A number of recent tribunal decisions have added to our understanding of how discovery actually works. In J Anderson, not only must the officer who raises an assessment subjectively hold a belief, but that belief must, objectively, be a reasonable belief. In R Tooth, the Upper Tribunal accepts the proposition that a discovery can become stale, with an implied time limit between the discovery being made and the assessment being issued. Additionally, the fact that a discovery has been made does not mean that HMRC has an unfettered right to raise an assessment. In A Cooke, the available information exemption meant that the taxpayer was protected by TMA 1970 s 29(5), as the entries on the tax return were enough for the hypothetical officer to realise that there was an insufficiency of tax. However, the case of Y Blum shows that the protection under s 29(5) is not unlimited. And J Addo shows the central importance of understanding how the burden of proof operates in discovery cases. Gordon provides further guidance on current prevailing practice.
Last year, I wrote a detailed article looking at the state of the law relating to discovery (‘Discovery’, Tax Journal, 8 June 2017). Nothing stands still in tax, however. In the 12 months or so since that article was written, a number of further decisions in the courts and tribunals have added to our understanding of how this important, yet highly complex, part of the tax system actually works. Your editor has therefore asked me to discuss some of those recent developments.
Most of the previous cases in this area have not been about whether a discovery has actually been made; instead, they have considered whether or not HMRC is shut out from making a discovery assessment by TMA 1970 s 29(4) or (5). We will come back to those sections later, but first I want to cover some interesting developments about whether or not a discovery has actually been made.
There is a long discussion of what constitutes a discovery in J Anderson [2018] UKUT 159. This was a case about a loss relating to a trade of football development. HMRC did not enquire into the relevant return within the time limit and therefore sought to deny the loss by raising a discovery assessment. The first question was whether a discovery had actually been made.
The Upper Tribunal (UT), as is almost inevitable in discovery cases, reviewed the authorities going right back to Kensington Income Tax Commissioners [1913] 6 TC 279. From that line of authorities, it enumerated the following key principles, which have to be applied in determining whether a discovery has been made:
Therefore, the first question to be decided was whether, at the time the discovery assessment was issued, the actual officer who raised the assessment had a belief that there was an insufficiency of tax. A mere suspicion that there may be a loss of tax is not sufficient; there has to be a belief. The tribunal said that a suspicion becomes a belief when it is based on logical conclusions derived from what is known.
On the date that the assessment in Anderson was raised, the officer knew about the scheme, knew the taxpayer had participated in it, and also knew there were potential implementation issues. The officer who raised the assessment did not give evidence before the First-tier Tribunal (FTT), so there was no direct finding of fact as to her state of mind. However, the UT agreed with the FTT that her state of mind when raising the assessment could be inferred from what she did; i.e. raising the assessment. This is perhaps a slightly awkward way of dealing with things. We can expect in future discovery cases that the taxpayer’s counsel will want to establish as a matter of fact exactly what the officer’s state of mind was when raising the assessment.
What then of the objective element of the test? The tribunal said that this was akin to a public law issue. Not only must the officer subjectively hold a belief, but that belief must, objectively, be a reasonable belief. This is an important safeguard. Because s 29(1) looks at the position of an actual officer, that officer might genuinely believe there is an insufficiency of tax, even if nobody else did. The objective part of the test scrutinises the reasonableness of that belief. In this case, the UT had no doubt that the officer’s belief was reasonable.
This case is thus important for two reasons. First, it confirms that a discovery has to be based on more than a suspicion but doesn’t have to be based on conclusive evidence. There has to a belief which is based on a logical conclusion from what is known. Secondly, it clarifies that the question of belief has two elements: the subjective view of the actual officer making the assessment; and an objective view as to whether that view was reasonably held.
It is vital to remember that this test is only relevant for s 29(1), which relates to the actual officer who raised the assessment. It has no relevance to the test in s 29(5), which is concerned with a hypothetical officer.
Having upheld the discovery assessment in principle, the tribunal in Anderson had no problem in confirming the decision of the FTT that the loss relief was not due.
There is no time limit within s 29 itself: the normal assessing time limits apply. However, one of the issues which has emerged recently is whether there is any sort of implied time limitation between the discovery being made and the assessment itself being issued.
In Charlton [2012] UKUT 770, the Upper Tribunal said: ‘If an officer has concluded that a discovery assessment should be issued, but for some reason the assessment is not made within a reasonable period after that conclusion is reached, it might, depending on the circumstances, be the case that the conclusion would lose its essential newness by the time of the actual assessment.’
As I mentioned in my article last year, there was a divergence of opinion about whether this concept, which has been termed ‘staleness’, does in fact have any relevance to discovery. Staleness was one of the matters in contention in the Upper Tribunal in R Tooth [2018] UKUT 38. Mr Tooth participated in the same scheme that was subject to litigation in Cotter [2013] UKSC 69, which readers will recall concerns the interaction between claims into returns and claims into enquiries. Because of the way in which Mr Tooth had completed his return (which is discussed in more detail below), the Supreme Court decision in Cotter meant that HMRC had not used the correct enquiry powers in his particular case and therefore had to close its enquiry into the claim without any adjustment. However, HMRC then had a second bite of the cherry by raising a discovery assessment.
This is where the question of staleness came into play. HMRC had discovered (i.e. had an objective and subjective belief) that there was an insufficiency in Mr Tooth’s return in 2009, when it opened an enquiry into the claim. However, the discovery assessment was not issued until 2014. The FTT had found as a fact that the officer made a discovery in 2014 when raising the assessment, but if so, the Upper Tribunal thought, it was exactly the same discovery that had been made in 2009 when the enquiry was opened. The UT was very clear that the same discovery could not be made twice, even by two different HMRC officers. As the discovery had been made in 2009, then it was clearly stale by 2014.
As the tribunal said: ‘Here, HMRC discovered the insufficiency in 2009. It was incumbent on HMRC, at that stage, to decide what to do consequent upon this discovery. It was not open to HMRC without more to make a discovery of insufficiency and then to “park” the question of issuing a TMA s 29 assessment until a later date.’
