Incompetent parties, gullible Tribunal

Failure to notify - a new twist (or twits)

Didn't find your answer?

Finola Owens v HMRC https://financeandtax.decisions.tribunals.gov.uk/judgmentfiles/j12997/TC%2009098.pdf is a rather odd decision from the FTT, in relation to penalties for failure to notify, as there is an unusual dispute between the parties about whether the taxpayer did in fact notify chargeability.   The case also involved discovery assessments for 5 years, 2 of which were out of date and are worth considering in the light of recent threads here. 

Was there a failure to notify?  The decision relates at [6] that:

“On 5 February 2018 the Appellant’s agent, Jon Vyse of Pearl Lily & Co Accountants (“the Agent”) wrote to Self-assessment, HM Revenue & Customs, BX9 1AS informing them of the Appellant’s liability to tax for the tax year 2016-17.   After supplying the Appellant’s name and National Insurance number the letter included the following:

“We are writing to give notice of liability to tax under TMA 1970 s 7(1) for tax year SA17.

We acknowledge that this notice is after the deadline of 5th October 2017 ...”

Similar letters were written for every year afterwards, all after the s 7 deadline except for 2020-21.”

The income concerned was rental income.

HMRC’s arguments on FTN included these choice morsels.  It is difficult to see the reasons why some of them were advanced, but I’ll address that in my comments subsequently.

[30] “… The Appellant and the Agent were aware of the chargeability and went as far as declaring it but the five letters did not constitute notifications pursuant to s7 TMA 1970 because they did not contain enough information to allow HMRC to make an informed decision as to how to proceed in assessing the type and quantum of the liability.”

[31] “…. The five notices of liability submitted by the agent were not invalidated by their lack of content, but are simply insufficient for the purpose of assessing liability to income tax and become declarations rather than notifications capable of fulfilling the requirements of s7 TMA 1970.”

Here the game is “spot the inconsistency”.  In [30] the letters were not notifications under s 7, because they contained insufficient information.  In [31] the notices are *not* invalidated by their lack of content, but then turn into mere declarations, not proper notifications under s 7.

[33] “The Appellant has never submitted a tax return and the grounds of appeal do not convey why the liability to tax acknowledged since 2017 remains unpaid. HMRC accepts that there is no prescribed form to notify liability pursuant to s7 TMA 1970. However, for the reasons stated above, they consider that the liability was admitted and declared but not notified to HMRC in a manner that would allow HMRC to assess its nature and quantum.

The declarations of liability or chargeability did not include the tax type or any attempt to quantify it.”

Now the liability was admittedly declared, though it is not clear whether the remark about there being no prescribed form for notifying implies that they were proper s 7 notices, but not informative enough to allow assessing.   

[34] …. “HMRC does not accept the Agent’s assertion that the mere act of declaring liability discharges the need to comply fully with the statutory duty to notify in an appropriate manner and pay tax in due time.”

Whatever [33] meant, [34] is clear that HMRC’s view is that mere declaration does not comply “fully” with the statutory duty – which in the context can only be that under s 7 TMA.    And the statutory duty to pay the tax on time (if there is one) has nothing to do with notifying liability.

[41]. “Section 36(1A)(b) of TMA 1970 provides that there is a time limit of 20 years for raising an assessment. In this case all assessments have been issued within this time limit.”

Well yes, but only if there was a failure to notify.

[44]. HMRC submitted that there is simply no “reasonable excuse” or other provision, such as “special circumstances”, in the legislation for amending or cancelling assessments issued under section 29 TMA 1970.

But as we know from the posts in this forum, a reasonable excuse for a failure to notify will invalidate any out of time DA – s 118(2) TMA.

Pausing there, what strikes me as seriously odd is why HMRC were so keen to establish that the notifications that the appellant made were not “proper” section 7 notifications.   There were two types of assessment in issue, the s 29 TMA DAs and the Schedule 41 penalty assessment.

In a case such as this where no returns were made, the only test for an in date assessment is whether s 29(1) is applicable.  If HMRC had, as they seem to say, insufficient information on which to base an assessment, their remedy is to use Schedule 36 FA 2008 to obtain it under threat of penalty, and this they in fact did in December 2021 (see [13]). 

In December 2021 all the years were in date except 2016-17, but [8] shows that HMRC sought information about the income on 5 March 2019, more than two years before 2016-17 went OOT, yet HMRC did nothing more until 22 October 2021, by which time the year was OOT.

