SPK- Spark NZ (TELCO)

[Aaron]

Sorry to morph an interesting question on earnings and dividends to a discussion on depreciation recovered but as I like an argument I should see it out to the end.

I stand corrected there would be no depreciation recovered on the sale of the shares in Tower co.

I see that the assets were transferred into tower co in the 2022 financial year and if I read the note 1.5 on page 86 of the 2022 annual report correctly then $198 mill of assets and $94mill of liabilities were transferred and per the note.

No gain or loss was recognised in the statement of profit or loss on classification of the above assets and liabilities to held for sale.

Interesting that a company worth $104mill ($198-$94) in 2022 could be sold in 2023 for $900mill. I have no idea of the rules around consolidated companies so maybe this is not a big deal. It also is irrelevant to the company going forward.

Maybe there is a nifty way to sell assets without incurring depreciation recovered tax issues as I had foolishly assumed.

P.s. just read Ferg's post 2320 after writing this. Maybe depreciation recovered is not as boring as I thought.

[bull....]

my understanding is depreciation once claimed it cant be taken back even on a sale.

[Ferg]

I see that the assets were transferred into tower co in the 2022 financial year and if I read the note 1.5 on page 86 of the 2022 annual report correctly then $198 mill of assets and $94mill of liabilities were transferred and per the note.

Maybe there is a nifty way to sell assets without incurring depreciation recovered tax issues as I had foolishly assumed.

Thanks for the update.

I checked the 2022 AR fixed assets disclosure and the cost of towers etc that were transferred was $970m* with accumulated depreciation of $891m with a book value of $79m. So ignore my $59m in my earlier post - I thought the transfer was in FY23 but I see now it was FY22 per your post.

Is this a nifty way to avoid depreciation recovered or a scheme designed to avoid tax? I expect Spark would have had expensive advice on this but if I were Nicola Willis I would be asking the CIR to look into the tax treatment of this transaction.

*This assumes there were no other asset disposals during that FY with a book value of zero.

[Aaron]

my understanding is depreciation once claimed it cant be taken back even on a sale.

Better hope you are never audited by IRD then.

Mind you as a full time share trader your computer and the 12 screens I heard about from SailorRob might be your only depreciable assets so no worries then.

[Aaron]

Thanks for the update.

I checked the 2022 AR fixed assets disclosure and the cost of towers etc that were transferred was $970m* with accumulated depreciation of $891m with a book value of $79m. So ignore my $59m in my earlier post - I thought the transfer was in FY23 but I see now it was FY22 per your post.

Is this a nifty way to avoid depreciation recovered or a scheme designed to avoid tax? I expect Spark would have had expensive advice on this but if I were Nicola Willis I would be asking the CIR to look into the tax treatment of this transaction.

*This assumes there were no other asset disposals during that FY with a book value of zero.

No doubt Spark would have complied to the letter of the law.

[bull....]

Better hope you are never audited by IRD then.

Mind you as a full time share trader your computer and the 12 screens I heard about from SailorRob might be your only depreciable assets so no worries then.

i was in reference to your transfer in 22 and yr 23
if it was done all in the same yr deprec recovered , otherwise a gain or loss on sale in subsquent yrs is recorded generally and no deprec recovery ( except on a loss maybe ) all irrelevant stuff this dep recovered stuff really as all just stuff going thru different pl accounts to reflect BV

[Aaron]

i was in reference to your transfer in 22 and yr 23
if it was done all in the same yr deprec recovered , otherwise a gain or loss on sale in subsquent yrs is recorded generally and no deprec recovery ( except on a loss maybe ) all irrelevant stuff this dep recovered stuff really as all just stuff going thru different pl accounts to reflect BV

Sorry misunderstood, all pretty irrelevant this debate in the greater scheme of things.

[Snoopy]

My snout is out! I am going full detective mode into sniffing out where Spark's imputation credits have come from. Like any good detective case, we have to start from what we know 'for sure', and work from there. The cashflow statement is normally a good place to start as 'cash is cash' and can't be fudged.

What we do know is that:

1/ A company can only gain NZ imputation credits by paying NZ income tax to the NZ government IRD.
2/ As at 30th June 2020, Spark's imputation credit balance was $0m. (AR2020 p94). IOW imputation credits were exhausted on that date.

So.......if we add up the tax paid from the cashflow statements subsequent to the EOFY2020 balance date, that will give us the maximum imputation available to Spark since.

