0:00
[SOUND]
[MUSIC]
Welcome back.
In this module, we'll continue our discussion about income taxes, but
instead of talking about deferred taxes we're going to be talking about
uncertain tax positions, another very important topic.
So let's start out with an introduction to uncertain tax positions.
So not all tax deductions survive.
There's different approaches to taking deductions on a tax return.
Some positions are very conservative.
Salary paid to employees are usually deductible and some are disputable.
How you account for transfer of payments, how you allocate taxes between
tax jurisdictions, maybe subject to different interpretations.
And some can be done right at aggressive and
R & D costs that are frequently cited for this.
So what is the accounting issue, where the premise when do you recognize, so
we want to see accounting issue, where the premise.
When do you recognize a deduction?
Do you recognize it when it's taken?
Or when it's accepted by the tax authority?
And how do you incorporate that uncertainty about whether or
be accepted by the tax authority into the measurement of the deduction?
So there's different ways to look at it.
One way to look at it could be that uncertain taxes are a contingent asset.
And then you would look at this is a gain contingency and you wouldn't include
the impact of that deduction in income until the uncertainty is resolved.
In other words when it's accepted by the tax authority.
So graphically you could look at it like this.
For the most part you're not going to accept it until
the probability of recognition becomes very high.
This is what the FASB would call a benefit approach.
You're going to wait until the probability is high and
recognize the benefit, otherwise there is no recognition.
Or you could look at uncertain taxes as a contingent liability.
You could say uncertain tax positions could be viewed as a contingent loss if
it's not accepted by the tax authority.
And at that you would record it when it
becomes probable that a liability exists to that taxing authority.
Well, let's look at that bar again.
Now we've just reversed the probabilities on it.
There would be no recognition if it's probable of denial.
But if it's not probable of denial, we're going to recognize and
the benefit of the tax deduction when it's taken.
So this would be an impairment approach,
this is the terminology the Feds would use.
So which is correct?
Either the tax deduction assets or liabilities.
If you consider it a contingent asset,
you would only recognize it if it's reasonably assured it will not be denied.
If it's considered a liability,
which the FASB compares to an impairment, then the benefit would be
recognized unless it was probable that the deduction would be denied.
So which would it be?
Note that different interpretations are both based on an amortized cost model.
So no tax benefit is recognized until the benefit is reasonably certain,
or no tax liability recognized unless the denial of a benefit is probable.
Both of those have a recognition threshold,
which is characteristic of an amortized cost model.
That's not the way you would do it under fair value.
Under a fair value model you do not have a recognition threshold,
you would deal with uncertainty through measurement.
So there's a difference.
These are a recognition threshold that's characteristic of an amortized cost model.
4:02
So what are the FASB decide when this problem was brought to them,
they decided there is conceptual support for looking at both ways.
You could say that's a contingent asset and
you could say it's a contingent liability.
Well, in order to have consistency in financial reporting the had to make
a decision go one way or the other.
So the Board decided an impairment approach would not be appropriate
when an enterprise cannot conclude to a specified confidence level,
it is entitled to the economic tax benefit.
So they're saying, we shouldn't look at this as an contingent liability,
if you can't say at least to a reasonable level of confidence that you're
going to be entitled to that economic tax benefit.
So they would've gone more with the benefit recognition approach,
that you have a specified confidence level as a precondition for recognition.
So they're going more with that benefit approach, which had the lower level.
However, because it would be a significant change in practice,
they weren't comfortable at leaving that recognition threshold up at
the upper end where the probability of recognition would be fairly low.
Instead, they decided to split the difference.
What they did was they moved the recognition threshold to
a specified confidence interval right in the middle, more likely than not.
This will be the threshold for recognizing a tax deduction when it's uncertain,
whether it will be accepted by the taxing authority.
So it's neutral between recogizing an asset and liability.
The FASB describes this as a benefit approach with a specified,
in other words a non-conceptual confidence level of more likely than not.
And that's what we'll be using in this, in this module.
6:02
So what would a fair value approach look like?
Well in a fair-value model, and certainly again, it's addressed through measurement,
not recognition.
In a fair value model, a market participant view of the probability of
discovery by the tax authority and probability of denial would be
factors that would help determine the amount to be recognized.
Well, what's wrong with that?
Well, first there's a public policy issue.
The fair value, if you're considering a market participant's risk of examination,
in other words, can I get away with it?
That didn't seem like a good place to go as an accounting principle.
So they weren't comfortable with that from a public policy issue.
Also the Board didn't want to write a standard with measurement criteria that
involved estimating the likelihood that a shady tax position would not be detected.
The second reason the Board didn't go with fair value is that it conflicts with
current GAAP for deferred taxes, which we just talked about in the previous module.
Recall that deferred taxes are based upon the future cash flows, but
not taking into account any future changes in tax law, and they're not discounted.
So if we used fair value, it would have to consider the time value money and
it would have to be discounted.
And also,
you would have to consider the impact of any possible future changes in tax law.
Again, the uncertainty would be in measurement rather than in a recognition
threshold, which is what happens when you have to wait until it's enacted.
So you would get a very different measure for
an uncertain tax position then you would from a deferred tax asset and liability.
Because current GAAP doesn't permit discounting or anticipation of changes
in the tax law, the Board ruled out fair value for these uncertain tax positions.
Well, what about almost fair value, say an expected outcome approach?
This is used in another places in GAAP.
They did consider that measurement that uses some of the inputs to a fair
value model measurement, but they would exclude discounting or
anticipated changes in the tax rate and exclude examination risk.
Nah, that didn't fly though, some board members objected to a measurement that
excludes those factors, though they could be significant.
And maybe there was a little bit of an all or nothing approach to it.
So fair value was rejected, we're staying within an amortized class model.
And that's how we ended up with what the FASB calls a non-conceptual answer.
And that's going to be a two-step approach.
The one, you're going to evaluate the tax positions for recognition against that
specified non-conceptual recognition threshold more likely than not.
And they're going to specify a measurement rule that we're going to
recognize the largest amount that is more likely than not of being recognized.
So there are multiple views of the nature of uncertain tax positions.
Some did not address uncertainty,
some viewed uncertainty like a contingent asset.
Some viewed uncertainty like a contingent liability,
no one really considered the fair value of a tax benefit or liability.
The FASB elected a non-conceptual approach using a specified confidence level,
and that's what we'll discuss next.