Again, the picture is slightly muddled by a lack of specific evidence as to the exact basis on which the discovery assessment was raised. Had that been the only point at issue, the matter might have been remitted back to the FTT for a specific finding of fact. Because the UT found for the taxpayer on other grounds as well (see below), that was not necessary. But the key point is that the UT has accepted the proposition that a discovery can become stale. HMRC can’t simply sit on a discovery and wait until a later date to make an assessment. This could be particularly important in cases where HMRC only issues an assessment to protect time limits.
Further evidence that the tribunals have now accepted that concept of staleness comes in a very recent case Gordon and others v HMRC [2018]UKFTT 307. The case concerned pensions transfers and in particular whether a transfer to a particular arrangement gave rise to an unauthorised payment charge. HMRC considered that it did and raised a discovery assessment. The tribunal accepted that the conditions for a charge were met but did not uphold the assessment on the basis that the s 29(1) conditions were not met. The discovery assessments were issued in March 2014 but HMRC were aware of the insufficiencies perhaps as early as 2011 and certainly by the Spring of 2012. That was in the words of the tribunal ‘too long’ and as a result the discovery had become stale at the time that the assessments were issued and accordingly the assessments were squashed.
That of course leaves open the question of how long a gap can be permitted between the making of the discovery and the issue of the assessment. In Tooth the gap was five years and here it was two years. Both were held to be too long. In Tersam Gakhall TC/2015/02046 the delay was about three and a half months and that was held not to prevent a discovery assessment being valid. Whatever views might be on the rights and wrongs of a particular decision it seems to be highly unsatisfactory from a public policy point of view that there is an implicit but undefined time limit here which is not explicitly articulated. The VAT legislation does have a mechanism of dealing with this type of situation. As well as the normal four year assessing time limit there is a second limit. If the assessment is to be made more than two years after the end of the VAT quarter then HMRC must issue an assessment within 12 months of the date that evidence of the facts that justify making the assessment comes to HMRC knowledge. That test does have its own problems, but surely it would be better to have some time limit in the discovery legislation so that all parties know where they stand.
Gordon also deals with the question of burden of proof. I discuss that issue later in this article.
The fact that a discovery has been made does not, of course, mean that HMRC has an unfettered right to raise an assessment. Aside from the normal assessing time limits, three protections are available to the taxpayer: available information; conduct; and prevailing practice. The tribunals have shed new light on all three of these protections during the last year.
We can deal with the available information exemption first.
The test here is whether at the point that the enquiry window closed, an officer of the board ‘could not have been reasonably expected, on the basis of the information before him, to be aware of [the matter which had been discovered]’ (TMA 1970 s 29(5)).
It is now well established that, for the purposes of s 29(5), we are not dealing with the officer who actually raised the assessment but with a hypothetical officer. The case law discussed in my previous article has tried to clarify what characteristics and knowledge that hypothetical officer is deemed to have. Once those characteristics are established, the information actually available from the taxpayer’s return (and perhaps other material, see below) is considered in the light of that knowledge. Would that hypothetical officer have been expected to have been aware of the situation?
In Lansdowne Partners [2012] STC 544, the test was expressed as ‘whether the officer has the characteristics of an officer of general competence, knowledge or skill, which includes a reasonable knowledge and understanding of the law’. In Charlton, there was a slightly different nuance, in that the hypothetical officer was require to have the ‘reasonable knowledge and understanding that would be reasonably be expected in an officer considering the particular information provided by the taxpayer’ (emphasis added); in other words, the hypothetical officer could be assumed to have the specific knowledge to deal with the particular matter in tax return.
This test was considered in the case of A Cooke [2017] UKFTT 844. The issue was that the client’s tax return had been prepared by an agent using commercial software. Foreign dividends had been received and double tax relief (DTR) had been claimed in full, rather than restricted to the treaty amount. The software did not flag up the restriction and the agent did not appreciate that there was any restriction. The error was made in two successive returns. HMRC enquired into the latter but missed the deadline for the former and so raised a discovery assessment. HMRC needed no new information to enable the assessment to be raised; it simply restricted the DTR claim to the treaty amounts.
The question was therefore whether or not the hypothetical officer could reasonably have been expected to have been aware that further amounts of tax were due, purely on the basis of information in the return. The tribunal found that the officer should have been aware. DTR is very common and an officer of general competence and skill should, using the test in Lansdowne, have been aware that there were limits on the amount of relief which could be claimed. The hypothetical officer under the Charlton test would have been assumed to have some expertise in DTR and would obviously have been aware that something was wrong. Therefore, the tribunal found that the taxpayer was protected by s 29(5). Everything was clearly on the return and HMRC did not need any additional information to enable it to raise the assessment. The entries on the return themselves were enough for the hypothetical officer to realise that there was an insufficiency of tax.
HMRC raised the point that it operated a system of ‘process now and review later’ for SA returns; and that the earlier of the two returns was not examined manually by an officer until after the enquiry limit had expired. The tribunal gave this argument very short shrift: ‘It may well be that s 29(5) does not fit as well as HMRC might like with [its] current working practices, and in particularly the electronic processing of returns, but that is no basis for interpreting the provision in a way which does not accord with what it actually says.’
This is an important point. If something is clearly wrong on the face of the return and HMRC does not raise an enquiry in time, then in the absence of careless or deliberate conduct, it cannot raise a discovery assessment to correct the position.
The recent case of Y Blum [2018] UKFTT 152 shows that the protection under s 29(5) is not unlimited. In that case, the taxpayer had two employments and double counted an amount of PAYE paid when filling in her tax return. As a result, she understated her liability. This was not picked up on enquiry and HMRC subsequently sought to raise a discovery assessment to bring the additional tax into charge.
The taxpayer claimed that s 29(5) applied. The tribunal held that it did not. Although the hypothetical officer could be assumed to know that the employer would have made an end of year return showing the true amount of PAYE, this did not mean that the officer had to be treated as inferring that the information on that form was different to that on the return. This is a critical distinction. In Cooke, the insufficiency was obvious from the return; in Blum, there was nothing on the face of the return which pointed to an insufficiency. The hypothetical officer was not expected to cross check the return against other information available to HMRC.