It may be then, that conscious of their failure to assess 2016-17 and 2017-18 in normal time, they realised they would have to show that s 36 TMA allowed them to assess those years.  This they could do so long as they could show that for those years there was a “a failure by the person to comply with an obligation under section 7”.  It seems from their contentions in the appeal that they were making a strenuous effort to deny that there had been a notification under s 7 TMA – hence their arguments that what was actually sent to them was a mere declaration  of chargeability, insufficient to satisfy the obligation under that section.

But, as shown above in the extract from paragraph [6] of the decision, the purported s 7 notification for 2016-17 was made on 5 February 2018, four months after the deadline.   The same applies to all the others, 2021-22 apart.  It appears HMRC are of the view that any notification of chargeability under section 7, whenever made, prevents section 36(1A)(b) from operating. 

If that is their view it’s a very surprising one.  Is it right?   Section 36(1A)(b) is the gateway to a 20 year time limit in any case where the loss of tax is attributable to the failure to comply with a s 7 obligation.  

The obligation on every person who is chargeable to income tax for any tax year and who hasn’t received a notice to file is to give notice to an officer of the Commissioners for HMRC that they are chargeable AND to give the notice within the notification period, ie before 5 October after the tax year.   Giving a notice after the end of that period is a failure, as is not giving a notice at all before the assessment is made.

Thus there was a relevant “failure” for the purposes of s 36(1A)(b) whether or not the letter of 5 February 2018 and others was a s 7 notification.

Of course it was.  There is nothing in s 7 that requires anything other than for the taxpayer-in-waiting to say that they are chargeable.  The next step envisaged and the reason for the 6 month deadline is the issue of a notice to file with a 31 January deadline if possible.   Nothing requires the notification to set out the amount or type of income.

Form SA1 which is one way of notifying chargeability simply asks you to say why you “need” to file a tax return.  It doesn’t require any breakdown of the types of income you have or the amount apart from a few general statement which they might or might not need to answer.

Is the HMRC spurious distinction between a mere declaration and a proper s 7 declaration relevant to penalties?   Paragraph 1 Schedule 41 FA 2008 allows a penalty to be assessed if:

“a person … fails to comply with an obligation specified in the Table below (a “relevant obligation”).

Obligation under section 7 of TMA 1970 (obligation to give notice of liability to income tax or capital gains tax).”

That means failure to comply with the obligation set out above, and failure means failure by 5 October after the year, not just a failure that still continues.

What did the Tribunal decide?  It came to the right answer in relation to the DAs, but was wrong about the penalty for 2020-21 in principle and arguably wrong about the percentage reductions in some other years.   But the Tribunal made some very questionable statements.  For example:

[55] “… However, the wording used by the Agent in his six letters to HMRC falls far short of the ordinary meaning of s7 TMA 1970.”

Nonsense, I’m afraid, for the reasons given above.

[56] “… Although the Appellant attended the hearing, she did not give evidence and the Tribunal was therefore unable to ascertain when she became aware that she owed tax for each of the five years under appeal.”

Why does it matter?  Notification was her only obligation, and that was done, albeit late.

[58] “ … All five discovery assessments were therefore issued well within the time limit. … The penalties could have been suspended had the Agent engaged with HMRC by submitting Self-Assessment Returns in each year.”

No they couldn’t – suspension does not apply to Schedule 41 penalties.  It is not surprising that neither the appellant nor HMRC mentioned the possibility (and HMRC usually do in Schedule 24 FA 2007 cases where it is relevant)

[61] “The Agent has not put forward any arguments as to why the Tribunal should find that there were any special circumstances. The Tribunal is unable to find any facts that would justify granting any reductions due to special circumstances.”

Fair enough, but what was missing in the FTT’s decision was any discussion of reasonable excuse under s 118(2) in relation to the OOT DAs .  The letter quoted in [6] set out at the start of this post went on to say:

“However, we believe our client has a reasonable excuse in that it has taken our client sometime to self- assess whether any tax is due and the estimate tax at risk is low. We therefore ask that a late notice penalty not be issued.”

At [34] HMRC are shown as submitting that:

“The Appellant and the Agent do not have a reasonable excuse not to have acted upon their knowledge that the Appellant was liable to income tax.”

And see what they (erroneously of course) say in [44] shown above.

The FTT swallowed these misstatements and garbled submissions by HMRC, while ignoring the agent’s attempt to invoke a reasonable excuse defence for the failure which would stop any penalty – paragraph 20 Schedule 41 FA 2008.

