known by Parliament to be existing, at the time. It is a fair presumption that Parliament's policy or intention is directed to that state of affairs… [W]hen a new state of affairs, or a fresh set of facts bearing on policy, comes into existence, the courts have to consider whether they fall within the Parliamentary intention. They may be held to do so, if they fall within the same genus of facts as those to which the expressed policy has been formulated. They may also be held to do so if there can be detected a clear purpose in the legislation which can only be fulfilled if the extension is made. How liberally these principles may be applied must depend upon the nature of the enactment, and the strictness or otherwise of the words in which it has been expressed. The courts should be less willing to extend expressed meanings if it is clear that the Act in question was designed to be restrictive or circumscribed in its operation rather than liberal or permissive. They will be much less willing to do so where the new subject matter is different in kind or dimension from that for which the legislation was passed. In any event there is one course which the courts cannot take under the law of this country: they cannot fill gaps; they cannot by asking the question, "What would Parliament have done in this current case, not being one in contemplation, if the facts had been before it?", attempt themselves to supply the answer, if the answer is not to be found in the terms of the Act itself”.

In R v Ireland in 1998 the House of Lords decided that causing or inflicting “actual bodily harm” in sections 18, 20 and 47 of the Offences Against The Person Act 1861 includes causing or inflicting “psychiatric illness”. Lord Steyn said at pp 158-159:

“The proposition that the Victorian legislator when enacting sections 18, 20 and 47 of the Act of 1861, would not have had in mind psychiatric illness is no doubt correct. Psychiatry was in its infancy in 1861. But … the 1861 Act is …’always speaking’…: the statute must be interpreted in the light of the best current scientific appreciation of the link between the body and psychiatric injury.”

In Quintavalle the question at issue was the interpretation of the word “embryo”. Creating human embryos outside the body was regulated by the Human Fertilisation and Embryology Act 1990. When the Act was passed the only known method of creating an embryo involved fertilisation and in the Act an embryo was defined as “a live human embryo where fertilisation is complete.” Subsequently, a new method of creating an embryo - cell nuclear replacement – had been developed by scientists which did not involve fertilisation. Although that development could not have been reasonably envisaged at the time the Act was passed, the House of Lords held that an embryo created by this new method was regulated by the Act. The words “where fertilisation is complete” were interpreted as laying down the time at which an embryo should be treated as an embryo rather than being integral to the definition of an embryo.

In Owens v Owens, a divorce case, the Supreme Court, upholding the Court of Appeal, decided that, although the marriage had broken down irretrievably, a divorce should not be granted to the wife because she had failed to prove that her husband’s behaviour was such that she could not reasonably be expected to live with him. Therefore, none of the grounds for divorce in section 1 of the Matrimonial Causes Act 1973 (re-enacting sections 1-2 of the Divorce Reform Act 1969) had been established. Particularly informative on the always speaking doctrine was Sir James Munby P’s judgment in the Court of Appeal. He said, at para 38, that because an Act is “always speaking”, one needed to construe it “taking into account changes in our understanding of the natural world, technological changes, changes in social standards and, of particular importance here, changes in social attitudes.” He explained that, in the family law context, what is covered by, for example, a “child’s welfare” (originally used in section 1 of the Guardianship of Infants Act 1925, now section 1 of the Children Act 1989) was to be judged by the standards of 2017, not those of 1925. It followed that the objective test in this case (“cannot reasonably be expected to live with [him]”) should be judged by the standards of 2017, not those of 1969. The relevant standards were not those “of the man or woman on the Routemaster clutching their paper bus ticket … in … 1969 … but the man or woman on the Boris Bus with their Oyster Card in 2017” (para 40). Yet even applying the standards of 2017, when a wife might be reasonably expected to be less tolerant than in 1969, she had failed to make out her case.

Finally, Test Claimants in the Franked Investment Income Group Litigation v Revenue and Customs Commissioners illustrates that the relevant changes may include changes in the law. Various companies had been required, by domestic legislation, to pay advance corporation tax, contrary to EU law. Those companies, the claimants, sought restitution of the tax paid that was legally not owed. Under the Limitation Act 1980, the normal limitation period for restitution of money paid is six years from the date of payment. But the claimants had been paying this tax for decades before they brought their actions for restitution. The claimants therefore sought to rely on the mistake exception to the running of time first laid down in 1939 and re-enacted in section 32(1)(c) of the Limitation Act 1980 that “where the action is for relief from the consequences of a mistake, the period of limitation shall not begin to run until the plaintiff has discovered the … mistake … or could with reasonable diligence have discovered it.” The claimants argued that their payments had been made by mistake of law and that they could rely on section 32(1)(c) to recover payments made more than six years before the actions were brought. It is important to appreciate that it was only in 1998, in Kleinwort Benson Ltd v Lincoln City Council [1999] 2 AC 349, that the House of Lords had decided that there could be restitution for payments made by mistakes of law as well as fact.

In what can be viewed as an application of the always speaking doctrine (although, at paras 218-219, the majority, Lord Reed and Lord Hodge with whom Lord Lloyd-Jones and Lord Hamblen agreed, stressed that this was merely an aspect of purposive interpretation so that it did not matter whether or not the always speaking principle was specifically in issue), the majority reasoned that section 32(1)(c) applied even though, at the time the 1939 and 1980 Limitation Acts were passed, the state of the law was such that only mistakes of fact triggered restitution. The state of the law had since moved on and the best interpretation of the Act should apply the purpose of the provision to the present, not the past, state of the law. The purpose was to postpone the running of time for any cause of action based on mistake. Therefore the claims for mistake of law, giving recovery of all payments made within six years of when the mistake could with reasonable diligence have been discovered, were allowed.

EU law: strict interpretation of exemptions and the effect of the standstill provision

Strict interpretation of exemptions

It is well established that zero-rating provisions must be interpreted strictly because they constitute exemptions to the general principle that all supplies of goods and services for consideration by a taxable person should be subject to VAT. They should not, however, be interpreted so strictly as to deprive the exemption of its intended effect. As stated by Lord Kitchin in SAE Education at para 42:

"In accordance with well-established principles, the terms used in articles 131 to 133 to specify exemptions from VAT must be construed strictly. Nevertheless, they must also be construed in a manner which is consistent with the objectives which underpin them and not in such a way as to deprive them of their intended effects."

See also Werner Haderer v Finanzamt Wilmersdorf Case C-445/05, [2008] STC 2171, para 18.

The need for strict interpretation is particularly marked where, as in this case, it does not involve mandated EU exemptions, but rather national law exceptions tolerated by EU law within the constraints of the EU standstill provision. As explained by the Advocate General in Talacre Beach, national exceptions must be “interpreted narrowly” (para 17) and, because they are not directed at the same objectives as EU mandated exemptions, “it is necessary to take particular care that the exceptions are not extended” (para 42). The need for a strict interpretation was endorsed by the CJEU (para 23).

The effect of the standstill provision

As has been mentioned at para 21 above, it is common ground that, in the present case, the standstill provision in article 110 means that the categories of zero-rating cannot be expanded or extended beyond those which existed on 31 December 1975.

The nature and effect of standstill provisions was considered in Talacre Beach. As already stated (see para 24(iii) above), the case concerned the UK’s zero-rating of caravans but not their contents. At the time the relevant standstill provision was that in article 28(2)(a) of the Sixth VAT Directive (as amended). In its judgment the CJEU stated as follows:

"20.

It is also common ground that the VAT Act specifically excludes some items supplied with the caravans from exemption with refund of the tax paid [ie zero-rating]. It follows that, so far as those items are concerned, the conditions laid down in art 28(2)(a) of the Sixth Directive, in particular the condition that only exemptions in force on 1 January 1991 can be maintained, are not fulfilled.

21.