There is another interesting point to emerge from the Blum case. Section 29(5) refers to information made available to the officer. Section 29(6) sets out what is meant by this phrase:
‘For the purposes of subsection (5) above, information is made available to an officer of the Board if—
a) it is contained in the taxpayer’s return under section 8 or 8A of this Act in respect of the relevant year of assessment (the return), or in any accounts, statements or documents accompanying the return;
b) it is contained in any claim made as regards the relevant year of assessment by the taxpayer acting in the same capacity as that in which he made the return, or in any accounts, statements or documents accompanying any such claim;
c) it is contained in any documents, accounts or particulars which, for the purposes of any enquiries into the return or any such claim by an officer of the Board, are produced or furnished by the taxpayer to the officer …; or
d) it is information the existence of which, and the relevance of which as regards the situation mentioned in subsection (1) above:
i. could reasonably be expected to be inferred by an officer of the Board from information falling within paragraphs (a) to (c) above; or
ii. are notified in writing by the taxpayer to an officer of the Board.’
HMRC argued that the employer’s return was not information provided by the taxpayer and therefore was not a document which was made available to an officer. The tribunal rejected this. Although sub-s 29(6)(d)(ii) did refer to information notified by the taxpayer, sub-s 29 (6)(d)(i) imposed no such restriction.
Information not supplied by the taxpayer could be taken into account if its existence could reasonably be inferred from other information provided by the taxpayer. Mrs Blum’s return had included employment pages and it was a reasonable inference that her employer would have made a return on form P14. It is surprising that HMRC ran this argument, because the UT in Charlton had previously interpreted the legislation in a way which did not support HMRC’s view.
This was a pyrrhic victory for Mrs Blum because she lost on the general s 29(5) point. Nevertheless, the tribunal’s interpretation of s 29(6) is worth bearing in mind should a similar situation arise.
The second protection is given by s 29(4). HMRC can raise a discovery assessment (regardless of the s 29(5) test) if the loss of tax is brought about carelessly or deliberately by the taxpayer or a person acting on his behalf. Or to put it another way – if the taxpayer has not been careless, HMRC cannot issue an assessment unless it can show that that s 29(5) does not protect the taxpayer.
This point was in issue in Cooke. We have already seen that the taxpayer had s 29(5) protection, so HMRC had to rely on carelessness in order to justify the assessment. The error was made by the accountant, who was clearly a person acting on the taxpayer’s behalf. So had the accountant been careless in the way that he had prepared the return? (There was never any suggestion that the behaviour had been deliberate.)
Carelessness is defined in TMA 1970 s 118(5) as a failure to take reasonable care. Most previous judgments, as the tribunal chair pointed out, have looked at the question of carelessness from the perspective of the taxpayer rather than the agent, but she thought that the same tests should apply when looking at the agent’s position. In particular, the test, though objective, is qualified by the particular circumstances of the case: ‘The question of reasonable care cannot be decided in a vacuum.’
In this particular case, the agent was a small general practice business with no dedicated tax partner. There were processes in place to check returns and the returns were produced on industry standard software. All of the correct figures were entered into the returns. However, the agent had not appreciated that there was a limit to the amount of relief which could be claimed: there was no flag in the software to indicate that there was a potential problem. So in all of the circumstances, the accountant could not be regarded as having failed to take reasonable care.
As a result, s 29(4) did not allow HMRC to raise a discovery assessment. As we have already seen, the taxpayer had s 29(5) protection and as a result the assessment was quashed.
What is interesting about Cooke is the different level of knowledge that each party argued that the other should have. As the judge put it: ‘This case presents something of a conundrum ... HMRC is arguing that the appellant’s accountant was careless in not identifying that the claims were excessive but at the same time an HMRC officer could not have been expected to pick the point up. The appellant argues the opposite – the accountant was not careless but an HMRC officer should have been able to spot the problem.’ As we have seen, the tribunal – quite reasonably in my opinion – thought that it was HMRC that should have had the higher level of knowledge.
The ambit of s 29(4) was also considered by the tribunal in Tooth. In order to understand the way that the tribunal analysed this part of the legislation, it is necessary to understand precisely what was included on the taxpayer’s return.
Mr Tooth entered into an avoidance scheme involving the creation of an employment loss. The transaction was undertaken in 2008/09 but the taxpayer carried back the loss to 2007/08. (The scheme was also the subject of the Cotter case and readers will recall that this concerned the mechanism under which loss carry backs should be returned and enquired into.) Mr Tooth sought to include the loss in his 2007/08 return, rather than through a separate claim. There was no mechanism for doing this using his accountants’ software (because the loss carry back does not actually affect the 2007/08 return), so Mr Tooth ‘forced’ the loss claim into his return by using the supplementary partnership pages. He made an extensive white space disclosure explaining exactly what he had done, in particular drawing attention to the fact that the loss was not a partnership loss but an employment loss. He also included the DOTAS number on his return.
The question before the UT was whether there was a deliberate inaccuracy in Mr Tooth’s return. If there was, that would enable a discovery assessment to be issued. This was formulated by the tribunal in the following way: ‘The question therefore arises as to whether an entry in a document that is explicitly based on a bona fide, albeit controversial interpretation of tax law, which subsequently proves to be wrong, can amount to an inaccuracy.’
The judge said that it couldn’t. He said: ‘An inaccuracy is something that is not accurate. Something is accurate if it confirms with the truth or with a given standard. In our judgment, where a taxpayer adopts a position in his return which, albeit controversial cannot (at the time of the return) be said to be wrong and takes the trouble to identify the position he has taken (and the fact that it is controversial) in that return cannot be guilty of an inaccuracy when, subsequently, it is established that the position taken by the taxpayer is wrong.’
The return may later turn out to be inaccurate, but the test has to be done at the time that the return is submitted.
Thus, the return was not inaccurate when it was made. Personally, I find this a somewhat surprising conclusion, given that it included relief for a loss which was not available, particularly when that loss was described as a partnership loss, which it manifestly wasn’t. My initial reaction was that the taxpayer was very fortunate and that the tribunal had taken an extremely benign view of how he had tried to ‘get round’ the system. I’ve spoken to other people who have taken a different view, though. Their perspective is that if Mr Tooth genuinely thought that the loss could be set off by being included in the 2007/08 return, then what he did was perfectly proper. It was not his fault that the software did not allow him to file according to his interpretation of the law without having to use the ‘wrong’ pages. Readers will no doubt form their own view.