As to the quantum of the penalties, HMRC had submitted:

“[47]. Paragraph 13 Schedule 41 FA2008 determines minimum amounts of penalties depending on whether a person’s disclosure was prompted or unprompted, and whether HMRC became aware of their failure less than 12 months after the due date for income tax unpaid as a result of the failure. The failure to notify penalties for tax years 2016-17 to 2020-21 detailed in the table at paragraph 22 have been charged at 20% These percentages reflect the timing and quality of the Appellant’s disclosure, and determination that their disclosure was prompted pursuant to paragraphs 12 and 13 Schedule 41 FA2008. The penalties were explained in the penalty explanation letter 18 October 2022. HMRC submitted that the penalties have been charged correctly.”

The Tribunal’s response was:

[62]. “As the five penalty charges were issued at the same time and were all calculated in accordance with statutory requirements, they also are valid.”

Were they all so calculated?  We know that all years were penalised and that the penalties in all years were 20% of the PLR.

The incorrect characterisation by HMRC of the 5 letters as not being sufficient notification means that even the letter sent on 5 October 2021, clearly in time on any sensible, not nonsensical, interpretation of section 7 was erroneously swallowed as not an in-time notification gullibly by the FTT.

That swallowing also meant that there is no attempt to determine whether Case A or Case B penalties have been applied (paragraph 13 Schedule 41) nor any discussion of HMRC’s assertion that the disclosure was prompted [47].   In the absence of any mention of what the percentage reduction was we cannot know for sure.  A 20% penalty for a Case A prompted failure means a 50% reduction; a Case B one means a 100% deduction.  It would be in accordance with HMRC’s submissions that this was a Case B case with maximum reduction for disclosure.  But the timing of what were s 7 notifications is such that I think that most if not all of the failures were Case A ones.

I also find it difficult to see on the basis of the letters and notifications that the disclosures were prompted.   

This brings me to the agent, who cannot escape unscathed by remarks from me.  In the grounds of appeal the errors of omission include:

  • Arguing about the penalties on the lines indicated above.
  • Not submitting RE for both penalties and OOT Das.

Errors of commission include saying:

  • all DAs are invalid - [36(1)]
  • all s 7 notifications were in time so all penalties invalid - [36(2) and (3)] – ludicrous.  RE was the way to go here.
  • DAs cannot be issued if there are no returns – [36(7) & (8)]
  • Citation of powers is necessary in all DAs – [36(10) & (12)]

The merely irrelevant category includes [36(4), (5), (6) & (9).

An inglorious decision argued by incompetents.   Food for Justin’s strictures about the FTT no doubt.  

Replies (23)

Please login or register to join the discussion.

RLI
By lionofludesch
11th Mar 2024 21:05

Ehhh, Richard, you do make me chuckle.

Thanks (0)
avatar
By More unearned luck
11th Mar 2024 22:00

"If HMRC had, as they seem to say, insufficient information on which to base an assessment, their remedy is to use Schedule 36 FA 2008 to obtain it under threat of penalty, and this they in fact did in December 2021 (see [13])." Or for 2017/18 and later years, NTFs.

HMRC's problem (if an RE had been proffered) is assessing the earlier years. This is reminiscent of the HICBC Wilkes problem. If HMRC had started sooner the exercise of comparing incomes to CB claims and had issued NTFs in suitable cases rather than DAs then the flaw in s 29(1) wouldn't have mattered nor would any ignorance of the law RE.

Thanks (0)
avatar
By kim.shaw-and-co.com
12th Mar 2024 01:12

richard thomas wrote:

An inglorious decision argued by incompetents.

One of the difficulties, especially looking at the level of tax liability and penalties involved in this Appeal as an example, may stem from the costs involved in engaging (perhaps) more experienced hands in the absence of fee protection insurance.

Whilst homeowners sometimes get legal expenses cover options extending to tax appeals associated with buildings and contents cover, tenants are less likely to. That leaves standalone cover or fee protection insurance offered to practices to market to their clients.

She may have purchased the latter, but one of the conditions of such policies is usually that Returns are filed within 3 months of the due date and cover extends to both a Compliance Check into the Return and appeals to FTT/UT in respect of an assessment and or penalty determination associated with a closure notice. Alternatively, an application by the taxpayer for a CN.

Maximum fees insured for a single claim rarely in practice take the well-represented insured any further than the FTT, being capped overall at £100k -£150k, but more often £100k. Situations where DAs are in point due to failure to notify are frequently excluded unless a Return has been filed, Compliance Check opened and the DA issued pursuant to conclusion of the 'Check'.

The chances of this Appellant having been represented on an insured basis, especially given she was represented by an Agent with a specialism in micro businesses of family owned companies (from a quick Google), may not be very high.

Prospects of appealing the FTT decision on an insured basis may be higher given the defects in the judgement identified above - you would probably know better than anyone on here the position as regards insurable grounds for Appeal of the FTT decision.