NZ Imputation Credits Available FY2021-FY2023.5	Income Tax Paid	Imputation Credits Available	Reference
FY2021	$188m		(AR2021 p65)
Tax Effects of Non-NZ Profits FY2021	($6m)		(AR2021 p100)
Taxes Paid in Foreign jurisdictions FY2021	($6m)		(AR2021 p100)
FY2022	$160m		(AR2022 p82)
Tax Effects of Non-NZ Profits FY2022	($7m)		(AR2022 p119)
FY2023	$190m		(AR2023 p92)
Tax Effects of Non-NZ Profits FY2023	($5m)		(AR2023 p130)
HY2024	$101m		(HYR2024 p6)
Tax Effects of Non-NZ Profits HYR2024	($2.5m)		(Estimate 0.5xFY2023)
Total	$613m	$613m

NZ Imputation Credits Paid Our FY2021-FY2023.5	Dividend Paid	Imputation Credit Attached (1)	Reference
FY2021 Previous Year Second half year dividend	$230m	$89.4m	(AR2021 p91)
FY2021 First half dividend	$231m	$89.8m	(AR2021 p91)
FY2022 Previous Year Second half year dividend	$233m	$90.6m	(AR2022 p110)
FY2022 First half dividend	$234m	$91.0m	(AR2022 p110)
FY2023 Previous Year Second half year dividend	$234m	$91.0m	(AR2023 p121)
FY2023 First half dividend	$252m	$98.0m	(AR2023 p121)
FY2024 Previous Year second half dividend	$249m	$96.8m	(HYR2024 p5)
Total		$646.6m

Imputation Credits Paid calculations

1/ 0.3888 is the number you apply to the declared dividend to get the IC's where they are fully imputed (0.3888 = 0.28/0.72)

-----------------------------

The imputation credits paid out cannot exceed the imputation credits available. So what gives? One word: timing.
The imputation credits paid out cannot exceed the imputation credits available, BUT Spark has until the 31st March date, following the reporting year end, to make up the difference: $646.6m-$613m = $34m.

This is similar to the personal tax IR3 system, whereby any shortfall in the previous year's tax ending in March does not have to be made up until the first quarter of the subsequent calendar year. So apparently Spark can give shareholders an 'imputation tax credit', without having handed that tax money over to the IRD. Only a 'promise to do so' in the future. What happens if they don't pay? Does the IRD nullify any unpaid for imputation credits already paid out to shareholders?

SNOOPY

[Snoopy]

The post I have requoted below was my best previous attempt to explain why Spark were able to pay out more in imputation credits, than the tax they had paid on previous earnings, over this particular timeframe.

'Deferred tax assets and/or liabilities' are always presented as a cumulative total on the balance sheet. For comparative purposes, I need to present the 'earnings history in a comparative cumulative format. The table below is the same data presented in post 2122, but in that different (cumulative) format. I use the EOFY2020 as a start point, because we know the imputation credit balance was nil on that date.

Another point to bear in mind is that the imputation credits paid out only need to match or be less than the imputation credits created on 31st March each year. And because the reporting half year ends in December and the reporting full year in June, a 31st March date does not correspond to any of the reference dates in the table below. That means any payments owing can be (actually must be) made up at a date that is not in the table below (a time period ended 31st March).

Declared Earnings verses Dividends Paid FY2021 & FY2s022 & FY2023 & HY2024	Cumulative Net Dividend {A}	Cumulative Imputation Credits	Cumulative Declared HY Net Earning s {B}	Cumulative Net Earnings Shortfall {A}-{B}	Cumulative Imputation Credit Shortfall (1)	Deferred Tax Assets (Liabilities)	Cumulative Deferred Tax Assets (Liabilities) from base
Start Point: End of FY2020	$0m	$0m	$0m	$0m	$0m	($61m) base
Sub Total (HY2021 balance date)	$230m	$89.4m	$147m	$83m	$32m	($54m)	$7m
Sub Total (FY2021 balance date)	$461m	$179.2m	$381m	$80m	$31m	($82m)	($21m)
Sub Total (HY2022 balance date)	$694m	$269.8m	$560m	$134m	$52m	($81m)	($20m)
Sub Total (FY2022 balance date)	$928m	$360.8m	$791m	$137m	$53m	($108m)	($47m)
Sub Total (HY2023 balance date)	$1,162m	$451.8m	$956m	$206m	$80m	$44m	$103m
Sub Total (FY2023 balance date)	$1,414m	$549.8m	$1,254m	$160m	$62m	$55m	$116m
Sub Total (HY2024 balance date)	$1,663m	$646.6m	$1,411m	$252m	$98m	$58m	$119m

Notes

1/ Calculating associated imputation credit 'I' from NPAT: I/(I+NPAT) = 0.28 => 0.28NPAT + 0.28I = I => I = (0.28/0.72)NPAT

-----------------------

Lots of numbers here. The important information to focus on are the two columns in bold.