Therefore, an exemption with refund of the tax paid in respect of those items would extend the scope of the exemption laid down for the supply of the caravans themselves. That would mean that items specifically excluded from exemption by the national legislation would be exempted nevertheless pursuant to art 28(2)(a) of the Sixth Directive.

22.

Clearly, such an interpretation of art 28(2)(a) of the Sixth Directive would run counter to that provision's wording and purpose, according to which the scope of the derogation laid down by the provision is restricted to what was expressly covered by the national legislation on 1 January 1991. As the Advocate General observed in paras 15 and 16 of her opinion, art 28(2)(a) of the Sixth Directive can be compared to a `stand-still' clause, intended to prevent social hardship likely to follow from the abolition of exemptions provided for by the national legislature but not included in the Sixth Directive. Having regard to that purpose, the content of the national legislation in force on 1 January 1991 is decisive in ascertaining the scope of the supplies in respect of which the Sixth Directive allows an exemption to be maintained during the transitional period".

The judgment makes it clear that (i) the purpose of the standstill provision tolerating the maintenance of zero rates is “to prevent social hardship likely to follow from the abolition of exemptions provided for by the national legislature but not included in the Sixth Directive”; (ii) in accordance with that purpose and its wording, the scope of derogation allowed by the standstill provision “is restricted to what was expressly covered by the national legislation” on (in the case before us) 31 December 1975; (iii) “the content of the national legislation in force” on that date “is decisive” in ascertaining “the scope of supplies” allowed; and (iv) the “scope of the exemption” laid down in that legislation is not to be extended.

In the Court of Appeal, Simler LJ (at para 64) observed that the standstill provision is “intended to preserve the status quo” as it was on 31 December 1975. That observation is largely correct, as is the notion of freezing the law as at that time (see the FTT at para 24(ii) above), but both phrases have to be treated with some care because some development is consistent with the purpose of the standstill provision (as we illustrate in the context of this case at para 58 below). The crucial question is the extent of the development.

The proper interpretation of “newspapers” in Item 2 (leaving aside fiscal neutrality)

The ordinary meaning of the word “newspapers” in context as at 31 January 1975

The VAT Act (which is the governing Act for the period with which this case is concerned) came into force on 1 September 1994. But as has been explained, the effect of the standstill provision is that, under EU law, zero-rating could not be extended after 31 December 1975. In interpreting the word "newspapers” in Item 2, the starting point (and this was common ground between the parties) is therefore the ordinary meaning of the word “newspapers” in its context as at 31 December 1975. At that date, “newspapers” referred only to printed newspapers. Those were the only kind of newspapers which existed at that time and digital editions lay many years in the future. Although there was no evidence directly in point, it would appear that the same would have followed even if one were to take the relevant date as being 1 January 1991 (in accordance with the Principal VAT Directive, as set out in para 21 above) and probably even if the relevant date had been 1 September 1994 when the VAT Act came into force. In order to succeed on its appeal News Corp therefore accepts – and indeed it is at the forefront of Mr Peacock’s submissions - that it has to rely on the always speaking principle and to show that, applying that principle and purposive interpretation, digital editions are properly to be viewed as falling within the zero-rated category of “newspapers”.

Purposive interpretation

As regards the purpose of the zero-rating of newspapers, Mr Peacock placed considerable reliance upon the FTT’s finding at para 17 of its decision (see para 13 above) that "...the social policy required by article 110 which lay behind the UK's decision to zero-rate newspapers and books etc. was the promotion of literacy, the dissemination of knowledge and democratic accountability by having informed public debate...". As Mr Peacock submitted, that purpose and policy is satisfied by digital editions in the same way as it is by printed editions. A purposive interpretation of the word “newspaper” leads, Mr Peacock submitted, to the clear conclusion that it includes digital editions.

However, with regard to social purpose, we agree with Ms Eleni Mitrophanous KC for HMRC that, in the realm of taxation, the fact that the same social purpose may be served by zero-rating two related items does not mean that they should or will be treated the same way. Taxation involves budgetary and political decision-making and how far to extend exemptions from tax is a budgetary choice. Such decisions are made on a granular basis. As she submitted, an interpretation which entails that more supplies promoting the social purpose are captured by the zero rate is not a reason for favouring that interpretation. This is illustrated by the example given by Simler LJ (at para 71) of an online rolling news service which no one suggests is a newspaper although it may be said to fulfil the same social purpose. The role played by identifying the social purpose of the legislation in enacting Item 2 zero-rating newspapers can go no further than ensuring that interpretations which do not meet that purpose are rejected. Put another way, identifying that the social purpose of the zero-rating of newspapers extends to digital newspapers is of limited assistance in deciding on the correct interpretation because that same purpose is over-inclusive in applying equally to many items (eg an online rolling news service) that cannot possibly be covered by the word “newspapers”.

There is another aspect of the relevant purpose that must be considered. From the perspective of EU law, and as made clear in Talacre Beach, the purpose of the standstill provision tolerating the maintenance of zero rates was to prevent social hardship likely to follow from the abolition of existing national law exemptions. Yet no social hardship could follow from the exclusion of digital editions from the ambit of the standstill provisions as, at the material time, nobody had access to them. Moreover, the zero-rating for newspapers was seen as a transitional phase with the ultimate purpose being harmonisation with no derogations at all. This purpose, consistently with the strict approach to exemptions and the effect of the standstill provision, indicates that a narrow meaning should be given to the word “newspapers” and, in particular, is best seen as having a limiting effect on how far one can apply the always speaking principle.

The application of the always speaking principle having regard to the constraints of EU law

Turning to the application of the always speaking principle that is at the heart of this appeal, it is clear that, in this case, that principle has to be applied having regard to the EU law constraints imposed by the standstill provision and the principle of strict interpretation of exemptions. As explained above, that is reinforced by a purpose of the law on VAT, as seen from the perspective of EU law, as being harmonisation with no derogations. Here these constraints mean that the always speaking principle is significantly limited so as to ensure that it does not conflict with the requirement for zero-rating for newspapers to be strictly construed and not extended.

In Royal College of Nursing Lord Wilberforce stated (in the passage set out at para 32 above) that “how liberally” the always speaking principle is to be applied depends “on the nature of the enactment”. He explained that: “The courts should be less willing to extend expressed meanings if it is clear that the Act in question was designed to be restrictive or circumscribed in its operation rather than liberal or permissive”. The EU law constraints mean that this is a case in which the court “should be less willing to extend expressed meanings” than in other cases where a liberal or permissive approach is called for. In this case the always speaking principle falls to be applied at the less liberal end of the scale.

Lord Wilberforce also made clear that the always speaking principle applies where the new state of affairs falls within “the same genus of facts” as that for which the legislation was passed, and that it will not do so if it is “different in kind or dimension”. How then does that apply in this case?

Mr Peacock contended that the defining characteristics of a “newspaper” in 1975 and now are that it comprises the publication of edition based, curated news. What matters is the content of the publication, not the medium by which it is delivered. A digital edition is such a publication.

In relation to the question of defining characteristics, Mr Peacock’s approach does not pay sufficient regard to what was meant by “newspaper” as at 31 December 1975. What, viewed at that time, would be regarded as being the defining characteristics of a newspaper? Plainly at that time “newspaper” would be understood to be news communicated through the medium of print in a physical form. This is reflected in the word itself, which refers to news on “paper”. Throughout the history of newspapers up to that time they had only existed in printed physical form. Although news was also communicated through the medium of radio and television, such news bulletins would not be understood to be or to be akin to newspapers. The medium of print in a physical form would have been regarded as a defining characteristic of a newspaper in 1975. It is also true that the other items in Group 3 were physical (in 1975), as Simler LJ explained in detail at para 69(i) and (ii) of her judgment.

