SPK- Spark NZ (TELCO)

[Snoopy]

When the FY2022 results came out, 'Spark Sport' was very much 'still in the picture' as part of the 'exciting growth story' (at HY2023, Spark exited Spark Sport for a one off write off of $52m). So how did the total of the exciting new growth divisions -at the time- shape up?

P87 of AR2022 is where the true calculation of profitability starts:

	Operating Revenues	less Product Costs	less Labour Costs (1)	less Other Operating Expenses (1)	equals EBITDA
'Other Operating Revenues'	$152m	$72m	$20m	$16m	$44m

Notes

1/ 'Labour Costs' and 'Other Operating Expenses' are estimated in fractional proportion (f) to the percentage of revenue turned over by the 'Other Operating Revenues' business unit.

f= 152/3694 = 4.114%; Labour Cost = 0.04114 x $495m = $20m, 'Other Operating Expenses' = 0.04114 x $381m = $16m

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


EBITDA is a good proxy for cashflow. But barring some trunk transmission assets, most of the equipment at Spark is not long lived. Indeed there is significant investment now replacing the old PSTN telephone system and continuing the 5G mobile roll out. In my assessment, this means EBIT is the more important measure.

Depreciation & Amortisation ('Other Revenue') = 0.04114 x $520m = $21m

EBIT= EBITDA - DA = $44m - $21m = $23m

The interest charge against 'Other Revenue' = 0.04114 x [$26m - $74m] = -$2m. So underlying Net Profit After Tax is:

NPAT = 0.72(EBIT - I)= 0.72($23m-$2m)= $15m

On page 89 of AR2022 we learn "Included in 'Other operating revenues' is revenue from Qrious (Artificial Intelligence, data and analytics), Internet of Things, Spark Sport, Connect 8 (the construction contractor, now fully brought back in house by buying out the Electra shareholding) and exchange building sharing arrangements." I had previously assumed this category included 'Spark Health' as well. But it could be the Spark Health referred to as a promising potential future revenue business unit has yet to start from a zero base.

Whatever, the NPAT estimate for all those promising future growth initiatives looks to have improved by 50%, even if the overall contribution to Spark profit remains small.

I have expressed some doubts on where growth in the operational business at Spark will come from, that is needed to support the current level of fully imputed dividend payments. To that end the 'growth arm' of Spark, captured under 'other revenue', will play a big part. So how is this 'other revenue' growth arm of Spark going?

At HY2023, Spark exited Spark Sport for a one off write off of $52m. So how did the total of the exciting new growth divisions over 2023, -including Spark Sport when it was a going concern - shape up?

P97 of AR2023 is where the true calculation of profitability starts. 'Other revenue' includes revenue from mobile infrastructure, Qrious (Artificial Intelligence, data and analytics),, the Internet of Things, Spark Sport (discontinued mid year), exchange building sharing arrangements and for the first time MTTR (digital identification start up). Connect8, a servicing and installation arm of Spark that was bought back in house in FY2022 has now been amalgamated with the Entelar business unit. Entelar is not listed as being part of 'Other Revenue'.

The much talked about business unit 'Spark Health' is not mentioned either. I therefore have to assume that 'Spark Health' is a grouping for customer service reasons, but for reporting purposes is distributed across many reporting segments. So there are no 'Spark Health profit figures' per se.
My assumption here was subsequently confirmed in HYR2024 on p9:
"Health results are included across a range of product categories."

	Operating Revenues	less Product Costs	less Labour Costs (1)	less Other Operating Expenses (1)	equals EBITDA
'Other Operating Revenues'	$241m	$110m	$32m	$28m	$71m

Notes

1/ 'Labour Costs' and 'Other Operating Expenses' are estimated in fractional proportion (f) to the percentage of revenue turned over by the 'Other Operating Revenues' business unit.

f= 241/3875 = 6.219%; Labour Cost = 0.06219 x $511m = $32m, 'Other Operating Expenses' = 0.06219 x $456m = $28m

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


EBITDA is a good proxy for cashflow. But barring some trunk transmission assets, most of the equipment at Spark is not long lived. Indeed there is significant investment now replacing the old PSTN telephone system and continuing the 5G mobile roll out. In my assessment, this means EBIT is the more important measure.

