I was working with a
younger accountant this busy season who is a fan of AI in tax research. He uses
it quite a bit. He also has a client who in turn has used AI to review his work.
This has not amused my friend, and I understand he intends to fire the client.
Irony, methinks.
There has always been research in tax practice, and AI is just the newest and shiniest model on the lot. My concern about AI is that previous research alternatives did not invent answers - that is, hallucinate. This can be a problem, especially for a young(er) practitioner learning the ropes. An experienced hand may recognize when AI leaves the pavement. That is small comfort, as I question whether an experienced hand would rely heavily on AI.
As we have commented
before: you don’t know what you don’t know.
Let’s look at the Clinco
case.
Peter Clinco was an
attorney in California. He mostly practiced real estate and business law. He
was also an entrepreneur and spent much of his time running MedCafe Westwood, a
restaurant and bar near the UCLA campus. It started off as a partnership, but
over the years Clinco wound up owning the place by himself. MedCafe had
approximately 60 employees but did not have strong accounting for sales and
tips. This would become an issue.
Clinco personally prepared
his 2015 tax return, although he filed it late (2018). He reported restaurant gross
receipts of approximately $1.6 million, with enough expenses to show a net loss
of $400 grand. We do not know whether filing late was an issue, but 2015 got
pulled for audit. There were two areas on that return the IRS clearly wanted to
look at:
- The restaurant
- Two rental properties
Why do I say “clearly?” See a tax return the way I do: where are your
subtractions – that is, your deductions? More specifically, where are your biggest
deductions? That is where an auditor would want to look, because that is where the
dollars – and audit adjustments – are.
The exam started in 2019,
when Clinco was already quite ill. The accountant stepped in for Cinco, but
this was after Clinco commented to the revenue agent that an estimated 10% of
the restaurant’s revenues were in cash.
Clinco planted a bug in
the auditor’s ear. The auditor responded by using a common-enough technique: comparing
known credit (and debit) card transactions to reported cash transactions. While
the ratios can vary (in this case, 90/10), it is a starting point. Sure enough,
the auditor decided something was off and expanded her audit.
What does it mean to “expand”?
Easy. She requested Forms 1099 issued to MedCafe. The IRS would have those as a
matter of routine.
She also requested copies
of bank statements.
COMMENT: The bank deposit analysis is virtually de rigueur for all Schedule C audits at this point (MedCafe was a Schedule C because Clinco owned 100%). The concept is easy: all deposits are income unless proven otherwise. Fail to prove otherwise and you have a problem. I had an audit – with deposit analysis – a few years ago. A son (my client) intermingled his business deposits with his father’s (both were contractors). Why? Who knows. It is not normal business practice; I had a difficult time understanding why he did this; the auditor had a difficult time believing either of us; and his foolishness made the audit much more difficult than it needed to be.
The IRS thought actual revenues
were about $3.8 million – approximately $2.2 million more than the $1.6 million
reported on the tax return.
Yep, you can see that train
a ‘coming.
What was Clinco’s first line
of defense?
COMMENT: Somewhere during this Clinco passed away. Technically the matter would have been pressed by his estate and agents.
Clinco challenged whether
one of the early procedural steps - the Notice of Deficiency – needed to be
signed by the IRS in fresh ink.
This is well-trod road
with (very) low risk of victory, but Clinco’s attorney (Mr. Wagner) brought
novelties to the party:
He cites ‘Cacchillo v Commissioner’ … as a case where the taxpayer challenged the validity of the notice of deficiency because it lacked an official signature. He claims we held the IRS’s failure to issue a valid signed notice of deficiency ousted us out of jurisdiction.”
Mr. Wagner claims ‘Cacchillo v Commissioner’ … overturned ‘Miller v Commissioner’ and ‘Tefel v Commissioner’ ….”
Here is the Court:
Neither of these cases exist as cited.”
OK.
There is no case named “Tefel v Commissioner …”
Going down folks.
The bouillabaisse of case names, reporter citations and legal propositions suggests something cooked up by AI.”
Hard landing imminent.
Their presence is unacceptable.”
So… I would say that the IRS was not required to
hard-sign the Notice of Deficiency.
On to the audit
adjustments.
Cash deposit analysis is
a long-standing technique. The Court granted an adjustment when Clinco could
prove that a deposit was not income, but it was not going to reject the entire
analysis.
Even money-losing businesses, however, can have unreported income.”
There was one more issue.
The IRS wanted some proof
for depreciation expense on two rental properties.
Normally, this is not outrageous
to provide. One gets a copy of the closing statement. Sometimes the
municipality itself maintains those records. Granted, it may not show later improvements
and whatnot, but it is a start.
Clinco went in a
different direction. Clinco argued that the IRS could not challenge
depreciation because they had allowed it in a different tax year.
Folks - with minimal
exceptions - this is not the way it works. The IRS not asking about your “fill-in-the-box”
deduction in year one does not mean they cannot ask about it in year three. This
is long-standing practice and predates me being in school.
Even AI should have picked
that up.
Our case this time was Peter
L. Clinco, Deceased, C. M. Barone-Clinco, Successor in Interest, and C. M.
Barone-Clinco, T. C. Memo. 2026-16