Another defining characteristic at that time was that the buyer of a newspaper obtained complete access to the news in that paper. There was no requirement of connectivity. Access did not depend on owning or buying something else, such as a device. If, for example, it had been possible in 1975 to obtain a facsimile of a newspaper on a television screen that would have been regarded as being a form of television news rather than a newspaper because you would have to own or rent a television in order to access it.

Those two characteristics – first, a physical printed form and, secondly, accessibility without a separate device – are a reflection of there being a conceptual difference between newspapers in 1975 and digital editions and of that difference being a radical one which opens up all sorts of possibilities for interactive communication. It is these features which, along with the strict and non-expansive interpretation required by EU law, render it clear, in our view, that digital editions fall outside “the same genus of facts” as newspapers in 1975. In other words, those underlying fundamental features, viewed in the light of the EU law constraints, mean that the difference between newspapers in 1975 and digital editions is one of kind not merely degree.

The fundamental conceptual difference between print newspapers and digital newspapers is that the former are goods whereas the latter are services. Indeed, as Simler LJ explained (at paras 23 and 72), the inclusion of services within Item 2 would create new legal issues such as deciding where the services are supplied. Harrier LLC v Revenue and Customs Commissioners (see para 25(iv) above), which Mr Peacock relied on, is distinguishable because that merely involved the inclusion of a different type of physical book albeit one produced by new technology. It was not concerned with the inclusion of services as would be involved with electronic books. We also consider that the UT’s reliance on section 30 of the VAT Act, as an indication that the distinction between goods and services is irrelevant, is misplaced. That provision authorises zero-rating on goods and services but it does not mean that that distinction is irrelevant in deciding the scope of Item 2. Certainly, given the significant difference between print and digital editions of newspapers, it cannot be said to be irrational to distinguish between the VAT treatment of printed newspapers and digital editions.

The rationality of such a distinction is borne out by the fact that it has been drawn in EU VAT law. Directive 2009/47/EC allowed reduced rates to be extended to newspapers (or books etc) supplied by physical means of support, such as CD-ROMs and memory sticks, but did not permit an extension of reduced rates to newspapers (or books etc) dependent on electronically supplied services. See also article 7 of Regulation (EU) No 282/2011 and, in the context of books, the decision of the CJEU in European Commission v Luxembourg (para 22 above). Although it is correct that, as Mr Peacock stressed, EU VAT law was directly dealing with reduced rates, rather than zero-rating as in the UK, it is clear that, prior to the period with which this case is concerned and through to the amendment to the Principal VAT Directive in 2018, the EU was drawing a clear distinction for VAT purposes between printed newspapers and digital editions of newspapers.

Moreover, the technological development that has led to digital editions is a radical one which takes one a long way from the physical item and opens up all sorts of possibilities for interactive communication that were not possible with print newspapers. These include videos and audio links and updating. Although the FTT found, and it is not in dispute before us, that the content of the digital editions is the same or very similar to physical newspapers, it is clear that digital newspapers are in other respects very different from print newspapers.

In our view, therefore, having regard to the constraints of EU law, the always speaking principle cannot be applied so as to interpret newspapers as covering digital editions. This is not to close off entirely the operation of the always speaking principle in this context. So, for example, it is not in dispute that new versions of print newspapers, eg produced by computer rather than hot metal presses and containing colour rather than black and white photos, would all be included even if computers and colour photos were not used (and were possibly not even contemplated) in 1975. Similarly, if paper was replaced by a substitute physical material, newspapers published on such material would also likely be included. These examples, however, are very different to digital editions and do not involve the adoption of an impermissibly expansive approach, as would be required to include digital editions.

Subsequent legislative treatment

Although unnecessary to our decision, it is also noteworthy that, subsequent to the specified period, both the EU Commission and the UK Parliament have made reforms explicitly dealing with electronically supplied newspapers in 2018 and 2020 respectively (see para 23 above). Applying a wide-ranging approach so that digital editions were already covered by the zero-rating for VAT prior to 2020 would entail accepting that, eg, Parliament was acting unnecessarily and indeed on a mistaken basis (see as evidence for this the explanatory memorandum referred to in para 23 above) when it explicitly “extended” the items (albeit with some exceptions) as from May 2020. That one can refer to subsequent legislation to resolve an ambiguity in earlier legislation has been accepted in several cases: see, eg, Cape Brandy Syndicate v Inland Revenue Commissioners [1921] 2 KB 403, 414; Commissioner of Inland Revenue v Hang Seng Bank [1991] AC 306, 323-324; Bennion, Bailey and Norbury on Statutory Interpretation 8^th ed, (2020) section 24-19.

Conclusion on the proper interpretation of “newspapers” in Item 2

For all these reasons, we agree with the Court of Appeal that “newspapers” in Item 2 is not to be interpreted as including digital editions. The ordinary meaning of “newspapers” in context as at 31 January 1975 referred only to print newspapers; and a contrary view does not follow from an examination of the purpose of the relevant provisions. The always speaking principle has to be applied narrowly given the constraints of EU law. The relevant “genus of facts” should be viewed as covering only physical items involving the medium of print and no connectivity requirement. Digital editions do not fall within this categorisation.

Fiscal neutrality

Under the principle of fiscal neutrality, goods and services which are “similar” should be treated the same way for VAT purposes - see Commissioners for Her Majesty's Revenue and Customs v The Rank Group plc C-259/10 and C-260/10) [2011] ECR I-10947.

News Corp’s case was that in view of the FTT’s finding that the digital editions are "fundamentally the same or very similar" to the printed editions from the point of view of the consumer (para 153), it would distort the market and breach fiscal neutrality for the digital editions to be standard-rated whilst the printed editions are zero-rated. As HMRC pointed out, however, this assumes that News Corp has already established what is necessary for it to succeed on the interpretation issue. If so, News Corp does not need to rely on the principle of fiscal neutrality. If not, the principle does not assist them as it is well established that it cannot be relied upon to extend the scope of an exemption – see, for example, Finanzamt Frankfurt am Main v-Hochst v Deutsche Bank AG [2012] STC 1951 (at para 45).

At the hearing it was acknowledged by Mr Peacock that the principle of fiscal neutrality could only avail News Corp if the court was to hold that some forms of digital edition were to be zero-rated while others were not to be. We have not so concluded and, in those circumstances, it is not necessary to address the potential applicability of the fiscal neutrality principle any further.

Conclusion

In conclusion, and for the reasons set out above, we would dismiss the appeal.

I agree with Lord Hamblen and Lord Burrows that the appeal should be dismissed. But I express my reasons separately as important questions of principle are raised - in particular, regarding the so-called “always speaking doctrine” in interpreting statutes - on which my reasoning does not coincide in all respects with theirs.

The central issue in the appeal is whether the digital editions (described by Lord Hamblen and Lord Burrows at paras 8 to 10 of their judgment) of The Times, The Sunday Times, The Sun and The Sun on Sunday supplied by News Corp during a period between 2010 and 2016 were “newspapers” within the meaning of that term in Schedule 8 to the Value Added Tax Act 1994 (the “VAT Act”), which applies “zero rates” of VAT to certain specified items. Those items include (at Item 2, Group 3, Schedule 8) “Newspapers, journals and periodicals.” Zero rates are more advantageous than pure exemptions from VAT because suppliers of zero-rated goods and services are not only exempt from liability to charge and pay VAT on their supplies of such goods and services but are entitled to recover input tax incurred in order to make those supplies.

When the United Kingdom joined the European Economic Community (EEC) in 1973, VAT was introduced in place of an earlier tax on consumption known as purchase tax. This was done to comply with European legislation, at that time contained in the Second Council Directive (67/228/EEC). The Directive had as its aim establishing a common system of VAT throughout the EEC. However, in what was intended to be a transitional measure, article 17 of the directive permitted member states to provide in their national legislation for “reduced rates or even exemptions with refund [ie zero rates],” provided such measures were taken “for clearly defined social reasons and for the benefit of the final consumer.” From its inception, the UK VAT legislation included “Newspapers, journals and periodicals” in a schedule of zero-rated items: see Item 2, Group 3, Schedule 4 to the Finance Act 1972.