Depreciation & Amortisation ('Other Revenue') = 0.06219 x $504m = $31m

EBIT= EBITDA - DA = $71m - $31m = $40m

The interest charge against 'Other Revenue' = 0.06219 x [$32m - $99m] = -$4m. So underlying Net Profit After Tax is:

NPAT = 0.72(EBIT - I)= 0.72($40m-$4m)= $26m

Whatever, the operational NPAT estimate for all those promising future growth initiatives looks to have improved by 73% year on year, even if the overall contribution to Spark profit remains small. But it is a bit hard to compare these figures with the similar numbers generated in FY2022 (see quoted text), because the categories of 'Other revenue' have subtly changed, and will be changing again for FY2024.

From HYR2024 p9
-------------------------
Spark has reclassified the comparative segment results to:
• Redistribute certain revenues between two new categories IT products (previously cloud, collaboration, managed data and networks) and IT services (previously service management and security
• Move Qrious, Internet of Things, and MATTR from other products into a new high-tech category
• Split data centres out from cloud, and split co-location out from other products to create a combined data centres category.
-------------------------

SNOOPY

[Snoopy]

Financial Year	Labour Expense	Finance Expense
2018	$513m	$77m
2019	$475m	$85m
2020	$511m	$94m
2021	$491m	$81m
2022	$495m	$74m

The table above shows labour expenses have been held down, and interest costs are at cyclical lows. But I would expect upward pressure on both these cost staples as we look to FY2023. Growth initiative research (post 1982) does not give any hint of a significant boost in top line revenue going forwards. Nevertheless mobile service revenue growth continues to exceed expectations, which largely covers the 'new class' growth initiatives and cloud services not performing to expectations

Part of the ability of a telecommunications company like Spark to grow profits, is their ability to continually take out costs. So how is Spark doing?
I foretold pressure on interest and labour costs (despite the ongoing trend to automation) in the Spark accounts for 2023. And this is exactly what has happened.

Financial Year	Labour Expense	Finance Expense
2018	$513m	$77m
2019	$475m	$85m
2020	$511m	$94m
2021	$491m	$81m
2022	$495m	$74m
2023	$511m	$99m
2xHY2024	$558m	$126m

New category growth initiatives (post 2131) show an incremental NPBT gain of $36m over FY2023. This has been more than wiped out by the increase in labour and finance expenses over the same period. Spark's 'new ventures' are not looking like the key for the company boosting earnings. Nevertheless mobile service revenue growth continues to exceed expectations, and the investment in data-centres is earnings positive. To date this more than covers the 'new category class' growth initiatives not performing to expectations. But it is a competitive market out there, both in mobile services and data-centres.

Cost control projections for FY2024 (I have simply taken the costs for HY2024 and doubled the figures) are not looking good. But CEO Jolie Hodson was very clear at the half year briefing that the 'cost out program' is more than on track. At this point I am giving our CEO the benefit of the doubt. But all will be revealed at financial year reporting time.

SNOOPY

[Snoopy]

Product Category	Operating Revenue (FY2023)	Product Margin (FY2023)	Product Margin %ge (FY2023)	Operating Revenue (FY2019)	Product Margin (FY2019)	Product Margin %ge (FY2019)
Mobile	$1,470m (+15.7%)	$984m	66.9%	$1,271m	$775m	61.0%
Voice	$231m (-52.5%)	$133m	57.6%	$486m	$310m	63.9%
Broadband	$626m (-8.61%)	$298m	47.6%	$685m	$344m	50.2%
Cloud, Security & Service Management	$436m (+9.00%)	$328m	75.2%	$400m	$327m	81.8%
Procurement and partners	$584m (+60.0%)	$67m	11.5%	$365m	$43m	12.3%
Managed Data, Networks & Services	$287m (+45.7%)	$132m	46.0%	$197m	$104m	52.8%
Other Operating Revenues	$241m (+111.4%)	$131m	54.4%	$114m	$51m	44.7%
Total	$3,875m (+8.00%)			$3,588m

FY2019 was the year before any Covid-19 revenue effects took hold. I can see the product categories with the highest 'Product Margin' percentage over FY2023 were:

1/ 'Cloud, Security & Service Management' ' and
2/ 'Mobile'.