Having found that there was no inaccuracy, the question of whether there was a deliberate inaccuracy did not apply. However, the tribunal did for completeness go on to consider this. It decided that it wasn’t deliberate: ‘The mere insertion of a figure into a document that is inaccurate may be a deliberate act but it is not necessarily a deliberate inaccuracy.’ HMRC needed to establish that there was a deliberate inaccuracy because otherwise it would have been out of time to raise the assessment. (It had missed the time limit for failure to take reasonable care.) As HMRC had failed to demonstrate that there was any inaccuracy, let alone a deliberate inaccuracy, the discovery assessment was quashed.
If this is decision is not reversed, it will be a useful protection for taxpayers. HMRC will always argue that any return which contains an entry which is not in accordance with law is de facto inaccurate. This decision shows that this is not necessarily correct.
The third defence against a discovery assessment is that there is an error or mistake in the basis on which a person’s liability ought to have been computed but the return was made on the basis of, or in accordance with, the practice generally prevailing at the time when it was made (s 29(2)).
This was an issue in the Gordon case referred to above. The taxpayers had transferred their pension funds to Wenns International Pension Scheme, which was a QROPS listed on its website and which had been given a reference number by HMRC. HMRC subsequently argued (and the tribunal agreed) that in fact the scheme did not meet the statutory definition of a QROPS. One of the taxpayers involved argued that because of the inclusion of Wenns fund on HMRC list of QROPS there was a prevailing practice that it was accepted as a qualifying fund and hence s 29(2) should apply to prevent HMRC from raising a discovery assessment.
The tribunal dismissed this argument. It said that the inclusion on the list was not relevant to the prevailing practice argument. It went on to say ‘if the point was not otherwise clear then the inclusion of the caveat on the QROPS list puts the point beyond doubt’. The caveat was a statement ‘Publication on this list should not be seen as confirmation by HMRC that it has verified all the information supplied ... if a scheme has been included on this list ... in circumstances where it did not satisfy the conditions ... any transfer that has been made to the scheme could potentially give rise to an unauthorised payment charge...’.
That must be right. For s 29(2) protection a taxpayer needs to rely on a clear unambiguous statement from HMRC.
The final case concerns burden of proof. In my previous article, I dealt with W Blumenthal [2012] UKFTT 497 and the difficulties which arise because of the shifting burden of proof. HMRC has the burden of proving that it has made a discovery. If it fails on this point, that is the end of the matter. But if it does meet that burden, then it is for the taxpayer to show that the assessment is wrong. If he/she does not meet that second burden of proof, the assessment is upheld.
J Addo [2018] UKFTT 93 is the second round in procedural litigation by this particular taxpayer. In the first instalment, the tribunal turned down her application to hold a separate hearing on the discovery issue and directed that the discovery issue and the substantive dispute should be heard together. This second instalment was about the procedure to be adopted at that hearing. HMRC did not object to the principle of presenting its case on discovery first. However, it was concerned about what would happen if, as anticipated, the taxpayer submitted that there was no case to answer after HMRC had set out its case. If the tribunal agreed, that would be the end of the matter; but what would happen if the tribunal concluded there was a case to answer, i.e. that there were grounds for raising a discovery assessment?
HMRC wanted the tribunal to give a direction that if Mrs Addo’s submission that there was no case to answer failed, she would be ‘put to her election’; i.e. that she would then not be able to argue about the underlying tax issue itself. In other words, she could not have her cake and eat it, by saying that there was no case to answer and then proceeding to answer that case.
The tribunal was not prepared to give such a direction and left it to the tribunal which would hear the ultimate appeal to determine what procedure to adopt if there was a no case to answer. This may appear to be an arcane dispute about dry points of procedure but it does show the central importance of understanding how the burden of proof operates in discovery cases.
Burden of proof was also an important element in the Gordon case. The judge reminded HMRC that it bore the burden of proof that a discovery had been made and that the discovery was not stale at the time that the assessment was raised. At the hearing itself HMRC produced very little evidence. The judge commented ‘In circumstances where it was clear that HMRC had identified an issue with Wenns by 2010 they needed to do something more to explain why no assessment was issued until March 2014 beyond simply saying ... that there was no evidence before the tribunal that a discovery was made before August 2013. The tribunal directed HMRC, after the hearing itself, to provide copies of any additional correspondence with the taxpayers. This was done and the judge described the additional correspondence as ‘illuminating’, which I think is a classic case of British understatement. Among the correspondence was a letter to one of the taxpayers dated March 2012 which actually referred to the transfer as an unauthorised payment. HMRC even attempted to correct the taxpayer’s return on that basis. In May 2012, HMRC made an unequivocal statment to one of the taxpayers that ‘HMRC had established that Wenns was not a legitimate pension scheme’. There were other documents as well, but the point is well made. HMRC had not produced evidence to discharge its burden of proof and when it was forced to produce that evidence by the tribunal it showed very clearly that that burden could not be discharged – quite the opposite. It is noticeable that the taxpayers were unrepresented in this case (indeed only one of them actually appeared before the tribunal in person) and it might have been the case that the normal process of disclosure of documents did not happen as rigorously as it would have done had the taxpayers been represented by counsel. Fortunately for them the judge was not prepared to let the point go and insisted on evidence being produced.
These two cases show just how important understanding the shifting burden of proof in discovery cases actually is. I suspect that we have not seen the last word on this difficult issue.
Whether by coincidence or design, the cases reported this year have covered most of the key problem areas of discovery. These include:
The law continues to evolve and there are important nuances in all of these cases. I would single out Tooth for special attention because of its approach to staleness. On the basis of that decision, I think that in future discovery cases it will be important that the taxpayer tests the HMRC officer who raised the assessment on what exactly what was discovered and when it was discovered. (Students of political history will recognise this as a variant of the famous quote by a member of the Watergate Senate Committee: ‘What did the president know and when did he know it?’)
Finally, is it notable that the taxpayer was victorious in three of these cases. Since what might be thought to be the high-water mark for taxpayers in Charlton, discovery cases have tended to be resolved in HMRC’s favour. Perhaps the pendulum is swinging. Every reader would accept that HMRC does need some powers to raise assessments to recover tax where the enquiry window has closed (most obviously in the case of deliberate understatement). However, there has been a widespread feeling in the tax community that the finality which the self-assessment system was intended to give was being undermined by the way in which HMRC was applying the discovery provisions. Perhaps we can now restore the balance to where most of us believe it should lie.