However, quite what premium she would be charged for that, and whether (taken together with her likely costs already incurred) it would objectively make sense for her to do so given this may be impacted by prospects of a costs award, I have no idea.

I am not defending the Agent in any way for having accepted the case and failed to run the appropriate arguments, or any other defect. Nor can I be sure that the Agent was not representing the Appellant under authority from an insurer (however I somewhat doubt it given the Agent's profile).

Unfortunately, the cost to the Appellant of paying 'full whack' fee costs of having her appeal presented more competently in the absence of insurance (if that is the choice she weighed up) may have been prohibitive relative to the tax and penalties in point.

Unless the Agent now accepts his likely errors of judgement and waives his fees for preparation of case and representation at the Tribunal, which in my view would likely be the 'decent' thing to do, the Appellant may already have considerable difficulty rationalizing her costs position.

Thanks (0)
avatar
By Justin Bryant
12th Mar 2024 09:26

Very topical indeed, but on this occasion I would be a bit less harsh on the FTT judge, as if there is woefully inadequate advocacy in the first place then the taxpayer is likely to get what they deserve (the burden of proof is basically on the taxpayer after all) and they should not have to rely on the judge to effectively argue their case (as they should know that many judges are rubbish even in front of an effective advocate - great and knowledgeable judges like RT who can go out of their way to dig taxpayers out of holes (quite properly in the interests of justice of course) are very much the exception as we all know.)

See a typical case of that happening here, where the taxpayer had an easily winnable big money case, but did things on the cheap without proper advocacy/representation and paid the price.

https://www.accountingweb.co.uk/any-answers/bonkers-cgt-judgment

It's not just tax tribunals of course. I see this in county courts all the time, where one party has a good advocate and the other doesn't have any and guess what happens more often than not (even when the party without an advocate has a good case)?

Thanks (1)
Replying to Justin Bryant:
avatar
By Justin Bryant
12th Mar 2024 13:58

BTW, I'm no fan of Master of the Rolls Sir Geoffrey Vos, but I agree dodgy decisions like this will be a thing of the past in a few years due to the advent of AI. See:
https://www.lawgazette.co.uk/news/ai-will-change-principles-of-common-la...

Thanks (0)
Replying to Justin Bryant:
RLI
By lionofludesch
12th Mar 2024 14:05

Ah - the King's chauffeur.

Thanks (3)
Replying to lionofludesch:
avatar
By Justin Bryant
12th Mar 2024 15:58

Blimey, I only just got that!

Thanks (1)
Replying to Justin Bryant:
avatar
By richard thomas
12th Mar 2024 14:35

So if in future AI was used to make a decision on the same factual matrix as Owens, what is to stop it having learned about how to decide such cases by using the Owens decision as part of that learning?

Thanks (2)
Replying to richard thomas:
avatar
By FactChecker
12th Mar 2024 14:39

The AI Decision Verifier - trained on a LLM specially prepared and maintained by the impartial and all-knowing ... ah, see what you mean! :=)

In the article for which Justin kindly provided a link, the more concrete and potentially plausible suggestion made by Vos appears to be for AI (suitably curated presumably) to provide a first-level triage service as a means of speeding up the flow to the point where humans are thought to be needed to consider the facts.
The unspoken danger, I guess, being that humans then design the processes such that if the potential plaintiff doesn't like triage conclusion ... then it becomes nigh on impossible to 're-join' the overall process (like the lack of opportunity to 'go around' an A&E filter that fails to acknowledge your symptoms)?

Thanks (2)
Replying to richard thomas:
avatar
By Justin Bryant
12th Mar 2024 15:55

I'm no AI expert and no-one can tell the future, but all I'm saying is that I think it's inevitable AI will be involved in some way, shape or form, even as just a sense check to avoid such dud judgments.

Thanks (0)
Replying to Justin Bryant:
avatar
By FactChecker
12th Mar 2024 16:05

Equally plausible (logically) .. a kind of inverted triage (check whether or not patient should have died given the symptoms at time of entry)!

Much easier to implement than Vos' 'prediction' - although obviously it wouldn't have the 'speeding-up' benefit that he envisages.
However, as per my tongue-in-cheek medical analogy, where does it get you in practical terms? Would all judgements then become 'subject to confirmation'?

Thanks (2)
Replying to FactChecker:
avatar
By Justin Bryant
12th Mar 2024 16:08

In about 20-50 years it's basically gonna be like 2001 Space Odyssey with Hal et al.

https://www.newscientist.com/definition/clarkes-three-laws/

Thanks (0)
Replying to Justin Bryant:
avatar
By Justin Bryant
15th Mar 2024 14:26

Looks like this AI re judges etc. could be happening in more like 20-50 months:
https://www.taxjournal.com/articles/robot-judges-the-use-of-predictive-j...

