Lessex Ltd v HM Inspector of Taxes

This is an appeal by the taxpayer company (the Company) against the imposition of a penalty of £1000 under the Finance Act 1998, Schedule 18, paragraph 17(3) (paragraph 17).

The only relevant facts are agreed. The Company filed a company tax return more than 3 months late. It had also filed the two previous returns more than 3 months late. Penalties of £200 were imposed on the company in respect of the late filing for the two previous accounting periods. The penalties laid down in paragraph 17 are flat-rate with no discretion as to the level to be imposed. It was agreed that there were no “merits” issues to be considered about the individual company or its conduct.

The parties also agreed that the question for me to determine in the appeal is whether the appropriate level of the late filing penalty to be determined in relation to the late delivery by the Company of its corporation tax return on those facts should be £1000 as contended by the respondents, applying paragraph 17. It is agreed that the only other decision on the legislation is a penalty of £200. There are no other answers.

That question should be straightforward. It is not. The reason is that the three successive late returns straddle the transfer of the corporation tax system from pay-and-file to self-assessment. The first of the late returns was liable to a penalty under the pay-and-file provisions. It was imposed under section 94 of the Taxes Management Act 1970. The third of the late returns was made after the self-assessment appointed day and was therefore under the self-assessment provisions of Schedule 18 to the 1998 Act. The penalty is imposed under paragraph 17.

It was not disputed by Mr Gordon for the Company that had the Company defaulted three successive times under the pay-and file system then the penalty for the third period under section 94 would be £1000. Nor was it disputed that if all three defaults had occurred under the self-assessment system then the penalty for the third default under paragraph 17 would also be £1000. He contended, however, that during the transition the higher penalty for successive late filings did not apply, and the maximum penalty that the respondent could impose was £200. This was because paragraph 17 contained faulty drafting such that it failed to make effective transitional provision for the higher penalty. I was told that this was not a test case but that it would affect other taxpayers.

The main part of Paragraph 17 provides:

Where all the accounting periods end after the self-assessment appointed day, these provisions apply. They set out a simple multi-level penalty scheme differentiating between returns that are late, perhaps only marginally, and those that are significantly late (three months) and also returns that are late on one or two occasions, and those that are repeatedly late (three or more successive occasions). The highest penalty is reserved for those that are both repeatedly and significantly late (£1000).

Where all the accounting periods ended before the self-assessment appointed day (that is, they were under the pay and file regime and made under section 11 of the Taxes Management Act 1970 in its then form), there was a directly equivalent provision in section 94 of the Taxes Management Act 1970. That provided the same multi-level flat-rate penalty scheme in directly equivalent circumstances. While the wording of section 94 is different to that of paragraph 17, nothing turns on this and I do not need to consider the differences.

To achieve a smooth transition, there needs to be a transitional provision to deal with a company that makes repeated late returns both under the pay and file system and under the self assessment system. At the extreme is the company that is always significantly late with its return. At the margin is the company that is late once or twice under the pay and file regime but is then late a third, or second and third, time under the self-assessment system.

The drafter provided for the transition by the following additional sub-paragraph to paragraph 17:

If we follow the instructions of sub-paragraph (4)(b) literally, then sub-paragraph (3) reads as follows:

That literal reading clearly does not achieve its intention. Instead it resurrects a repeated reference to section 94, now repealed; and it fails to deal directly with the transition. There cannot be a penalty under section 94 for periods after the self-assessment appointed day. So sub-paragraph (3)(d), as amended by sub-paragraph (4)(b) to read as above, can never apply to any company. Put simply, the inclusion of the words “and (d)” in sub-paragraph (4)(b) create a nonsense. It is a case where, as Horace said of Homer, “Indignor quandoque bonus dormitat” (in Wickhams’s translation: but if Homer, usually good, nods, I think it shame). Or, in the vernacular, it is a rare example of parliamentary counsel getting it unambiguously wrong.

Before considering the effect of this error, it is appropriate to define its extent more precisely. This is because the parties took rather different views of the nature of the error.

the company is again liable to a penalty under this paragraph in respect of the third period.