A number of recent tribunal decisions have added to our understanding of how discovery actually works. In J Anderson, not only must the officer who raises an assessment subjectively hold a belief, but that belief must, objectively, be a reasonable belief. In R Tooth, the Upper Tribunal accepts the proposition that a discovery can become stale, with an implied time limit between the discovery being made and the assessment being issued. Additionally, the fact that a discovery has been made does not mean that HMRC has an unfettered right to raise an assessment. In A Cooke, the available information exemption meant that the taxpayer was protected by TMA 1970 s 29(5), as the entries on the tax return were enough for the hypothetical officer to realise that there was an insufficiency of tax. However, the case of Y Blum shows that the protection under s 29(5) is not unlimited. And J Addo shows the central importance of understanding how the burden of proof operates in discovery cases. Gordon provides further guidance on current prevailing practice.
Last year, I wrote a detailed article looking at the state of the law relating to discovery (‘Discovery’, Tax Journal, 8 June 2017). Nothing stands still in tax, however. In the 12 months or so since that article was written, a number of further decisions in the courts and tribunals have added to our understanding of how this important, yet highly complex, part of the tax system actually works. Your editor has therefore asked me to discuss some of those recent developments.
Most of the previous cases in this area have not been about whether a discovery has actually been made; instead, they have considered whether or not HMRC is shut out from making a discovery assessment by TMA 1970 s 29(4) or (5). We will come back to those sections later, but first I want to cover some interesting developments about whether or not a discovery has actually been made.
There is a long discussion of what constitutes a discovery in J Anderson [2018] UKUT 159. This was a case about a loss relating to a trade of football development. HMRC did not enquire into the relevant return within the time limit and therefore sought to deny the loss by raising a discovery assessment. The first question was whether a discovery had actually been made.
The Upper Tribunal (UT), as is almost inevitable in discovery cases, reviewed the authorities going right back to Kensington Income Tax Commissioners [1913] 6 TC 279. From that line of authorities, it enumerated the following key principles, which have to be applied in determining whether a discovery has been made:
Therefore, the first question to be decided was whether, at the time the discovery assessment was issued, the actual officer who raised the assessment had a belief that there was an insufficiency of tax. A mere suspicion that there may be a loss of tax is not sufficient; there has to be a belief. The tribunal said that a suspicion becomes a belief when it is based on logical conclusions derived from what is known.
On the date that the assessment in Anderson was raised, the officer knew about the scheme, knew the taxpayer had participated in it, and also knew there were potential implementation issues. The officer who raised the assessment did not give evidence before the First-tier Tribunal (FTT), so there was no direct finding of fact as to her state of mind. However, the UT agreed with the FTT that her state of mind when raising the assessment could be inferred from what she did; i.e. raising the assessment. This is perhaps a slightly awkward way of dealing with things. We can expect in future discovery cases that the taxpayer’s counsel will want to establish as a matter of fact exactly what the officer’s state of mind was when raising the assessment.
What then of the objective element of the test? The tribunal said that this was akin to a public law issue. Not only must the officer subjectively hold a belief, but that belief must, objectively, be a reasonable belief. This is an important safeguard. Because s 29(1) looks at the position of an actual officer, that officer might genuinely believe there is an insufficiency of tax, even if nobody else did. The objective part of the test scrutinises the reasonableness of that belief. In this case, the UT had no doubt that the officer’s belief was reasonable.
This case is thus important for two reasons. First, it confirms that a discovery has to be based on more than a suspicion but doesn’t have to be based on conclusive evidence. There has to a belief which is based on a logical conclusion from what is known. Secondly, it clarifies that the question of belief has two elements: the subjective view of the actual officer making the assessment; and an objective view as to whether that view was reasonably held.
It is vital to remember that this test is only relevant for s 29(1), which relates to the actual officer who raised the assessment. It has no relevance to the test in s 29(5), which is concerned with a hypothetical officer.
Having upheld the discovery assessment in principle, the tribunal in Anderson had no problem in confirming the decision of the FTT that the loss relief was not due.
There is no time limit within s 29 itself: the normal assessing time limits apply. However, one of the issues which has emerged recently is whether there is any sort of implied time limitation between the discovery being made and the assessment itself being issued.
In Charlton [2012] UKUT 770, the Upper Tribunal said: ‘If an officer has concluded that a discovery assessment should be issued, but for some reason the assessment is not made within a reasonable period after that conclusion is reached, it might, depending on the circumstances, be the case that the conclusion would lose its essential newness by the time of the actual assessment.’
As I mentioned in my article last year, there was a divergence of opinion about whether this concept, which has been termed ‘staleness’, does in fact have any relevance to discovery. Staleness was one of the matters in contention in the Upper Tribunal in R Tooth [2018] UKUT 38. Mr Tooth participated in the same scheme that was subject to litigation in Cotter [2013] UKSC 69, which readers will recall concerns the interaction between claims into returns and claims into enquiries. Because of the way in which Mr Tooth had completed his return (which is discussed in more detail below), the Supreme Court decision in Cotter meant that HMRC had not used the correct enquiry powers in his particular case and therefore had to close its enquiry into the claim without any adjustment. However, HMRC then had a second bite of the cherry by raising a discovery assessment.
This is where the question of staleness came into play. HMRC had discovered (i.e. had an objective and subjective belief) that there was an insufficiency in Mr Tooth’s return in 2009, when it opened an enquiry into the claim. However, the discovery assessment was not issued until 2014. The FTT had found as a fact that the officer made a discovery in 2014 when raising the assessment, but if so, the Upper Tribunal thought, it was exactly the same discovery that had been made in 2009 when the enquiry was opened. The UT was very clear that the same discovery could not be made twice, even by two different HMRC officers. As the discovery had been made in 2009, then it was clearly stale by 2014.
As the tribunal said: ‘Here, HMRC discovered the insufficiency in 2009. It was incumbent on HMRC, at that stage, to decide what to do consequent upon this discovery. It was not open to HMRC without more to make a discovery of insufficiency and then to “park” the question of issuing a TMA s 29 assessment until a later date.’