The annual report does not do a descriptive 'product category' report to back up the numbers. But I am curious to know the distinction between 'Cloud, Security & Service Management' (+9% revenue) and 'Managed Data, Networks & Services' (+46% revenue). I am thinking the former is 'off the customer premises' while the latter is 'on the customer premises'. But with the ever improving cloud, I imagine such a distinction, if it exists, must be getting blurred. I would have expected 'Cloud, Security & Service Management' would be growing the faster of the two. Which just shows how little I know about this industry, because the opposite happened. Product margin percentages were down five points for both, despite what I am imagining is increasing scale. The effects of competition? Nevertheless, there remains good money to be made in this space.

I find it interesting that if you add the mobile and voice revenues together for the same year, the totals come out to a similar amount (a 3% decline). One way of interpreting that would be to say that 'mobile' is the direct replacement for 'voice' (the old PSTN system), in usage, but also in earnings terms. Despite the rapid decline in revenue from voice (-52% over five years), the percentage product margin has only declined by six percentage points. And legacy 'voice' remains the third highest percentage product margin grouping that Spark sells.

I was very surprised to see revenue from broadband actually decline over the five years. Yet management talk at result briefings would suggest that Spark is more interested in 'running a steady ship' rather than growing this aspect of the business. This is probably related to the product margin percentage for fibre broadband being low in comparison to other front line product group offerings. IOW it makes sense to put your effort into selling product with better margins. Spark have been running a campaign to try and get more 'fixed broadband customers' (those wholesaled from the Chorus network) onto the Spark owned 'mobile broadband network'. If 'mobile broadband' customers are being product grouped under 'mobile' not 'broadband', then this could explain the 'broadband' product line decline. It is notable that 'mobile' is the only specifically delineated product classification to increase its product margin percentage over the comparative five years.

The lowest margin product classification is 'procurement and partners'. This isn't surprising as it basically means picking up someone else's product or service and 'on selling' it. Mobile phone handset sales would be one example. I don't know if the product costs here include all the mall and main-street 'Spark' retail shop rental costs dotted around the country. But I expect they might. The margin may not have jumped. But revenue over the five years most certainly did, by 60%.

'Other operating revenues' showed good improvement. But as a catch all category for 'other activities', I suspect the composition of this product category has changed over the years. So in this instance, I am not sure if it is meaningful to compare figures across the five years.

Overall revenue growth has been 8.00% (that is total, not annual) over five years, while normalised operating profit growth was 6.78% (total not annual) over the same period. You can see that Spark is not a growth investment per se, (although it has performed well on an international comparative against other established telecommunications retailers). Then again, you are not paying a large growth premium to own the shares either. So where does Spark sit on the investment prospect line up? If you want to own a good income paying shares, that will pay you that income across the business cycle, then I would say Spark is a good one to form a cornerstone of such a portfolio.

SNOOPY

[Snoopy]

This is a comparison between 3 methods of measuring net profit after tax. The first one, most quoted in the glossy Spark presentations is 'Net Profit After Tax'. This is recorded in this table below under the monicker 'Net earnings'. This is how the 'often highlighted in presentations' NPAT is described in Spark's 'Statement of Profit and Loss and Other Comprehensive Income'.

Nevertheless in any year, there are often transactions or expenses that are 'one offs', that take away from the general picture of how the core of the business is doing. As an investor, I look for repeatable results. So I like to look through these 'one offs', so that I can get a better understanding of long term earnings trends. To enable me to do this, I calculate something I call 'normalised earnings', which is the second NPAT measure that I have tabulated.

The third NPAT measure is called 'Total Comprehensive Income' (TCI). This income measure takes 'Net Earnings' but further incorporates changes in values of hedge contracts based around currency swaps and interest rates swaps. Spark is partly financed by overseas capital, which is borrowed at overseas interest rates over period(s) of several years. Hedging contracts can provide certainty of NZ dollar denominated interest payments while these loans are outstanding, and certainty on the NZ dollar capital repayment that will ultimately be required to pay back loan capital at the end of each overseas loan term. Nonetheless, 'annual reporting' is required to reflect fair market values of these contracts should they be suddenly and prematurely cancelled. These annual adjustments are required to be incorporated in annual profit figures by NZ accounting standards, despite there being no plans by Spark to terminate these arrangements early. Typically these annual adjustments are volatile in size and may be positive or negative. This is why TCI is a more volatile measure of net profit than the other two methods I have described. To further add to the volatility, equity investments in companies initiated by Spark management, - as measured by a change in market value of such investments over a year - , are also incorporated into the TCI profit calculation.