But, as Mr Gordon also showed, there are other problems. There is an internal conflict between the opening words of sub-paragraph (4) and sub-paragraph (4)(b) . The opening words apply the sub-paragraph to sub-paragraph (3) if and only if “the first or second period” is (italics mine) a pre-self-assessment period. That is overcompacted wording which, read literally, may also entirely fail. If the third period is after the self-assessment appointed day, then the second period cannot be under section 94 unless the first period is also under that section. This could be handled by reading the “or” as “and”. But that creates another problem, because it then excludes the reference to the first period only. That might again be dealt with by reading the wording as applying to the first period and also where relevant the second period. But that reveals a further underlying error in the logic of sub-paragraph (4)(b).

A reference to the “first period” only does not fit with the precise instruction in sub-paragraph (4)(b). The amendments prescribed in sub-paragraph (4)(b) leave sub-paragraph (3)(c) applying only to a third period, not to second and third periods. In other words, the opening words of sub-paragraph (4), if read broadly rather than literally as a precondition for the application of sub-paragraph (4)(b), lead to the result that it would apply in cases where returns for both the first and second periods are penalised under section 94, but not where only the first period only is penalised under that section.

The conclusion is that the opening words of sub-paragraph (4) are open to an interpretation that effectively raises the penalty to £1000 in the case where a company has defaulted on three successive occasions, two of which are under section 94 and the third under paragraph 17, but not the case where the first is under section 94 and the second and third both under paragraph 17. Such an interpretation is in my view patently absurd. It would raise the penalty to £1000 in the case of a company that makes a first late filing under self-assessment after two or more penalties for late filing under section 94, but prevent the penalty being more than £200 if the same company then made a fourth late filing the following year! That is the reverse of the obvious purpose of the provision. Whatever interpretation is put on paragraph 17(4), it must in my view apply equally to the case where there is no prior penalty under paragraph 17 and the case where there is one prior penalty under that Paragraph.

If that result is to be achieved, it requires a broader re-reading of paragraph 17(4) than that suggested by Mr Hill. It is not enough just to ignore the signposting “and (d)” and the inappropriate “again”. Some broad sense must also be made of the initial wording of sub-paragraph (4) and the failure properly to amend sub-paragraph (3)(c).

In my judgment, this means that a “proper” interpretation of paragraph 17 has to adopt one of two courses. It could recognise not only that there is an error in the signposting and drafting of sub-paragraph (4)(b) but also that the existing wording of the opening condition of the sub-paragraph covers only some of the possible cases to which the sub-paragraph should apply, omitting others. So the interpretation could adopt the broad view that the intention of the paragraph as a whole is obviously that the same multi-rate penalty regime should apply to all transitional cases as it does to all cases before and after transition. It must therefore repair the omissions as well as omit or adjust the commissions.

Alternatively, the interpretation could recognise that the sub-paragraph is so faulty that it should be taken at its face value and discounted completely as a provision that entirely fails to impose the higher penalty in any case because its internal consistencies are such that it cannot actually apply to any transitional case at all. I can see no rational basis for an intermediate interpretation that applies a different regime to a company that makes late returns two or more times under the pay-and-file regime and then a third time under the self-assessment system as against another company – or the same company – making a second late return under the self-assessment system after two (or, in the case of the same company, three) previous consecutive late returns.

So the question is whether I interpret sub-paragraph (4) literally, conclude that it does not work and ignore it in its entirely, or whether I interpret it as a transparently and obviously faulty attempt to achieve a transparent and obvious objective, to be read with that objective in mind rather than by reference to the exact wording.

Both parties presented me with elaborate and sustained arguments about interpretation to substantiate their two diametrically opposed views. They based themselves in large part on differing presentations of well-rehearsed arguments about how to interpret ambiguous legislation. But, with respect to both of them, there is a difference between the errors in paragraph 17 and those in most of the cases to which my attention was drawn. As I put to both of them at the hearing, there can be no ambiguity in the formula “and (d)”. It cannot have a wider or a narrower meaning, nor can I turn to a dictionary to help me. It is a signpost. The question is not how I interpret the signpost but whether I follow it simply because it is there, or whether I recognise that the signpost is obviously pointing in the wrong direction and ignore it. Further, there is no ambiguity about the failure of the opening wording of the sub-paragraph to link with sub-paragraph (4)(b). The inconsistencies are logical rather than semantic. And the need to read the amendment to sub-paragraph (3)(c) broadly is a logical point, not a semantic one.