Again, the picture is slightly muddled by a lack of specific evidence as to the exact basis on which the discovery assessment was raised. Had that been the only point at issue, the matter might have been remitted back to the FTT for a specific finding of fact. Because the UT found for the taxpayer on other grounds as well (see below), that was not necessary. But the key point is that the UT has accepted the proposition that a discovery can become stale. HMRC can’t simply sit on a discovery and wait until a later date to make an assessment. This could be particularly important in cases where HMRC only issues an assessment to protect time limits.
Further evidence that the tribunals have now accepted that concept of staleness comes in a very recent case Gordon and others v HMRC [2018]UKFTT 307. The case concerned pensions transfers and in particular whether a transfer to a particular arrangement gave rise to an unauthorised payment charge. HMRC considered that it did and raised a discovery assessment. The tribunal accepted that the conditions for a charge were met but did not uphold the assessment on the basis that the s 29(1) conditions were not met. The discovery assessments were issued in March 2014 but HMRC were aware of the insufficiencies perhaps as early as 2011 and certainly by the Spring of 2012. That was in the words of the tribunal ‘too long’ and as a result the discovery had become stale at the time that the assessments were issued and accordingly the assessments were squashed.
That of course leaves open the question of how long a gap can be permitted between the making of the discovery and the issue of the assessment. In Tooth the gap was five years and here it was two years. Both were held to be too long. In Tersam Gakhall TC/2015/02046 the delay was about three and a half months and that was held not to prevent a discovery assessment being valid. Whatever views might be on the rights and wrongs of a particular decision it seems to be highly unsatisfactory from a public policy point of view that there is an implicit but undefined time limit here which is not explicitly articulated. The VAT legislation does have a mechanism of dealing with this type of situation. As well as the normal four year assessing time limit there is a second limit. If the assessment is to be made more than two years after the end of the VAT quarter then HMRC must issue an assessment within 12 months of the date that evidence of the facts that justify making the assessment comes to HMRC knowledge. That test does have its own problems, but surely it would be better to have some time limit in the discovery legislation so that all parties know where they stand.
Gordon also deals with the question of burden of proof. I discuss that issue later in this article.
The fact that a discovery has been made does not, of course, mean that HMRC has an unfettered right to raise an assessment. Aside from the normal assessing time limits, three protections are available to the taxpayer: available information; conduct; and prevailing practice. The tribunals have shed new light on all three of these protections during the last year.
We can deal with the available information exemption first.
The test here is whether at the point that the enquiry window closed, an officer of the board ‘could not have been reasonably expected, on the basis of the information before him, to be aware of [the matter which had been discovered]’ (TMA 1970 s 29(5)).
It is now well established that, for the purposes of s 29(5), we are not dealing with the officer who actually raised the assessment but with a hypothetical officer. The case law discussed in my previous article has tried to clarify what characteristics and knowledge that hypothetical officer is deemed to have. Once those characteristics are established, the information actually available from the taxpayer’s return (and perhaps other material, see below) is considered in the light of that knowledge. Would that hypothetical officer have been expected to have been aware of the situation?
In Lansdowne Partners [2012] STC 544, the test was expressed as ‘whether the officer has the characteristics of an officer of general competence, knowledge or skill, which includes a reasonable knowledge and understanding of the law’. In Charlton, there was a slightly different nuance, in that the hypothetical officer was require to have the ‘reasonable knowledge and understanding that would be reasonably be expected in an officer considering the particular information provided by the taxpayer’ (emphasis added); in other words, the hypothetical officer could be assumed to have the specific knowledge to deal with the particular matter in tax return.
This test was considered in the case of A Cooke [2017] UKFTT 844. The issue was that the client’s tax return had been prepared by an agent using commercial software. Foreign dividends had been received and double tax relief (DTR) had been claimed in full, rather than restricted to the treaty amount. The software did not flag up the restriction and the agent did not appreciate that there was any restriction. The error was made in two successive returns. HMRC enquired into the latter but missed the deadline for the former and so raised a discovery assessment. HMRC needed no new information to enable the assessment to be raised; it simply restricted the DTR claim to the treaty amounts.
The question was therefore whether or not the hypothetical officer could reasonably have been expected to have been aware that further amounts of tax were due, purely on the basis of information in the return. The tribunal found that the officer should have been aware. DTR is very common and an officer of general competence and skill should, using the test in Lansdowne, have been aware that there were limits on the amount of relief which could be claimed. The hypothetical officer under the Charlton test would have been assumed to have some expertise in DTR and would obviously have been aware that something was wrong. Therefore, the tribunal found that the taxpayer was protected by s 29(5). Everything was clearly on the return and HMRC did not need any additional information to enable it to raise the assessment. The entries on the return themselves were enough for the hypothetical officer to realise that there was an insufficiency of tax.
HMRC raised the point that it operated a system of ‘process now and review later’ for SA returns; and that the earlier of the two returns was not examined manually by an officer until after the enquiry limit had expired. The tribunal gave this argument very short shrift: ‘It may well be that s 29(5) does not fit as well as HMRC might like with [its] current working practices, and in particularly the electronic processing of returns, but that is no basis for interpreting the provision in a way which does not accord with what it actually says.’
This is an important point. If something is clearly wrong on the face of the return and HMRC does not raise an enquiry in time, then in the absence of careless or deliberate conduct, it cannot raise a discovery assessment to correct the position.
The recent case of Y Blum [2018] UKFTT 152 shows that the protection under s 29(5) is not unlimited. In that case, the taxpayer had two employments and double counted an amount of PAYE paid when filling in her tax return. As a result, she understated her liability. This was not picked up on enquiry and HMRC subsequently sought to raise a discovery assessment to bring the additional tax into charge.
The taxpayer claimed that s 29(5) applied. The tribunal held that it did not. Although the hypothetical officer could be assumed to know that the employer would have made an end of year return showing the true amount of PAYE, this did not mean that the officer had to be treated as inferring that the information on that form was different to that on the return. This is a critical distinction. In Cooke, the insufficiency was obvious from the return; in Blum, there was nothing on the face of the return which pointed to an insufficiency. The hypothetical officer was not expected to cross check the return against other information available to HMRC.
