In Upper Churchill Deal A Grievous Act of Misrepresentation I presented a brief non-technical summary of this far more lengthy and technical post. The purpose of this post is to more fully define the value proposition of Churchill Falls and how the MOU negatively impacts that value.
The MOU is quite broad but this assessment is limited mainly to the Churchill Falls (CF) renewal portion of the agreement, more precisely the purchase of power by Hydro Quebec (HQ) from the Churchill Falls Labrador Company (CFLCo), in which HQ holds 34% ownership. The MOU only defines the revenue HQ will pay to CFLCo for the energy they take.
The MOU offers no details on revenues to come from NL Hydro’s significant usage of CF energy and it provides no forecasts of profit margins to be shared by HQ and NL Hydro. While Premier Furey is promoting the MOU by throwing out large numbers on the potential dividends that will flow to Government (GNL) from NL Hydro, there is absolutely no documentation in the MOU or elsewhere to substantiate those estimates.
The MOU offers no context of alternative solutions and how alternatives may have swayed their final solution. Everything about how they got to this point is a closely guarded secret of GNL and NL Hydro. If the deal is so great why not provide all the details?
Among the few useful things within the MOU that can be assessed are a table allocating power between HQ and NL Hydro and another showing the revenue to be paid by HQ to CFLCo during the 51 year term. Independent analysts have to fill in a lot of blanks with assumptions to critique the renewal’s value proposition – far more than is typical for an undertaking of this scale.
As part of the Government’s PR blitz for the MOU, a chart was posted on the MOU website a week after the original MOU announcement. This official GNL-offered chart is a foundational piece of the analysis and is copied below:
Source: https://www.ourchapter.ca/
GNL claims the average price over the term is 5.9 c/kWh but they don’t advise on what basis. The simple average of the chart shown is 16.3 c/kWh so that isn’t a match. However, using a spreadsheet to map out unit rates over time, and adjusting them for an assumed 3% inflation rate (not provided in the MOU or elsewhere) the average becomes 5.9 c/kWh. The proponents of the MOU must be using 3% as their forecast average inflation value.
Two major observations jump off the chart.
One is that we can see that most of the price benefit arrives after the mid-point of the term and certainly well past 2041. The price appears to be below market value for decades to come. It begs the question, why might the revenue benefits be deferred in this fashion?
The other is the very strange shape of the curve compared to usual long-range planning exercises which normally forecast price on a very straightforward linear basis. We expect to see straight lines or a steady slight curve. This one has a weird finish where price escalation slows down after 2060. Some undeclared intentions lie beneath the surface.
The next step is to use a speadsheet to lay out the GNL price curve and compare it to a standard price escalation curve. The chart below features the GNL price curve (blue line) and overlays a simple line (orange dashed) with steady 8% annual escalation that begins at the same 1.63 c/kWh price point in 2025:
The fit is reasonable for about 35 years into the term but fails badly after about 2060. The escalation factor at the end of the term is clearly much lower. Also, 8% is an unusually high escalation factor that makes no sense in itself. No other long term power purchase comes to mind that uses such a rate. There’s still some hidden issue to find.
Another approach is needed to model the later part of the term and its lesser price escalation. The chart below models a price curve (orange dashed line) with 2% annual escalation sharing the same end point as the GNL rate chart of 37.24 c/kWh in 2075:
This chart provides a great match for the final 15 years or so of the MOU rate forecast. That end makes sense and 2% is a sensible number for escalation. The first two-thirds of the term are way off though.
The choice of 2% wasn’t a wild guess. The sections of the MOU related to major new-build plants (Gull Island and the proposed 1100MW powerhouse addition to CF) cite a plan to set rates for cost recovery using a 2% escalation factor for the entirety of their 65-year terms.
Those new projects demonstrate that the parties are entirely comfortable with a 2% escalation of price over a very long term on some very high costs. It’s not surprising to find that a 2% escalation factor is used in the CF renewal rate forecast but why only in a part of its overall range?
Replacement Cost of CF Energy = the MOU’s New Construction Projects
The gap in the chart above lead to some analysis of what it the true value of CF and what would HQ do if it had to replace CF with some other project or set of projects offering comparable energy and firm power capability. The answer is it would have to build them somewhere and at a very high capital cost.
Competing alternatives such as power imports and wind power might appear to be lower cost at first but these are not firm power sources. To ensure reliable power delivery to its customers, HQ would need additional projects to firm up the power from non-firm sources. Those would likely be the very type of hydropower projects described in the MOU: Gull Island and the expansion of Churchill Falls.