Before I turn to the approach I think it right to take, I record that it is accepted by both sides that the wording of sub-paragraph 17(4) is patently badly drafted. And there can be no doubt about what is intended. The penalties before and after the transition are, so far is possible, identical. Further, they remain identical through the transition to the occasional late filer. Literal and contextual readings of sub-paragraph (4) create a penalty reduction, as compared with the penalties before and after the transition, only for the persistent late filer; though, depending on how it is read, this may be all persistent late filers or only some of them and it may apply for only one of the transition years or both. I also record that because both the purpose and the errors are so clear, this is not a case for looking at any parliamentary record. As this provision is in a Schedule, not a section, there is unlikely to be any relevant record but even if there were I would not consider it of assistance.

Faced with the potential nonsense caused by these drafting errors, Mr Gordon strongly argued that I should treat paragraph 17(4) as it stands. It is literally a nonsense and I should ignore it. Both because it is a tax law, and because it is a penalty provision – and even more because it is a penalty provision in a tax law - I must give the taxpayer the benefit of any doubt, and I must take the words at face value even if, as a result, they do not apply to anything. The nonsense was created by the legislation, not by any interpretation of it, and the taxpayer is entitled to the benefit of that nonsense.

Mr Gordon supported his approach by a formidable array of authorities.

“First, the words are to be given their ordinary meaning”. That I fully accept. I have taken Mr Gordon’s view about the “ordinary” meaning of the sub-paragraph in my analysis of the faults in the drafting. I agree with him that the opening words of the sub-paragraph must be given full weight in relation to the operative provisions that follow. Where I disagree is with his conclusion that the ordinary meaning leaves the sub-paragraph with a narrow focus. I find myself pushed to the conclusion that the “ordinary” meaning is no meaning at all. In other words, a proper analysis of the language shows that the errors in it are such as to deprive it of an “ordinary meaning” when read literally as a whole. The errors here are not a matter of the meaning of the words used. There is no ambiguity in “again” or “third”, even if there is in “or”. And there is no direct (rather than referential) meaning at all in “and (d)”. Rather, it is a matter, as noted above, of faulty signposting combined with a failure to deal with the matter logically, whatever meaning is given to the individual words or to their combinations. Nor for the same reason am I assisted in this case by Lord Donavan’s warning that “moral precepts are not applicable to the interpretation of revenue statutes”, which again I fully accept. It was agreed that there are no “merits” issues in this case.

“Secondly … one has to look merely at what is clearly said. There is no room for any intendment. There is no equity about a tax. There is no presumption as to tax. Nothing is to be read in, nothing is to be implied. One can only look fairly at the language used.” That is the famous aphorism of Rowlatt J. Despite the endorsement of Lord Donovan in Mangin, I am not sure how far this survives (even if limited to the direct taxes) unaffected by the shift to a more contextual approach in taxation of recent decades. As Mr Gordon accepted, this approach was revisited by Lord Wilberforce in Ramsay v IRC [1982] AC 300. In particular, Lord Wilberforce emphasised that:

“what are “clear words” is to be ascertained upon normal principles: these do not confine the courts to literal interpretations. There may, indeed should, be considered the context and scheme of the relevant Act as a whole, and its purpose may, indeed should, be regarded.”

But in this case, again with great respect to guidance I fully accept, it is of limited assistance because the words are clear – it is their interrelation that is faulty. What is “clearly” said as a whole does not make logical sense because it is both internally inconsistent and referentially inconsistent.

“Thirdly, the object of the construction of a statute being to ascertain the will of the legislature it may be presumed that neither injustice nor absurdity was intended. If therefore a literal interpretation would produce such a result, and the language admits of an interpretation would avoid it, then such an interpretation may be adopted.” Mr Gordon argued that the cancellation of the third penalty by a literal reading of the words of the sub-paragraph did not create an absurdity. It was simply a failure of the provision. Alternatively, if it is an absurdity that the provision fails to impose the higher penalty on the taxpayer for the year in question, then it is not an absurdity that can be dealt with by a proper construction of the language. I have already accepted one aspect of this submission, though not in the way Mr Gordon put it to me. I regard it as absurd in this sense if the only literal construction leads to the situation that the Company is liable to a higher penalty for the first time it is late with its self-assessment return but not (if that were to be the case) the second time.