Subsequently, the Sixth Council Directive (77/388/EEC) prohibited member states from introducing new zero rates but allowed those zero rates in force on 31 December 1975 to be maintained. Article 110 of the Principal VAT Directive (2006/112/EC), which was the applicable European legislation during the relevant period between 2010 and 2016, contained a similar “standstill” provision permitting member states to maintain zero rates in force on 1 January 1991.

Changes in the applicable European legislation during the period that the UK was a member state included an amendment to the Sixth Council Directive made in 1992 to allow member states to apply a reduced rate (of not less than 5%) to supplies of specified categories of goods and services, which included books, newspapers and periodicals. (As regards those items, this change did not directly affect the UK as those items were already zero-rated.) In 2002, following a decision to introduce new harmonised VAT rules for “electronically supplied services,” such services were excluded (by Council Directive 2002/38/EC) from those to which a reduced rate of VAT could be applied. The provisions which allowed member states to apply a reduced rate to supplies of (among other items) books, newspapers and periodicals, but excluding “electronically supplied services,” were carried over into the Principal VAT Directive. In 2009, the words “on all physical means of support” were inserted after the reference to “books” in the list of items to which reduced rates could be applied.

Council Directive (EU) 2018/1713, which post-dates the relevant period, allowed member states which were then applying reduced rates or zero rates to certain books, newspapers or periodicals “supplied on physical means of support” to extend this treatment to such books, newspapers or periodicals when supplied electronically. In consequence, an amendment was made to Group 3 of Schedule 8 to the VAT Act with effect from 1 May 2020 to add, as a new Item 7, “the publications listed in Items 1 to 3 when supplied electronically.” It is thus not in dispute that, since 1 May 2020, digital editions of newspapers have been zero-rated.

It is common ground that the capability of delivering the content of a newspaper to a consumer in electronic form did not exist and was not in contemplation when the original provision for the zero-rating of newspapers was enacted in the UK in 1972, nor for that matter when the VAT Act was enacted in 1994. An issue in the appeal is what relevance this fact has in deciding whether the term “newspapers” in Item 2, Group 3, Schedule 8 of the VAT Act is to be interpreted as including the digital editions after they were introduced in 2010.

On the one hand, HMRC argues that the term “newspapers” as used in the Finance Act 1972 and when the first standstill provision took effect on 31 December 1975 cannot be taken to have included digital editions because digital editions did not exist at that time and were not within the ordinary meaning of the term. Therefore, as no subsequent extension of the scope of zero-rating was permitted by European law until after the relevant period, the term “newspapers” as used in the VAT Act cannot be interpreted as including the digital editions.

On the other hand, News Corp argues that it is an established principle of English law that legislation should be interpreted as “always speaking” so as to take into account changes - such as developments in technology - that occur after its enactment. Counsel for News Corp submit that applying this principle as part of a purposive approach to interpretation leads to the conclusion that the term “newspapers” should be read as including the digital editions. They further submit that such an interpretation does not involve any extension of the list of zero-rated items, as “newspapers” (however that term is properly to be construed) have always been zero-rated since VAT was first introduced in the UK by the Finance Act 1972.

I am not persuaded by either of these arguments. For the reasons that I am about to give, the fact that digital editions and the technology required to produce them did not exist in 1972 (or 1994) in my opinion makes no difference from a legal point of view to whether such editions do or do not fall within the meaning of the term “newspapers” in the VAT Act. No inference one way or the other as to whether the term encompassed the digital editions when they began to be supplied in 2010 can be drawn from that fact.

The metaphor that legislation is “always speaking” is arresting but enigmatic. The first use of the expression has been traced to a treatise by a barrister, George Coode, On Legislative Expression: Or, the Language of the Written Law (1845): see Neal Goldfarb, “‘Always Speaking’? Interpreting the Present Tense in Statutes” (2013) 58 Canadian Journal of Linguistics 63. In this treatise Coode criticised the practice which was common at that time of drafting legislation in the future tense. An example still on the statute book is section 20 of the Offences against the Person Act 1861, which provides that “[w]hosoever shall unlawfully and maliciously wound or inflict any grievous bodily harm upon any other person … shall be guilty of a misdemeanor …” This drafting practice seems to have been motivated by a concern to make it clear that the law was intended to apply to cases which occurred after its enactment. Coode argued that this concern was misplaced because the law is “always speaking” (p 23) and recommended instead the use of the present indicative in drafting legislation on grounds of clarity and simplicity. Coode’s essential point was that it is unnecessary to use the future tense because rules expressed in the present tense are naturally understood as applicable to events which occur at any time while the legislation is in force. It is, after all, the purpose of legislation to create such rules.

Coode’s treatise was extremely influential and his recommendation to use the present tense has been widely adopted in legislative drafting. It was taken up by, among others, the great Victorian Parliamentary Counsel, Lord Thring, who wrote in his book Practical Legislation (1877) at p 32:

“An Act of Parliament should be deemed to be always speaking and therefore the present or past tense should be adopted, and ‘shall’ should be used as an imperative only, and not as a future.”

The principle that legislation is “always speaking” as an explanation of how verbs in the present tense should be interpreted has been codified in a number of jurisdictions, including Northern Ireland. Section 31(1) of the Interpretation Act (Northern Ireland) 1954 provides:

“Every enactment shall be construed as always speaking and if anything is expressed in the present tense it shall be applied to the circumstances as they occur, so that effect may be given to each enactment according to its true spirit, intent and meaning.”

It seems ironic that the metaphor that an enactment is “always speaking” has itself changed its meaning and come to be used to express a different idea from the one originally expounded by Coode. Broadly stated, this idea is that legislation should be interpreted in a way that allows for, or adapts it to, changes that have occurred since the legislation was enacted. An interpretation of this kind is referred to in Bennion, Bailey and Norbury on Statutory Interpretation, 8th ed (2020), section 14.1, as an “updating construction.” As there stated:

“Acts are usually regarded as ‘always speaking’. Here, it is presumed that the legislature intends the court to apply a construction that allows for changes that have occurred since the Act was initially framed (an ‘updating construction’).”

Other expressions used in the literature on statutory interpretation to convey this idea include “ambulatory meaning” (see eg JF Burrows, “The Problem of Time in Statutory Interpretation” [1978] NZLJ 253), “dynamic interpretation” (see WN Eskridge, Dynamic Statutory Interpretation (1994)) and “evolutionary interpretation”.

It is unclear how the description of legislation as “always speaking” acquired a sense among English lawyers so different from that in which it was originally used. This may, at least in part, reflect the influence of Bennion on Statutory Interpretation. In the first edition published in 1984, section 146, and still in the most recent edition of this book, the commentary conflates what may be called the old and new uses of the “always speaking” metaphor without any recognition that they are different concepts. Be that as it may, the idea that legislation should as a general rule be given an “updating”, “ambulatory”, “dynamic”, “evolutionary” or (if the term is preferred) “always speaking” interpretation has come to be seen as orthodox. For example, in In re McFarland [2004] UKHL 17, [2004] 1 WLR 1289, para 25, Lord Steyn felt able to say:

“It is now settled that legislation, primary or secondary, must be accorded an always-speaking construction unless the language and structure of [the] statute reveals an intention to impress on the statute a historic meaning. Exceptions to the general principle are a rarity.”

See also earlier dicta of Lord Steyn in R v Ireland [1998] 1 AC 147, 158, and R (Quintavalle) v Secretary of State for Health [2003] UKHL 13, [2003] 2 AC 687, para 23. In his 2017 Hamlyn Lectures, Thinking About Statutes, p 21, Professor Burrows (as he then was) described the proposition that a statute is “always speaking” as “trite law” - although he went on to state, aptly if I may say so, that “what this precisely means is open to debate.”