Thanks (0)
Replying to Justin Bryant:
avatar
By richard thomas
15th Mar 2024 14:44

It doesn’t look like it at all.

Thanks (0)
Replying to richard thomas:
avatar
By Justin Bryant
18th Mar 2024 11:15

Well it does to me (reading between the lines if necessary) and here's Vos's speech (see para 46, where he's clearly not pooh-poohing the idea at all): https://www.judiciary.uk/speech-by-the-master-of-the-rolls-ai-transformi...

Thanks (0)
Replying to Justin Bryant:
By Ruddles
18th Mar 2024 11:53

I've read between every single line and I still can't see anything that supports a 20-50 month timeframe.

Thanks (1)
Replying to Ruddles:
avatar
By Justin Bryant
18th Mar 2024 14:18

I obviously meant nearer then than the prior minimum 20 year figure. If it was anything less then a decade then that would equally be more accurate (probably) is what I mean.

Thanks (0)
Replying to Justin Bryant:
avatar
By richard thomas
18th Mar 2024 13:23

Sir Geoffrey Vos says it may not be long before people are asking why not, not it not be long before it's being done.

Like Ruddles I can nothing in the Tax Journal article to support your supposed timeframe - most of it is warnings against.

Thanks (1)
Replying to richard thomas:
avatar
By Justin Bryant
18th Mar 2024 14:14

Let's agree to differ (people read between lines differently after all). I see the Big 4's tax advice is already basically AI-based:
https://www.icaew.com/insights/viewpoints-on-the-news/2024/mar-2024/pwc-...

Thanks (0)
Replying to Justin Bryant:
avatar
By FactChecker
18th Mar 2024 14:43

Should that be "AI-based" or "AI-biased"? :=)

The 'interesting' bit is "PwC has created the assistant in partnership with OpenAI, developer of landmark platform ChatGPT" ... which primarily means they have created (and no doubt will continue to feed) a proprietary LLM.
So depending on who you go to for your advice, in their futuristic vision you may get entirely different answers ... even on the 'basics' of tax returns.

Not sure how clients benefit (or who is responsible if the advice is generated from an LLM owned by the adviser) ... all I can see (rather like HMRC) is partners gleefully rubbing their hands at the thought of increased profits through fewer staff?

Thanks (2)
Replying to FactChecker:
avatar
By Justin Bryant
18th Mar 2024 15:11

Yes; that's of course right (and I'd be surprised if you'd get mostly identical tax returns without AI).

Thanks (0)
Replying to Justin Bryant:
avatar
By Tax Dragon
19th Mar 2024 15:58

Justin Bryant wrote:

people read between lines differently after all

AI reads the lines, men read between them, women go by body language... and IIRC it's fools who rarely differ.

Thanks (0)
Replying to Tax Dragon:
avatar
By FactChecker
19th Mar 2024 17:46

Interesting summation ... but, ignoring the gender trap (which I know wasn't your primary point anyway), my inability to make my earlier point more clearly has led to Justin and you agreeing!

FWIW I'm of course aware that the combination of law and human interpretation will often (if not quite always) mean that you can get different answers by paying different professionals - not for any deliberate reason of bias but because humans aren't automatons.

My obviously badly made point was meant to be that PWC (and others) are hardly likely to promote the idea that a move to ever greater reliance on AI (at least for the groundwork) will replicate these human 'differences'.

So far, abetted by the developers, the message is the reverse - that AI will deliver 'accurate, consistent and correct' answers ... after all these are reliable machines ... without referencing my point that (in order to maintain a competitive edge) it's not in their interest for all AI to give the same answers, so they pour money into building their own LLMs.
The objective has nothing to do with a better service for clients, and everything about reducing costs whilst developing an edge over competitors doing the same.

So the differences will no longer be just human/chance but will become deliberate - and that can lead to very dark scenarios.

Not directly relevant here, but I've just been reading the case details of a severely depressed person who was used as a guinea-pig to test/refine the efficacy of an AI ChatBot being 'trained' to recognise human emotions and reflect back their tone (as evidence of 'empathy') within the conversation.
There's a lot of money being poured into this area right now ('the future of social care for the lonely and elderly'), so there was considerable concern (for the wrong reasons) when the person committed suicide.
On looking back through all the 'conversations', the researchers found that the Bot was so focussed on the empathy bit (just algorithms to it as it neither feels nor understands emotions) that ... it rapidly descended into 'supporting' the person by agreeing that life was awful - and eventually proposed a suicide pact!

Thanks (2)