Mr Gordon further cautioned me against a forced construction in the words of Lord Reid in Luke v IRC [1963] AC 557. When faced with the problem of having to “do some violence to the words” to “achieve the obvious intention”, he cautioned:

“It is only when the words are absolutely incapable of a construction that will accord with the apparent intention of the provision and will avoid a wholly unreasonable result, that the words of the enactment will prevail.”

This, he submitted, was just such a case and the words should prevail.

“Fourth, the history of an enactment and the reasons which led to its being passed may be used as an aid to construction.” Mr Gordon argued that this did not assist me. I have already noted that the purpose of the provision is transparent both from the rest of the wording in paragraph 17 and from its necessary cross-reference to section 94, and that any parliamentary history is of no assistance. I agree for this reason that this approach takes me no further in this case. The context and purpose are already there in paragraph 17, and I need look no further.

Mr Gordon then argued that his conclusion was supported, in the alternative the same approach should be derived, by reference to the approach taken by judges to penalty provisions. In particular, he referred to the “principle against doubtful penalisation” from Part XVII of Bennion’s Statutory Interpretation: A Code (4^th edition). I indicated at the hearing that I fully accepted that principle without argument, and Mr Hill also indicated that he accepted the approach. It was also agreed that this case was not one where the penalty was within Article 6 paragraph 3 of the European Convention on Human Rights. Nonetheless it was, Mr Gordon stressed, a penalty. The respondent had no right to demand £1000 on this legislation – only £200.

Finally, he urged me, in balancing all the factors to be taken into account, to put the main weight on two aspects of the case. First, I should adopt the literal meaning in preference to others because, again adopting Bennion, “the literal meaning weighs heavily.” Secondly, justice required that I accept the taxpayer’s argument or, rather, not to do so would be injustice. Both went against a tidying up of the wording.

Mr Hill accepted that sub-paragraph 17(4) was faulty. In particular, the reference to “and (d)” was an obvious mistake. My task, he submitted, was to make the paragraph work while doing the least damage to its language. Faced with the alternatives of striking out “and (d)” and striking out the whole of sub-paragraph 17(4) I should take the narrower approach of ignoring only the faulty signposting. It is obvious that sub-paragraph (4) is a straddle provision – it can have no other focus. So it should not be interpreted to knock out that effect. Likewise, the purpose of the sub-paragraph is obvious. This would also apply to the faulty drafting of the opening words. To interpret it to negate it completely would frustrate that obvious purpose entirely. So the question is, what can be done about the errors?

Mr Hill called in aid another principle of Bennion’s Statutory Interpretation. This is the presumption of statutory construction that the legislator intends the court to apply a construction which rectifies any error in the drafting of the enactment, where it is required in order to give effect to the legislator’s intention (section 287 of the Code).

He took me to authorities that showed that the judiciary could and, in appropriate cases, should both read words in where that was necessary and also remove or ignore them. He drew in particular on Sharp v Perreira [1998] 4 All ER 145 as an example of reading words in and AG v Hislop [1991] 1 All ER 911 and the old case of Stone v Yeovil (1876) LR 1 CP 691, 701 as examples of ignoring them. The latter case was particularly of relevance here because it was one of ignoring a faulty cross-reference. The opinion of Lord Bridge in the more recent case of McMonagle v Westminster City Council [1990] 2 AC 716 was strong authority for the issue underlying these approaches. Judges should not simply leave it to the legislature to sort out obvious errors.

With regard to the principles of construction, this was clearly an example for using what used to be called the “mischief rule”. It was the clearest of examples of Lord Donovan’s third principle in Mangin. The reason why sub-paragraph (4) was in the statute was obvious, and effect should be given to it. This was a “familiar” principle (to quote from Lord Wilberforce in Ramsay). Further, effect was to be given to it without any special significance attaching to the fact that the provision appeared in a tax statute. Lord Wilberforce had laid to rest the old idea that there was something special about interpreting provisions in taxing Acts. Nor was there anything in Luke to suggest that I should take the literal approach here. Lord Reid’s view was that of the minority in the case. He was in any event looking to the purpose of the provision in question. That was clear in this case.

Turning to the presumptions, he agreed that there was a presumption against applying a doubtful penalty provision and that this applied to civil as well as to criminal penalties. But in his view that had no effect in this case. It only arose as a last resort when other approaches to resolving a problem had failed. It is effectively a tie breaker, and the tie does not need breaking in this case. In any event, this is not