I did not question the general proposition - although I did question what it precisely means - in R (on the application of ZYN) v Walsall Metropolitan Borough Council [2014] EWHC 1918 (Admin), [2015] 1 All ER 165, paras 39-48. But, with the benefit of the arguments on this appeal, I would go further. It now seems to me that the general rule or presumption articulated by Bennion and others that statutes are to be interpreted as “always speaking” is stated at too high a level of generality to be meaningful. Rather, there are different types of change that may occur after a statute is enacted to which different considerations apply.

This is not the place in which to attempt a full analysis, but I think it useful for present purposes to distinguish four different types of change.

One type is linguistic. Any natural language such as English is always in flux. The ordinary or conventional meaning of words changes over time. Where a word used in a statute changes its ordinary meaning or acquires a new additional meaning, I can see no justification for “updating” the meaning of the word. The rational assumption is that words used by Parliament were intended to bear the meaning they had when they were used and not whatever different meaning they might happen to acquire subsequently because ordinary usage changes. To take one of many examples of linguistic change given in the first edition of Bennion on Statutory Interpretation (1984) pp 367-8, a statute enacted in the reign of Henry VII required a member of the King’s household accused of conspiring to murder the King or any Lord of the realm to be tried by a jury of “twelve sad men”: 3 Hen 7 c 14 (1487). The word “sad” then meant sober and discreet. It would be absurd when the ordinary meaning of the word later changed to have interpreted the statute as requiring a jury to consist of twelve dolorous individuals.

An anomalous case in which a term used in a statute was construed as having its subsequently acquired rather than its original meaning is R v Munks [1964] 1 QB 305. The defendant had arranged two wires with the intention and result that, when his wife opened a French window to enter the house, she received a severe electric shock. He was convicted of an offence under section 31 of the Offences against the Person Act 1861 of placing “any spring gun, mantrap, or other engine” calculated to inflict grievous bodily harm. The word “engine” in 1861 meant any product of human ingenuity. The defendant’s conviction was quashed by the Court of Appeal, however, on the ground that, particularly as this was a penal statute, the word “engine” should be given what the court took to be “the natural meaning today” of the word as a mechanical contrivance or machine, which did not cover the contrivance used by defendant. I agree with the view expressed in Bennion, Bailey and Norbury on Statutory Interpretation, 8th ed (2020), at p 515, that this reasoning was erroneous. Had the word “engine” been ambiguous when the statute was enacted, it might have been justifiable to prefer the narrower meaning on the ground that penal statutes should be strictly construed. But the fact that the word had subsequently acquired a narrower meaning could not justify treating the statute as having become narrower in its scope.

A different type of change that occurs over time is change in social attitudes and values. Legislation often uses language that requires a value judgment to be made in applying it to factual situations. Examples of words or phrases that require such judgments are “reasonable”, “safe”, “obscene”, “welfare”, “cruelty”, “the public interest”: the list is endless. I think it right that there is in such cases a general rule or presumption that contemporary social values are to be applied rather than those which existed when the legislation was enacted. The rationale for this is that Parliament can be taken to have anticipated that values change over time; and the rational intention to attribute to Parliament is that the application of legislation should reflect this, and not that those whom it affects should remain in thrall to social attitudes and values which were current at the time of its enactment but have since become outdated. When evaluative terms such as “reasonable”, “safe”, “obscene” etc are used, they should therefore generally be understood to denote what is “reasonable”, “safe”, “obscene” etc by contemporary and not historic standards.

A number of the cases in which reference has been made to the so-called “always speaking principle” are of this type. An example is Owens v Owens [2017] EWCA Civ 182, [2017] 4 WLR 74; affirmed [2018] UKSC 41, [2018] AC 899. This was a contested petition seeking a divorce on the ground that, in the words of section 1(2)(b) of the Matrimonial Causes Act 1973, “the respondent has behaved in such a way that the petitioner cannot reasonably be expected to live with the respondent.” In the Court of Appeal, at paras 39-41, Munby P made the point that the test of what a spouse “cannot reasonably be expected to live with” was to be applied by reference to contemporary standards of reasonableness, not the standards of the time when the legislation was enacted. On the further appeal to the Supreme Court Lord Wilson likewise observed, at para 30, that:

“although its interpretation by these courts remains correct even after 40 years, the application of the subsection to the facts of an individual case is likely to change with the passage of the years.”

It is important to recognise that, as both Munby P and Lord Wilson made clear, in cases of this kind it is not the meaning of the statute which changes but its application to new cases. Lord Wilson quoted the words of Lord Bingham of Cornhill in the Quintavalle case, at para 9:

“The meaning of ‘cruel and unusual punishments’ has not changed over the years since 1689, but many punishments which were not then thought to fall within that category would now be held to do so.”

In Birmingham City Council v Oakley [2001] 1 AC 617, 631, Lord Hoffmann made the same point, using the same example, when he said:

“The concept of cruelty is the same today as it was when the Bill of Rights 1688 ... forbade the infliction of ‘cruel and unusual punishments’ (section 10). But changes in social standards mean that punishments which would not have been regarded as cruel in 1688 will be so regarded today.”

This approach can be extended to the interpretation of statutory provisions which, although not directly requiring the application of an evaluative standard, use terms which are imbued with social attitudes and values. An example is a line of cases in which provisions of the Rent Acts giving rights of succession to a “member of the tenant’s family” living in the rented property have been interpreted to reflect changing understandings of the family: see eg Fitzpatrick v Sterling Housing Association Ltd [2001] 1 AC 27. Again, however, although drawing the distinction is not always straightforward, it is important to distinguish between changes in social values and changes in meaning. In Fitzpatrick Lord Slynn of Hadley proceeded by analysing the characteristics of a “family” as the term was used in the original legislation enacted in 1920 and then asked whether “two same-sex partners could satisfy those characteristics so as today to fall within the word ‘family’.” Having regard to changes of attitude in society towards same-sex relationships, he concluded that they could. After reaching this conclusion he said, at p 39C:

“It is accordingly not necessary to consider the alternative question as to whether by 1999 the meaning of the word in the 1920 Act needs to be updated. I prefer to say that it is not the meaning which has changed but that those who are capable of falling within the words have changed.”

Lord Slynn’s preferred reasoning respects the distinction drawn, for example, in the passages quoted at para 86 above between the meaning of statutory language and its application to things. As already stated, I cannot accept that it would be justifiable to attribute a new, updated meaning to a word used in a statute. That would cross the constitutional line between interpretation, which is the role of the courts, and legislation, which is not. Had it been necessary in Fitzpatrick to consider the “alternative question” identified by Lord Slynn, therefore, the proper conclusion would have been that any such “updating” must be left for Parliament.

A third type of change that may occur after legislation has been enacted is advance in scientific knowledge. As with changes in social attitudes and values, it is reasonable to attribute to the legislature the intention that the words used should be applied in a way which reflects modern scientific understanding. Thus, in R v Ireland [1998] 1 AC 147 the House of Lords held that, having regard to the best current scientific appreciation of the link between the body and psychiatric injury, the term “bodily harm” in the Offences against the Person Act 1861 could include psychiatric injury.

The facts of the present case involve a further type of change, namely, technological change. It has often been suggested that the “always speaking doctrine” applies where there have been developments in technology. I do not think it right, however, to view changes of this type in the same way as changes in social values or in scientific knowledge. Whereas there is a presumption, in deciding what rights and obligations a statute creates, that current social values and scientific knowledge should be applied, the advent of new technology seems to me an essentially neutral factor. There is no equivalent justification for any general rule or presumption that a statute is intended to apply to a newly invented object or process. In the case of inventions, there is no reason why the law should favour novelty for its own sake. The most that can be said is that the fact that an object or process did not exist and could not have been foreseen when a statute was enacted is not a reason to regard the statute as inapplicable and that a purposive interpretation may lead to the conclusion that the statute applies to it. But there is nothing singular about this. It is the very essence of any general rule that its scope is not limited to specific cases contemplated by its authors and that it is capable of being applied to cases which they did not (or even could not) foresee. General words are not mere compendia of particular instances envisaged by those who use them: they extend beyond any set of particular instances to reach new instances. When unforeseen situations occur, the question whether a statutory provision applies may be easy or difficult to answer. But in answering that question, no special significance attaches to the fact, where it is the fact, that the situation has arisen as a result of a new technological development.