The first column in bold is the 'imputation credit shortfall'. This is the quantum of imputation credits that have actually been paid out, but which are attached to 'earnings in excess of normalised income'. How is this possible? It is possible if a company already has a store of imputation credits 'in the IRD bank' (so to speak) that are available to 'top up' the imputation credits paid out in relation to normalised operational earnings.

Does Spark have such a top up source available? This is where 'bold column 2' comes in. If the cumulative imputation credits in this second column exceed the corresponding cumulative number in the first column, then there is nothing to stop Spark using those 'imputation credits in the IRD bank' to pay out that imputation credit shortfall. Looking at the table, we can see that this was indeed possible based on the 'column 2 bold balances' exceeding the 'shortfall imputation credits' over the HY2023, FY2023 and HY2024 periods.

A subsequent post by JeffW has lead me to pause and rethink this.

The deferred tax asset or liability represents expected future tax liabilities or benefits, that will in time are expected to increase or decrease income tax payable. For example, accrued Employee Holiday Pay is not deductible for tax purposes at balance date (unless paid within 63 days), but is recognised as wages in the P&L and a liability in the balance sheet. So if a company has say $1m of Holiday Pay liability accrued, it has a deferred tax asset of $280,000 being the future tax benefit (or future deduction if you like) to the company when the holiday is paid out. Conversely, if the company has depreciated an asset by say $1m, but then revalues it by $600,000 then it anticipates a future depreciation recovery of the $600,000 and therefore recognises a Deferred Tax Liability of $168,000 being 28% of the $600k.

Summing this up in my own words, although a deferred tax asset is real (it is on the balance sheet after all) it is 'time delayed'. This means I can't 'cash in' my 'deferred tax asset' with the IRD (or get it recognised by anyone else) until that time delay passes. That in turn means that my quoted exercise above, where I attempt to justify the 'more than have been earned' imputation credits paid out by netting those off against a 'positive incremental deferred tax asset balance' is a load of cr@p and a complete crock.

Post 2328 shows the real answer. This being that the difference is 'in effect' being (temporarily) funded by IRD who allow Spark to pay up their tax due in delayed time, and the benefits of that payment to be passed on to shareholders before it is received.

But are my above quoted workings really wrong? In an accounting sense I would say 'yes'. But in an investment sense? The point I am trying to make here is that these 'incremental deferred tax assets' that I have tabulated will be real at some point in the future. The IRD knows this. So my argument is that the 'apparent generosity' of the IRD in allowing Spark nine months from their balance date to settle up the balance of their 'imputation credit account', is offset by the IRD knowing that Spark has been given an 'incremental deferred tax asset' which is in effect a loan 'the other way' from Spark to the IRD.

(Note: Before anyone rips into me for this interpretation, I do know that from the IRD perspective, the outstanding imputation credit catch up payment is technically not outstanding, because the IRD rules say it isn't even due until 31st March following the date the Spark results are declared. But I am looking at these payments from an 'investor perspective', not from the perspective of the IRD).

There is no legal requirement for the 'cumulative incremental imputation credit shortfall' to be balanced out by the, 'incremental from base level, rise in value of the deferred tax asset' as I have suggested in my quoted tabulated calculations above. As far as the IRD is concerned, neither of these two things even exist. They are just 'timing issues' that have to be reported in the published accounts done to IFRS standards, as if they existed.

Nevertheless from an investor perspective I think my exercise is worthwhile, because it is built on what is likely to happen 'eventually'. Balancing out debts and assets that 'don't exist now' but 'will exist when time shifts' (even if the 'becoming reality dates' don't necessarily match) is one way of future proofing your selected share investment (Spark) to indicate that your investment isn't on the path to becoming over-yielded.

SNOOPY

[Snoopy]

The imputation credits paid out cannot exceed the imputation credits available. So what gives? One word: timing.
The imputation credits paid out cannot exceed the imputation credits available, BUT Spark has until the 31st March date, following the reporting year end, to make up the difference: $646.6m-$613m = $34m.

Part 1 has shown 'how' Spark has raised extra imputation credits to pay fully imputed dividends (Quick answer: the IRD has allowed Spark to pay off their tax bill late, thinking in Spark reporting terms).

But another question remains. Have the dividends paid from FY2021 been sustainable? Or has that level of dividend only been possible because of a series of one off events?