Year	Normalised NPAT	Total Comprehensive Income (TCI)	Shares EOFY	Normalised eps (tax paid)	Net Earnings ps (tax paid)	TCI ps (tax paid)	dps (tax paid)	EOFY NTA
FY2021:	($384m - 0.72($28m - $16m) ) = $375m	$354m	1867m	20.1cps	20.6cps	19.0cps	25.0cps	80.5cps
FY2022:	($410m - 0.72($26m) ) = $397m	$427m	1872m	20.9cps	21.9cps	22.8cps	25.0cps	78.8cps
FY2023:	0.72x($1152m-($583m-$54m)) - 0.72($33m) = $425m	$1,093m	1845m	23.0cps	61.5cps	59.2cps	26.0cps	105.1cps
HY2024:	$157m-0.72($17m+$2) = $143m	$131m	1815m	7.9cps	8.7cps	7.2cps	13.5cps	92.0cps
Total				71.9cps	112.7cps	108.2cps	89.5cps

Notes

1/ Normalised earnings = Net Earnings (+/-) One off Adjustments. Normalisation details are in post 2073, and in note 5 below.
2/ Normalised eps = (Normalised Earnings) / (Total Shares on Issue, EOFY)
3/ Dividend payments recorded are those paid during the financial year in question (not dividends declared in that financial year).
4/ Gross dividend payments I have calculated in post 1968. After tax dividend figures are the respective gross numbers multiplied by 0.72.
5/ Gains of $17m of property plant and equipment sales and $2m on lease changes have been removed from the declared HY2024 profit.

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

Looking at this 'three year and one half year picture', (the time from EOFY2000 where we know for sure the imputation credit balance of Spark was zero): The cumulative dividend is straddled between 'cumulative normalised income' (24.4% above this) and 'total cumulative comprehensive income' (20.9% below). This means the dividend payments have been more than a shareholder might reasonably expect. But well down on what the company can afford to pay. The principal reason for this is the large 'one off' profit of $583m (31.6cps) from the Connexa (cell phone tower) sell down. Take that away and the dividend payments look anything but sustainable.

You may be wondering how the asset backing per share fell so dramatically between EOFY2023 and EOHY2024 (by the total value of the dividend paid), when 'positive earnings' over the period should have added to that net asset value. The reason is this 'missing capital' was spent on share buybacks which are not recorded in the table.

The one off Connexa sale has certainly provided a lot of capital to paper over some broken shareholder growth dreams (the most recent being the demise of Spark Sport). That Connexa deal even provided a deferred income tax asset. That means that Spark did not have to pay as much tax up front going forwards to reap the imputation credit benefits of that tax. The story behind increased sustainable earnings is that money spent so far in data centres and particularly on the roll out of 5G will start to earn an investment return in the near term. Personally I remain skeptical about these 'build it and customers will come' earnings projections. Everything like this has execution risk. The way I read my table 'poor execution' could lead to the dividend falling by 20% as the cushioning effect of the Connexa sell down flows through.

That is a pretty negative end to my overview of the Spark dividend situation. Readers may be wondering why I have just topped up my Spark shareholding if I think this way. An unsaid assumption behind my 'capitalised dividend valuation model', that determined the fixed the price I was prepared to pay for Spark shares, is that you have to trust management. If management are paying out a dividend at a certain rate, then you have to trust that there is an 'operational business plan' to justify such a level of payment. It is usually management who are in the best position to judge this. Trusting management is the risk that all ordinary shareholders take. An investors trust may end up being misplaced. But this risk factor is why Spark shares at $4.70 are trading on a gross dividend yield of around 7.5%, while Spark SPF600 bond investors with a more certain return profile are only getting a gross market return of 5.5%. For me the extra risk I am taking holding the shares over the bonds is well worth taking, to plug into that higher yielding income stream.

SNOOPY