Take the hypothetical example of a law which prohibits vehicles in a public park: see HLA Hart, “Positivism and the Separation of Law and Morals” (1958) 71 Harvard Law Review 593, 607 (“the most famous hypothetical in the common law world”: Frederick Schauer, “A Critical Guide to Vehicles in the Park” (2008) 83 New York University Law Review 1109). Some objects clearly fall within its scope - for instance, motorcycles. Other cases are harder: for example, does the law apply to skateboards? I cannot see that in answering that question it makes any difference whether the law was made in modern times or is an old law made before skateboards were invented. In either case the interpreter will need to ask whether as a matter of ordinary language the word “vehicle” is capable of being used to refer to a skateboard and also to identify the purpose of the law and ask whether a skateboard falls within the mischief at which it is aimed. The legislative history might be relevant in answering the latter question and whether skateboards existed when the law was made may be relevant when considering the legislative history. But, if so, this is merely a contingent fact and has no bearing on the correct approach to interpretation.

There are of course many examples of cases in which courts have had to decide whether words used in a statute apply to objects which did not exist when the statute was enacted. Traditionally, courts have approached such cases in just the same way as they would approach any case where it is uncertain whether a particular type of object falls within the scope of a legislative provision. For example, in the late nineteenth and early twentieth centuries the question arose in various cases whether the term “carriage” in laws made before bicycles had been invented should be interpreted as including a bicycle. Unsurprisingly, the answer given to that question varied according to the nature, purpose and detailed wording of the particular legislation. In Taylor v Goodwin (1879) 4 QBD 228 a Divisional Court decided that a bicycle came within a law which made it a criminal offence to drive “any sort of carriage … furiously so as to endanger the life or limb of any passenger” (the latter word being used in its archaic sense of a person travelling on foot). The court reasoned that the word “carriage” was capable of including a bicycle and that the furious driving of a bicycle was within the mischief at which the law was aimed. On the other hand, in Williams v Ellis (1880) 5 QBD 175 another Divisional Court held that a bicycle was not a “carriage” within the meaning of a law which authorised a toll to be charged for using a road. In Simpson v Teignmouth and Shaldon Bridge Co [1903] 1 KB 405, 413, the Earl of Halsbury LC, comparing these two cases, commented that they marked the distinction “between an enactment which is contemplated as being for the protection of the public, and which must therefore be given the widest possible scope, and a taxing Act.”

A case sometimes cited as an authority on the application of the “always speaking principle” to technological development is Royal College of Nursing of the United Kingdom v Department of Health and Social Security [1981] AC 800. The Abortion Act 1967 legalised abortion “when a pregnancy is terminated by a registered medical practitioner” in certain circumstances. After the Act was passed, a new technique was developed which involved inducing an abortion over a long period of up to 30 hours by administering a drug. The issue was whether the Department of Health was correct to advise that it was lawful for the drug to be administered by a nurse, acting on the instructions of a doctor. The House of Lords, by a 3-2 majority, held that it was and that the words “terminated by a registered medical practitioner” included a situation where the pregnancy was terminated by treatment carried out under a doctor’s control. All three members of the majority reached that conclusion by analysing the language and purpose of the Act and without any mention of any “always speaking principle”. A passage from the speech of Lord Wilberforce, who dissented, was quoted with approval in the Quintavalle case and is quoted by Lord Hamblen and Lord Burrows at para 32 above. That passage seems to me to be a helpful statement of how a court should approach any case where it is necessary to decide whether statutory language applies to a set of facts which its authors did not actually have (or are unlikely to have had or could not have had, it matters not) in contemplation. I do not read it as articulating anything which can usefully be described as an “always speaking doctrine”.

The notion that some special principle of interpretation is engaged in cases involving technological innovation may derive from an erroneous supposition that applying legislation to things which did not exist when the legislation was created must involve some change in its meaning. An example of such thinking would be the following reasoning. The meaning of the word “carriage” in the law prohibiting the furious driving of a carriage cannot have included a bicycle when the law was enacted because bicycles had not been invented at that time and the makers of the law could not have foreseen their existence. Therefore, if the term “carriage” as used in that statute is to be interpreted as applying to a bicycle, it is necessary to regard the statute as having changed its meaning.

Such thinking is fallacious because it fails to attend to the important distinction, which I have been emphasising, between the meaning of a word and its application. As Lord Bingham put it in R v G [2003] UKHL 50, [2004] 1 AC 1034, para 29: “the context or application of a statutory expression may change over time, but the meaning of the expression itself cannot change.” The distinction is one that has long been recognised in the philosophy of language, where it is classically described as the distinction between sense and reference. Understanding the sense (or interpreting the meaning) of a word enables us to classify objects but is not the same as designating the class of objects to which the word refers. For example, the general terms “creature with a heart” and “creature with kidneys” may refer to the same set of objects but they do not mean the same: see WV Quine, From A Logical Point of View (1980) p 21. As already discussed, it is in the very nature of general terms that they reach beyond any set of particular instances and are capable of applying to new instances not envisaged by their authors.

For these reasons, I do not consider that anything which can be described as an “always speaking doctrine” has a role to play in this case. There is no reason why the fact that digital editions did not exist when provision for the zero-rating of “newspapers” in the UK was originally made in 1972 (or when the VAT Act was enacted in 1994) should give rise to any presumption that the term should be read as including the digital editions after they were invented. When technological progress occurs, the proper approach is simply to ask in accordance with ordinary principles of interpretation whether the newly invented object falls within the meaning of the statutory language, interpreted in the light of the legislative purpose.

It is equally unsound to draw an opposite inference that, because digital editions did not exist when the legislation was enacted, they cannot fall within the meaning of the term “newspapers” as used in the legislation. To ask whether in 1972 (or 1994) the word “newspapers” would have been understood as referring to the digital editions is to ask the wrong question. The answer is obviously “no”, as at that time newspapers only existed in printed form. It follows that the term could not then refer to digital editions; but it does not follow that digital editions when they were invented could not fall within its meaning. In 1972 newspapers (at least in the UK) were exclusively printed using hot metal type, whereas today printed editions are produced by creating a computerised version of the whole publication which is then transferred to paper using an entirely different technology. That does not prevent modern printed editions from being classified as “newspapers”. Nor would it do so if, for example, they were no longer printed on paper but on some other, newly developed material.

As a matter of ordinary language, the word “newspapers” is capable of being used in a narrow sense which treats its physical, printed form as a defining characteristic. But it is also - and I think just as readily - capable of being used in a broader sense which includes digital editions. When tablet editions of The Times were first produced in 2010, somebody could quite naturally have said - and I am sure that many readers of the tablet edition did say - “I read my newspaper on a tablet.” The same applies to the other forms in which the digital editions were produced during the relevant period. Such use of the term did not, as it seems to me, involve giving the word “newspaper” a new meaning. If in 1972 a digital edition had been imagined by someone with the power to predict the future, it could just as well have been described then as a new form of “newspaper”. I therefore do not consider that the question whether the digital editions are “newspapers” can be answered simply by consulting the ordinary meaning of the word.