There is another interesting point to emerge from the Blum case. Section 29(5) refers to information made available to the officer. Section 29(6) sets out what is meant by this phrase:
‘For the purposes of subsection (5) above, information is made available to an officer of the Board if—
a) it is contained in the taxpayer’s return under section 8 or 8A of this Act in respect of the relevant year of assessment (the return), or in any accounts, statements or documents accompanying the return;
b) it is contained in any claim made as regards the relevant year of assessment by the taxpayer acting in the same capacity as that in which he made the return, or in any accounts, statements or documents accompanying any such claim;
c) it is contained in any documents, accounts or particulars which, for the purposes of any enquiries into the return or any such claim by an officer of the Board, are produced or furnished by the taxpayer to the officer …; or
d) it is information the existence of which, and the relevance of which as regards the situation mentioned in subsection (1) above:
i. could reasonably be expected to be inferred by an officer of the Board from information falling within paragraphs (a) to (c) above; or
ii. are notified in writing by the taxpayer to an officer of the Board.’
HMRC argued that the employer’s return was not information provided by the taxpayer and therefore was not a document which was made available to an officer. The tribunal rejected this. Although sub-s 29(6)(d)(ii) did refer to information notified by the taxpayer, sub-s 29 (6)(d)(i) imposed no such restriction.
Information not supplied by the taxpayer could be taken into account if its existence could reasonably be inferred from other information provided by the taxpayer. Mrs Blum’s return had included employment pages and it was a reasonable inference that her employer would have made a return on form P14. It is surprising that HMRC ran this argument, because the UT in Charlton had previously interpreted the legislation in a way which did not support HMRC’s view.
This was a pyrrhic victory for Mrs Blum because she lost on the general s 29(5) point. Nevertheless, the tribunal’s interpretation of s 29(6) is worth bearing in mind should a similar situation arise.
The second protection is given by s 29(4). HMRC can raise a discovery assessment (regardless of the s 29(5) test) if the loss of tax is brought about carelessly or deliberately by the taxpayer or a person acting on his behalf. Or to put it another way – if the taxpayer has not been careless, HMRC cannot issue an assessment unless it can show that that s 29(5) does not protect the taxpayer.
This point was in issue in Cooke. We have already seen that the taxpayer had s 29(5) protection, so HMRC had to rely on carelessness in order to justify the assessment. The error was made by the accountant, who was clearly a person acting on the taxpayer’s behalf. So had the accountant been careless in the way that he had prepared the return? (There was never any suggestion that the behaviour had been deliberate.)
Carelessness is defined in TMA 1970 s 118(5) as a failure to take reasonable care. Most previous judgments, as the tribunal chair pointed out, have looked at the question of carelessness from the perspective of the taxpayer rather than the agent, but she thought that the same tests should apply when looking at the agent’s position. In particular, the test, though objective, is qualified by the particular circumstances of the case: ‘The question of reasonable care cannot be decided in a vacuum.’
In this particular case, the agent was a small general practice business with no dedicated tax partner. There were processes in place to check returns and the returns were produced on industry standard software. All of the correct figures were entered into the returns. However, the agent had not appreciated that there was a limit to the amount of relief which could be claimed: there was no flag in the software to indicate that there was a potential problem. So in all of the circumstances, the accountant could not be regarded as having failed to take reasonable care.
As a result, s 29(4) did not allow HMRC to raise a discovery assessment. As we have already seen, the taxpayer had s 29(5) protection and as a result the assessment was quashed.
What is interesting about Cooke is the different level of knowledge that each party argued that the other should have. As the judge put it: ‘This case presents something of a conundrum ... HMRC is arguing that the appellant’s accountant was careless in not identifying that the claims were excessive but at the same time an HMRC officer could not have been expected to pick the point up. The appellant argues the opposite – the accountant was not careless but an HMRC officer should have been able to spot the problem.’ As we have seen, the tribunal – quite reasonably in my opinion – thought that it was HMRC that should have had the higher level of knowledge.
The ambit of s 29(4) was also considered by the tribunal in Tooth. In order to understand the way that the tribunal analysed this part of the legislation, it is necessary to understand precisely what was included on the taxpayer’s return.
Mr Tooth entered into an avoidance scheme involving the creation of an employment loss. The transaction was undertaken in 2008/09 but the taxpayer carried back the loss to 2007/08. (The scheme was also the subject of the Cotter case and readers will recall that this concerned the mechanism under which loss carry backs should be returned and enquired into.) Mr Tooth sought to include the loss in his 2007/08 return, rather than through a separate claim. There was no mechanism for doing this using his accountants’ software (because the loss carry back does not actually affect the 2007/08 return), so Mr Tooth ‘forced’ the loss claim into his return by using the supplementary partnership pages. He made an extensive white space disclosure explaining exactly what he had done, in particular drawing attention to the fact that the loss was not a partnership loss but an employment loss. He also included the DOTAS number on his return.
The question before the UT was whether there was a deliberate inaccuracy in Mr Tooth’s return. If there was, that would enable a discovery assessment to be issued. This was formulated by the tribunal in the following way: ‘The question therefore arises as to whether an entry in a document that is explicitly based on a bona fide, albeit controversial interpretation of tax law, which subsequently proves to be wrong, can amount to an inaccuracy.’
The judge said that it couldn’t. He said: ‘An inaccuracy is something that is not accurate. Something is accurate if it confirms with the truth or with a given standard. In our judgment, where a taxpayer adopts a position in his return which, albeit controversial cannot (at the time of the return) be said to be wrong and takes the trouble to identify the position he has taken (and the fact that it is controversial) in that return cannot be guilty of an inaccuracy when, subsequently, it is established that the position taken by the taxpayer is wrong.’
The return may later turn out to be inaccurate, but the test has to be done at the time that the return is submitted.
Thus, the return was not inaccurate when it was made. Personally, I find this a somewhat surprising conclusion, given that it included relief for a loss which was not available, particularly when that loss was described as a partnership loss, which it manifestly wasn’t. My initial reaction was that the taxpayer was very fortunate and that the tribunal had taken an extremely benign view of how he had tried to ‘get round’ the system. I’ve spoken to other people who have taken a different view, though. Their perspective is that if Mr Tooth genuinely thought that the loss could be set off by being included in the 2007/08 return, then what he did was perfectly proper. It was not his fault that the software did not allow him to file according to his interpretation of the law without having to use the ‘wrong’ pages. Readers will no doubt form their own view.
