The theory of share buybacks, with IPL as an example: Part 2
Moving onto the IPL annual report for 2023, there are many years where IPL has 'earned' more from the appreciation of the value of their property portfolio, rather than the underlying rent being paid by tenants on those properties. Property revaluations and (gulp) devaluations are connected to movements in various financial indicators, but are not readily predictable in the short term. That is why it is prudent for IPL management to maintain an 'equity buffer'. This is a way to ensure that a sudden 'property market downturn', will not trigger a 'banking covenant breach' and a concomitant sudden rush to refinance the company on the banking syndicate's terms.
My way of looking at IPL is to see how well the operational earnings (i.e. largely rent) expressed as 'return on assets' covers the bank (and IPL bondholders in this instance) financing costs. From the FY2023 financial statements:
Operational Profit before tax |
$35.207m |
less Current Tax Expense |
($4.972m) |
equals Underlying NPAT |
$30.235m |
When looking at the value of investment properties on the books, most reporting companies use an 'average value of assets' managed throughout the year. I prefer to use just the value of properties at the end of the financial year, because that is the value that will be carried forward into next years accounts and all prior adjustments are historical. The value of Investment properties on the IPL balance sheet as at EOFY2023 is: $1,070.451m. So the underlying return on assets may be calculated as follows:
ROA = $30.235m/$1,070.45m = 2.82%
We are told that the average weighted cost of debt per annum at at 31st March 2023 is 4.0% (AGMPR2023 Slide 8). I am not sure exactly how IPL calculates this. How does this compare with my own 'calculated interest paid' rate from the published accounts? To work out this, I first take the three published loan balances we know about over FY2023 (SOFY2023, HY2023 and EOFY2023) and average them:
($351.530m+$387.576m+$387.037m)/3 = $375.381m
Net finance expense = $16.195m
=> indicative interest rate paid = $16.195m/$375.381m= 4.31%
( I call the above calculation indicative, because we do not know the variations in the loan balance on the other 365-3=362 days of the year where loan data is not published. Nevertheless, there isn't too much difference between 4% and 4.31%)
So straight away we can see that even with a write-down in property values in some years (e.g FY2023), over the medium term we are relying on capital gain to make this funding model work.
In analytical terms, there is something to be said for considering a 'cost of equity capital'. This will help determine if for a particular investment your capital is wisely allocated. However in 'cash terms' there is no cost of equity capital. I say that, because you don't have to pay for the privilege of investing your own money. That means sometimes it is useful to look at a different figure, such as 'Return on Borrowings', or ROB, where your own equity capital is removed from the underlying asset:
ROB = $30.235m/($1,070.45m-$675.02m) = 7.65%
From the above we can see that the underlying IPL 'return on borrowed money' is significantly more than the interest rate being paid on company borrowings. This means there is no imminent 'operational cashflow threat' that might see Investore fold. If ROB > ROA this would also suggest to me that increasing company borrowings as the consequence of a share buyback is, within certain limits, a 'sensible strategy' for IPL to follow, as a rule.
SNOOPY
The theory of share buybacks, with IPL as an example: Part 3
Quote:
Originally Posted by
Snoopy
This would suggest to me that increasing company borrowings as the consequence of a share buyback is, within certain limits, a 'sensible strategy' for IPL to follow, as a rule.
The most current released LVR for Investore as at EOFY2023 was 36.5%. How was this calculated?
Value of Investment Properties (from FY2023 balance sheet): $1,070.451m
Borrowings (from FY2023 balance sheet): $385.037m
=> LVR = $385.037m / $1,070.451 = 36.0%
It looks like 0.5% has gone missing (reward offered for anyone who can find it - no not really). But if $385.037m of loans represents a 36.5% LVR, by how much would the property values have to decrease before that new LVR banking covenant of 52.5% would be breached?
$385.037m / 'Covenant Breaking Depreciated Value' = 52.5% = 0.525
=> 'Covenant Breaking Depreciated Value' = $385.037m / 0.525 = $733.404m
This means the 'cumulative critical write down' amount going forwards that will cause the LVR banking covenant to be breached is:
$1,071.451m - $733,404m = $338.047m
A further property valuation reduction of: $338.047m / $1,071.451m = 30% (round figures) from the 31st March 2023 balance date is the critical figure . This would likely require the one year government bond rate to rise from 4.11% (as it was on 31-03-2021) to 4.11% x 1.3 = 5.34%. (1) As at 28-06-2023 the government bond rate is 4.54%. I think it is unlikely that the one year government bind rate will rise further to 5.34% at this stage in the interest rate cycle. But we can't be 100% sure. This could explain why management are taking some pro-active action selling off those Nelson and Blenheim Countdown properties. Looking at the way interest rates have kept rising since the 31st March 2023 balance date, I do predict more property write downs at the September 2023 half year reporting date for IPL. Whether that will 'spook the market again', I guess we will find out,
Note (1) I am not suggesting that calculating any potential future property write down is as simple as just comparing one year government bond rates at comparative period end points. I am suggesting that as a quick approximation this simple calculation is a reasonable indicator as to what the actual property value adjustment might be.
SNOOPY
discl: do not hold, but is on my acquisition radar