Nor can it be answered by a linguistic analysis of the words used to describe other items included in the same group as “newspapers” (Group 3 of Schedule 8 to the 1994 Act and its predecessors). The first item specified in Group 3, “books”, is subject to the same potential ambiguity as “newspapers”, in that the word is capable of being used in a sense which either includes or excludes books supplied in electronic format. The Court of Appeal thought it significant that Item 6 is defined as “Covers, cases and other articles supplied with Items 1 to 5 and not separately accounted for.” This indicates that Items 1-5 were all envisaged as including physical objects capable of being put into covers, cases and other similar articles. But it does not indicate that only physical objects were included in Items 1-5. Item 4 is defined as “Music (printed, duplicated or manuscript)”. The purpose of the limiting words was evidently to confine zero-rated supplies to sheet music expressed in musical notation and to exclude music in audio form. The term “duplicated”, however, seems to me capable of being read as including sheet music reproduced in digital form. Other points made by HMRC about the language used to identify items in Group 3 other than the term “newspapers” are equally inconclusive.

Whether the term “newspapers” encompasses the digital editions is not, therefore, a question which can be resolved by a linguistic analysis. To ascertain the meaning of the term it is necessary - as it always is - to consider the nature and purpose of the legislation in which the term is used.

Once the “always speaking principle” is put to one side, the central argument advanced by News Corp is based on the undisputed finding of the First-tier Tribunal (“FTT”) that the social policy which lay behind the UK’s decision to zero-rate “newspapers” (along with books and other items in Group 3) was “the promotion of literacy, the dissemination of knowledge and democratic accountability by having informed public debate”: see [2018] UKFTT 129 (TC), para 17. News Corp also relies on the FTT’s finding that the digital editions “were essentially … the same as or very similar to the newsprint editions” in terms of content: see para 155. Putting these two findings together, News Corp argues that the purpose of the statute in zero-rating “newspapers” is satisfied by the digital editions in the same way as it is by the printed editions. The method of delivery, which is the only material difference between them, is irrelevant to that purpose. It is submitted that in these circumstances there is no justification for interpreting the term “newspapers” as including only printed and not digital editions.

This argument was attractively presented by Mr Jonathan Peacock KC on behalf of News Corp. But I agree with the other members of the court that it cannot be accepted, essentially for a combination of two reasons. The first concerns the nature and purpose of the VAT legislation and the second concerns the potential significance of the medium in which the digital editions are supplied.

In explaining the importance of identifying the purpose of a statute, Lord Bingham said in the Quintavalle case, at para 8:

“Every statute other than a pure consolidating statute is, after all, enacted to make some change, or address some problem, or remove some blemish, or effect some improvement in the national life. The court’s task, within the permissible bounds of interpretation, is to give effect to Parliament’s purpose.”

If the statute which the court is being asked to construe had been a statute enacted for the purpose of promoting literacy, the dissemination of knowledge and democratic accountability by having informed public debate, then giving the term “newspapers” a broad interpretation which includes digital editions might well have been justified. The reasoning in Victor Chandler International Ltd v Customs and Excise Commissioners [2000] 1 WLR 1296, on which News Corp relies, might also then have been apposite.

The issue there was whether the term “advertisement” in section 9(1)(b) of the Betting and Gaming Duties Act 1981, which made it an offence to issue, circulate or distribute in the United Kingdom “any advertisement or other document” inviting the making of bets, applied to advertisements broadcast on Teletext for viewing on television screens in the United Kingdom. The Court of Appeal held that it did. They rejected an argument that the phrase “advertisement or other document” applied only to advertisements issued in hard copy rather than in electronic form on the ground that the clear purpose of the provision was to protect the revenue derived from betting duty by prohibiting offshore bookmakers from advertising in the United Kingdom for business; that advertisements transmitted in electronic form were squarely within the mischief at which the provision was aimed; and that the words used were capable of being read in a sense which gave effect to Parliament’s purpose.

Like the Betting and Gaming Duties Act 1981, the VAT Act is a taxing Act. It was enacted with the aim of raising revenue by taxing supplies of goods and services and doing so in way which complied with the European law governing VAT. Unlike the statutory provision in issue in the Victor Chandler case, however, the provision in issue here was not enacted in order to further the basic aims of the legislation. Rather, it is a provision which detracts from those aims. An important feature of the legal landscape is the requirement to interpret such a provision narrowly.

It is well settled that where the European VAT legislation provides for mandatory exemptions (for example, in articles 132 and 133 of the Principal VAT Directive), the terms used to specify the exemptions must be interpreted strictly. That is because they constitute exceptions to the general principle that VAT is to be levied on all goods and services supplied for consideration by a taxable person: see eg English Bridge Union Ltd v Revenue and Customs Commissioners (C-90/16) [2017] STC 2317, para 20 (holding that competitive bridge did not constitute a “sport” within the meaning of article 132(1)(m)); A & G Fahrschul-Akademie GmbH v Finanzamt Wolfenbüttel (C-449/17) EU:C:2019:202, para 19; SAE Education Ltd v Revenue and Customs Commissioners [2019] UKSC 14, [2019] STC 768, para 42.

Where, as in this case, the relevant exception is one created by national law, the rationale for interpreting it strictly is even greater. That is because such an exception derogates not only from the general principle that VAT is to be levied on all goods and services but also from the aim of harmonising the laws of member states and establishing a common system of VAT across the EU. Thus, in Talacre Beach Caravan Sales Ltd v Customs and Excis Commissioners(Case C-251/05) [2006] STC 1671 the Advocate General said that the national exceptions “depart in various respects from the system of the directive and must therefore be interpreted narrowly” (para 17) and that, because they lie outside the harmonised framework and are not directed at the same objectives as the exemptions mandated by the European directive, “it is necessary to take particular care that the exceptions are not extended” (para 42). The correctness of these statements cannot be doubted.

The cases (mentioned at para 106 above) holding that the exemptions mandated by the European VAT legislation must be interpreted strictly also say that the requirement of strict interpretation does not mean that the terms used to specify the exemptions should be construed “in such a way as to deprive them of their intended effect.” As this qualification is almost self-evident, I take it also to apply to the exceptions established by national law, provided their scope does not exceed the discretion afforded to member states by the European legislation. This begs the question, however, of what was the intended effect of the provision which zero-rated newspapers.

At all relevant times the UK could only zero-rate items if the decision to do so had been made “for clearly defined social reasons and for the benefit of the final consumer” (see paras 67-68 above). As elucidated by the European Court of Justice in Commission of the European Communities v United Kingdom (Case 416/85) [1990] 2 QB 130, paras 12-17, these conditions imposed only very weak constraints. The concept of “social reasons” embraced any measures introduced primarily for general social purposes and not principally for industrial, sectoral or fiscal reasons. Within those broad bounds, the determination of social policy was a matter of political choice for the member state. Zero-rating measures taken in pursuance of a member state’s social policy could only be challenged if it could be shown that the social reasons relied on were not sufficiently “clearly defined” or could not justify the measure or if the measure was totally disproportionate to the social reasons advanced: paras 12-14.

Thus, in 1972 the UK had a very broad discretion to decide which (if any) supplies of goods and services should be given preferential tax treatment through zero-rating. It was a matter of political choice for Parliament to assess what weight to give to diverse interests and contributions to social welfare across the full panoply of goods and services supplied to consumers. Once reduced rates became available as well as zero rates, the assessment became even more complex.

It is common ground that the social policy reasons which lay behind the decision to zero-rate the items described in what became Group 3 of Schedule 8 to the VAT Act were the promotion of literacy, the dissemination of knowledge and democratic accountability by having informed public debate (see para 101 above). It cannot be inferred, however, that Parliament’s intention was to zero-rate the supply of all goods and services which potentially advanced those policy goals for the benefit of the final consumer. Indeed, that was clearly not the case, as it is not suggested that Group 3 contains an exhaustive list of items which could be said to do so. Leading counsel for HMRC, Ms Eleni Mitrophanous KC, gave as examples of items which could be said to advance those goals, but were not zero-rated: reading cards, dictionary cards and services which provide continuously updated news.