If we add up the tax paid from the cashflow statements subsequent to the EOFY2020 balance date, that will give us the maximum imputation available to Spark since.

NZ Imputation Credits Available FY2021-FY2023.5	Income Tax Paid	Imputation Credits Available	Reference
FY2021	$188m		(AR2021 p65)
Tax Effects of Non-NZ Profits FY2021	($6m)		(AR2021 p100)
Taxes Paid in Foreign jurisdictions FY2021	($6m)		(AR2021 p100)
FY2022	$160m		(AR2022 p82)
Tax Effects of Non-NZ Profits FY2022	($7m)		(AR2022 p119)
FY2023	$190m		(AR2023 p92)
Tax Effects of Non-NZ Profits FY2023	($5m)		(AR2023 p130)
HY2024	$101m		(HYR2024 p6)
Tax Effects of Non-NZ Profits HYR2024	($2.5m)		(Estimate 0.5xFY2023)
Total	$613m	$613m

But what happens if we replace the 'actual tax paid' with the tax that would have been paid if there were no abnormal events. To work out that, we need to know the 'normalised profit' for all periods which I have calculated in post 2073
https://www.sharetrader.co.nz/showth...=1#post1022531
For the normalised profit of HY2024 refer to post 2134.

Normalised NZ Imputation Credits Available FY2021-FY2023.5	Normalised Income Tax Paid	Normalised Imputation Credits Available	Reference
FY2021	$105m		(post 2073)
Tax Effects of Non-NZ Profits FY2021	($6m)		(AR2021 p100)
Taxes Paid in Foreign jurisdictions FY2021	($6m)		(AR2021 p100)
FY2022	$111m		(post 2073)
Tax Effects of Non-NZ Profits FY2022	($7m)		(AR2022 p119)
FY2023	$119m		(post 2073)
Tax Effects of Non-NZ Profits FY2023	($5m)		(AR2023 p130)
HY2024	$40m		(post 2134)
Tax Effects of Non-NZ Profits HYR2024	($2.5m)		(Estimate 0.5xFY2023)
Total	$349m	$349m

Note

1/ The 'normalised tax paid' on 'normalised profits' is derived by taking each 'normalised profit' and multiplying it by 0.28.

--------------------

The tables show a $613m-$349m= $264m imputation credit shortfall in the normalised figures. To try to get a feel for why the 'declared tax' and 'normalised tax' are so different:

a/ FY2021 I have removed $28m of 'other gains' (that includes $1m from the sale of a minority shareholding in long term investment 'Now New Zealand' (a boutique broadband retailer), $9m from the sale of Property Plant and Equipment (primarily mobile plant and equipment) and $18m from a gain on lease modifications and terminations). I have offset this against a one off refund of $16m of 'wire and maintenance' charges, that were wrongly charged to some fibre broadband and wireless customers. All this is a sub total profit adjustment of $39m.

b/ For FY2022 I have removed $26m of 'other gains' (that includes $10m from the sale of Property Plant and Equipment (primarily mobile plant and equipment)) and $16m from a gain on lease modifications and terminations (for a sub-total profit adjustment of $42m

c/ For FY2023 I have removed the $20m gained from selling property plant and equipment (primarily mobile network equipment) and the $13m gained on lease modifications and terminations have been removed (for a sub-total profit adjustment of $33m).

The total tax paid over FY2023 was $190m. The total comprehensive income was $1,093m. This suggests a tax rate of $190m/$1,093m = 17%. 17% is way lower than the 28% statutory rate. This suggests there is little tax related to the capital events of 'selling the cell towers' (+$583m) and writing off Spark Sport (-$54m). If those capital adjustments were taken out of the FY2023 profit figures, the underlying tax rate reduces to:
$190m/ ($1,093m-$583m+$54m) = 34%. This suggests these capital adjustments were reflected in the actual tax bill. So there is no need to adjust for these again in the normalised tax bill.

d/ For HY2024 Gains of $17m of property plant and equipment sales and $2m on lease changes have been removed (for a sub-total profit adjustment of $19m).

Incremental profits in 'total comprehensive income' over 'net earnings' are:

FY2021: $354m - $384m = -$30m
FY2022: $427m - $410m = +$17m
FY2023: $1,093m-$1,135m = -$42m
HY2024: $131m - $157m = -$26m

That sums to an $81m lower 'total comprehensive income' than 'net income' over our study period. That is the opposite effect that I was looking for to explain why actual tax paid was so much larger than 'normalised tax that should have been paid'.