Having found that there was no inaccuracy, the question of whether there was a deliberate inaccuracy did not apply. However, the tribunal did for completeness go on to consider this. It decided that it wasn’t deliberate: ‘The mere insertion of a figure into a document that is inaccurate may be a deliberate act but it is not necessarily a deliberate inaccuracy.’ HMRC needed to establish that there was a deliberate inaccuracy because otherwise it would have been out of time to raise the assessment. (It had missed the time limit for failure to take reasonable care.) As HMRC had failed to demonstrate that there was any inaccuracy, let alone a deliberate inaccuracy, the discovery assessment was quashed.
If this is decision is not reversed, it will be a useful protection for taxpayers. HMRC will always argue that any return which contains an entry which is not in accordance with law is de facto inaccurate. This decision shows that this is not necessarily correct.
The third defence against a discovery assessment is that there is an error or mistake in the basis on which a person’s liability ought to have been computed but the return was made on the basis of, or in accordance with, the practice generally prevailing at the time when it was made (s 29(2)).
This was an issue in the Gordon case referred to above. The taxpayers had transferred their pension funds to Wenns International Pension Scheme, which was a QROPS listed on its website and which had been given a reference number by HMRC. HMRC subsequently argued (and the tribunal agreed) that in fact the scheme did not meet the statutory definition of a QROPS. One of the taxpayers involved argued that because of the inclusion of Wenns fund on HMRC list of QROPS there was a prevailing practice that it was accepted as a qualifying fund and hence s 29(2) should apply to prevent HMRC from raising a discovery assessment.
The tribunal dismissed this argument. It said that the inclusion on the list was not relevant to the prevailing practice argument. It went on to say ‘if the point was not otherwise clear then the inclusion of the caveat on the QROPS list puts the point beyond doubt’. The caveat was a statement ‘Publication on this list should not be seen as confirmation by HMRC that it has verified all the information supplied ... if a scheme has been included on this list ... in circumstances where it did not satisfy the conditions ... any transfer that has been made to the scheme could potentially give rise to an unauthorised payment charge...’.
That must be right. For s 29(2) protection a taxpayer needs to rely on a clear unambiguous statement from HMRC.
The final case concerns burden of proof. In my previous article, I dealt with W Blumenthal [2012] UKFTT 497 and the difficulties which arise because of the shifting burden of proof. HMRC has the burden of proving that it has made a discovery. If it fails on this point, that is the end of the matter. But if it does meet that burden, then it is for the taxpayer to show that the assessment is wrong. If he/she does not meet that second burden of proof, the assessment is upheld.
J Addo [2018] UKFTT 93 is the second round in procedural litigation by this particular taxpayer. In the first instalment, the tribunal turned down her application to hold a separate hearing on the discovery issue and directed that the discovery issue and the substantive dispute should be heard together. This second instalment was about the procedure to be adopted at that hearing. HMRC did not object to the principle of presenting its case on discovery first. However, it was concerned about what would happen if, as anticipated, the taxpayer submitted that there was no case to answer after HMRC had set out its case. If the tribunal agreed, that would be the end of the matter; but what would happen if the tribunal concluded there was a case to answer, i.e. that there were grounds for raising a discovery assessment?
HMRC wanted the tribunal to give a direction that if Mrs Addo’s submission that there was no case to answer failed, she would be ‘put to her election’; i.e. that she would then not be able to argue about the underlying tax issue itself. In other words, she could not have her cake and eat it, by saying that there was no case to answer and then proceeding to answer that case.
The tribunal was not prepared to give such a direction and left it to the tribunal which would hear the ultimate appeal to determine what procedure to adopt if there was a no case to answer. This may appear to be an arcane dispute about dry points of procedure but it does show the central importance of understanding how the burden of proof operates in discovery cases.
Burden of proof was also an important element in the Gordon case. The judge reminded HMRC that it bore the burden of proof that a discovery had been made and that the discovery was not stale at the time that the assessment was raised. At the hearing itself HMRC produced very little evidence. The judge commented ‘In circumstances where it was clear that HMRC had identified an issue with Wenns by 2010 they needed to do something more to explain why no assessment was issued until March 2014 beyond simply saying ... that there was no evidence before the tribunal that a discovery was made before August 2013. The tribunal directed HMRC, after the hearing itself, to provide copies of any additional correspondence with the taxpayers. This was done and the judge described the additional correspondence as ‘illuminating’, which I think is a classic case of British understatement. Among the correspondence was a letter to one of the taxpayers dated March 2012 which actually referred to the transfer as an unauthorised payment. HMRC even attempted to correct the taxpayer’s return on that basis. In May 2012, HMRC made an unequivocal statment to one of the taxpayers that ‘HMRC had established that Wenns was not a legitimate pension scheme’. There were other documents as well, but the point is well made. HMRC had not produced evidence to discharge its burden of proof and when it was forced to produce that evidence by the tribunal it showed very clearly that that burden could not be discharged – quite the opposite. It is noticeable that the taxpayers were unrepresented in this case (indeed only one of them actually appeared before the tribunal in person) and it might have been the case that the normal process of disclosure of documents did not happen as rigorously as it would have done had the taxpayers been represented by counsel. Fortunately for them the judge was not prepared to let the point go and insisted on evidence being produced.
These two cases show just how important understanding the shifting burden of proof in discovery cases actually is. I suspect that we have not seen the last word on this difficult issue.
Whether by coincidence or design, the cases reported this year have covered most of the key problem areas of discovery. These include:
The law continues to evolve and there are important nuances in all of these cases. I would single out Tooth for special attention because of its approach to staleness. On the basis of that decision, I think that in future discovery cases it will be important that the taxpayer tests the HMRC officer who raised the assessment on what exactly what was discovered and when it was discovered. (Students of political history will recognise this as a variant of the famous quote by a member of the Watergate Senate Committee: ‘What did the president know and when did he know it?’)
Finally, is it notable that the taxpayer was victorious in three of these cases. Since what might be thought to be the high-water mark for taxpayers in Charlton, discovery cases have tended to be resolved in HMRC’s favour. Perhaps the pendulum is swinging. Every reader would accept that HMRC does need some powers to raise assessments to recover tax where the enquiry window has closed (most obviously in the case of deliberate understatement). However, there has been a widespread feeling in the tax community that the finality which the self-assessment system was intended to give was being undermined by the way in which HMRC was applying the discovery provisions. Perhaps we can now restore the balance to where most of us believe it should lie.