Given the breadth of the discretion and the nature of the assessment, it is inevitable that the selection of particular items identified as qualifying for zero-rating was in a sense arbitrary - not in the sense that there were no reasons for the choices made, but in the weaker sense that the reasons did not compel one particular set of choices rather than another.

Ms Mitrophanous KC gave a number of examples of differences in tax treatment between items for which there is no clear rhyme or reason. For instance, oranges are zero-rated but juice squeezed from oranges is taxed at the standard rate of VAT; chocolate cakes are zero-rated but chocolate coated biscuits are taxed at the standard rate; during the relevant period women’s sanitary products were subject to a reduced rate of VAT but incontinence products were taxed at the standard rate. The pattern of the legislation approaches what Ronald Dworkin described as a “chequerboard” statute, which treats different cases differently without a clear principled reason for doing so: see Law’s Empire (1986) pp 178-184. With a statute of this kind the scope for purposive interpretation of the words used to ascertain their intended effect is extremely limited. There is not - and does not need to be - an overall logic or coherence in the items specified by Parliament which would permit reasoning by analogy.

An appreciation of the nature of the statutory provision zero-rating newspapers blunts the force of News Corp’s argument about the intended effect of the provision, but it is not a complete answer. It is one thing to recognise that the designation of zero-rated items is a legislative policy choice not governed by any intrinsic logic. It would be quite another to treat the selection of items as arbitrary in the sense of being irrational. In the absence of very clear language, it would be wrong to attribute to Parliament an intention to draw a distinction for which no plausible reason can be conceived. The requirement to interpret the exception for “newspapers” strictly would not justify interpreting the term as limited to printed editions and excluding digital editions if there were no rational basis for differentiating between them.

I was initially impressed by the argument made by News Corp that, given the finding of the FTT that the content of the digital and printed editions was fundamentally the same or very similar during the relevant period, there could be no sensible reason for treating them differently in terms of their VAT rating. The pivotal assumption of this argument is that the method by which that content was delivered to the consumer could not reasonably be regarded as relevant to the decision to zero-rate newspapers.

I have been persuaded, however, that this assumption is unsound and that the form as well as the content of the item supplied is a significant feature. It is not necessary to endorse Marshall McLuhan’s mantra that “the medium is the message” to recognise that the medium by which information is communicated is not something neutral but has far-reaching social implications and effects, albeit that they are easy to overlook because the content of a communication is the natural focus of attention. (McLuhan compared the content of a medium to “the juicy piece of meat carried by the burglar to distract the watchdog of the mind”: Understanding Media: The Extensions of Man (1964) p 19.) The revolutionary impact on human history of printing technology has been the subject of many studies: see eg Elizabeth Eisenstein, The Printing Press as an Agent of Change (1979). It is already apparent that digital media are having effects on society likely to be no less profound. Whether the concept of a newspaper will survive the digital revolution in any shape or form remains to be seen. I think it wrong to assume, however, that even a digital facsimile of a printed newspaper is functionally equivalent to a printed copy.

For a start, a digital edition is only accessible by a person who owns a computer or other electronic device (which is not itself a zero-rated item and must be acquired separately). While today the possession of such a device in the form of a mobile phone has become almost universal, during the relevant period that was far from being the case. The need for such a device had implications for the size and social profile of the class of consumers who stood to benefit from any decision to zero-rate digital editions. Another potentially relevant difference is that the supply of a digital edition is much less costly than the supply of printed editions because it does not involve the expenses of producing and distributing a physical item. At a more abstract level, for the purposes of VAT, the electronic delivery of a digital edition constitutes a supply of services whereas the delivery of a printed edition is a supply of goods. While the rules on VAT are in principle intended to tax the supply of goods and services in the same way, access to information which only exists in electronic form is a different type of consumption which offers different benefits to consumers from the supply of a physical object. For example, website editions viewed through an internet browser might be considered to have affinities with other information services provided over the internet which are as great as (or greater than) their affinities with newspapers in print form. In these circumstances a decision to treat electronically supplied services differently for VAT rating purposes from the supply of goods with the same or very similar content could not be regarded as irrational.

That conclusion is further supported by the decision of the Court of Justice of the European Union (Grand Chamber) in Rzecznik Praw Obywatelskich (RPO) v Marszalek Sejmu Rzeczypospolitej Polskiej (C-390/15) [2017] BVC 13. In that case the Grand Chamber rejected an argument that the provisions of the Principal VAT Directive which prohibited member states from applying a reduced rate of VAT to the supply of books electronically, while permitting them to apply a reduced rate to the supply of books on all physical means of support, offended against the EU law principle of equal treatment. The court approached the question on the footing that “when the EU legislature adopts a tax measure, it is called upon to make political, economic and social choices, and to rank divergent interests or to undertake complex assessments. Consequently, it should, in that context, be accorded a broad discretion …” (para 54). Judged from that perspective, the court held that the difference in VAT treatment was justified by the legitimate objective of establishing a separate set of rules for electronic services as a whole and that it was a proportionate means of pursuing that objective to treat the supply of digital books differently from physical books. In the same way, it seems to me that it was a permissible choice for a national legislature, in deciding which items to zero-rate, to limit the zero-rating of “newspapers” to physical items and not to include supplies in electronic form.

I conclude that it is not possible to say that the intended effect of adopting an exceptional tax treatment for “newspapers” must be taken to have included applying that treatment to the digital editions. Given the requirement to interpret the term strictly, the narrower possible meaning of the term “newspapers” must therefore be preferred, which treats printed form as a defining characteristic.

This is not to say that it was desirable or good social policy to tax digital editions differently from printed editions. Ultimately, indeed, the decision was taken - after such an extension was permitted by Council Directive (EU) 2018/ 1713 - that digital as well as printed editions should be zero-rated, and this has been the law since 1 May 2020. But that decision was one for Parliament to take subject to the constraints, for as long as they applied, of EU law. It was not a decision pre-empted by the zero-rating of “newspapers” in 1972.

A further type of change that may occur after a statute has been enacted, which raises different considerations again from the types of change that I have already discussed, is change in the law. When and for what reasons subsequent legislation can affect the meaning of an already existing statutory provision is a large question. There is no doubt that it sometimes can. In this case the various legislative changes (described at paras 68-70 above) made after “newspapers” were zero-rated do not in my view affect the correct interpretation of that term, with one exception.

This is the amendment of the VAT Act in 2020 to add, as a new Item 7 in Group 3 of Schedule 8, “the publications listed in Items 1 to 3 when supplied electronically”. The implication of this addition is that Items 1 to 3 do not themselves cover books, newspapers and periodicals when supplied electronically. This point could not of course have been made during the relevant period, as at that time Item 7 had not yet been added. There is, however, a line of cases which I considered and applied in R (on the application of ZYN) v Walsall Metropolitan Borough Council, paras 52-66, holding that an ambiguity in statutory language may be resolved by subsequent legislation: see eg Inland Revenue Commissioner v Hang Seng Bank Ltd [1991] 1 AC 306, 323-4 (and cases there cited). The ambiguity of the term “newspapers” in Item 2, Group 3, Schedule 8 is in my view, as discussed above, resolved by the requirement to interpret that provision strictly. But, had there been remaining ambiguity, it would have been legitimate to treat the meaning of the term as fixed by the subsequent amendment of the provisions of Group 3, just as it might have been fixed by a subsequent authoritative decision of a court.

For these reasons, I agree with Lord Hamblen and Lord Burrows that the term “newspapers” in Item 2, Group 3, Schedule 8 of the VAT Act does not include the digital editions, with the result that the supplies of the digital editions by News Corp during the relevant period were not zero-rated. I therefore concur in the decision that the appeal should be dismissed.