HQ likely has a list of potential projects within Quebec but perhaps those might cost a little bit more because if they were cheaper, they’d be prioritizing to build those first instead of wanting to proceed with their next set of projects in Labrador.
The minimum benchmark for value is therefore replacement cost: the same price needed to go ahead with the new-build projects.
GNL has offered barely enough numbers on the new projects to allow some high-level analysis. The detailed findings are not presented here (this post is already much too long to include it – maybe another time) but the required answer soon emerged. It turns out the 2% escalating rate schedule required for viability of the new-build projects is the very same orange dashed line in the chart above.
The finding made great sense that all parties would want to see energy prices from both “old” CF and the new-build projects converge before the end of the CF renewal in 2075. This appears to be no coincidence but by design.
This replacement cost model with 2% escalation over the 51-year term from 2025-2075 has an average value of 23.7 c/kWh or 10.6 c/kWh when adjusted for 3% inflation. According to GNL, the MOU is returning 5.9 c/kWh average. The MOU average rate is 44% lower than the new-build benchmark for replacement cost. Why?
Is HQ Giving Up Any Value of the Existing Contract? – No
The orange dashed line is a price expectation if the existing CFLCo contract expired Dec.31, 2024. We know that’s not the case. HQ has an ironclad grip on super-cheap CF energy up to August 31, 2041. Several court challenges have defended the contract and it is only reasonable that the people of Quebec should expect to see its benefits fully delivered to them and not abandoned prematurely.
The next iteration of the spreadsheet model develops a value comparison between the MOU and a plain and simple renewal in 2041. The chart in rate terms looks like this:
The tradeoff is easily apparent. The simple 2041 scenario yields virtually nothing until the expiry date when the rate jumps almost one-hundredfold overnight. The MOU by comparison is phasing in rate increases much earlier and more gradually with rate parity not achieved until about 20 years after 2041.
Multiplying rates by the expected volume of energy (note the HQ energy volume dips in 2051 and 2061 leading to blips in the orange line), shows much the same trend in this chart of annual revenues:
Within this chart, two areas of interest are marked as areas A and B.
Area A between the two lines defines the early advantage of the MOU: the extra cash received from HQ vs waiting for expiry of the existing contract in 2041. Area B defines an opportunity cost: revenue not received in the MOU relative to waiting until the existing contract expiry. A simple eye-test shows that area B is much larger than area A.
The next spreadsheet model sets out to prove the mathematical relationship between areas A and B. The hypothesis is that area A constitutes a loan to be paid back with interest as area B. The MOU advises 5.822% is the cost of capital. Will area B compute to be the payback for area A with interest charges accounting for their difference?
Within this iteration of the spreadsheet model, the year 2041 was split 2/3 and 1/3 to reflect the August 31 expiry date. This results in some slight differences from official GNL numbers which refer to pre-2041 returns and would seem not to include the first 8 months of 2041. The distinction is that GNL claims to receive $17 Billion before 2041 but this analysis more precisely defines area A up to August 31, 2041 by which point the net new in-flows are $18 Billion. This is the size of area A. All transfers are modeled as debt with interest owed much like a line of credit where unpaid interest is added annually to increase the total owed. The chart below illustrates both the principal amount received and the added amounts of compounded interest:
With the accrued interest charges, the estimated for cumulative debt owed on the night of August 31, 2041 has swelled to $27.5 Billion.
The payback portion of the term begins on September 1, 2041 and is estimated to run 20 years (it may be longer but the payments on the tail end that may be missing are miniscule enough to ignore) as shown in this chart:
The payment schedule was deliberately distributed to match the shape of area B. The first year, 2041, is smaller because it includes only 4 months.
The blue part of the columns represents the portion of the $27.5 Billion debt paid down each year. The orange part of the columns is an estimate of annual interest charges on the balance of the loan that remains due. Cumulative interest expenses add up to nearly $15 Billion.
Overall, the loan simulation shows that $18 Billion in credit will cost $42.5 Billion.
The final step was to check whether the notional calculated loan payback value would equal the missing revenue of area B.
It does so with impeccable accuracy. That just doesn’t happen by accident. It provides the only plausible explanation for the irregular shape of the MOU rate curve: HQ is providing a credit facility of $18 billion at a finance rate of 5.822% on top of a simple 2041 renewal.
Unraveling the financial engineering shows that it is a highly customized and well-hidden loan program facilitated by HQ.
NL Hydro has given HQ two things in the MOU that have gone unadvertised:
- preservation of the full value of the existing contract
- additional residual benefits associated with the 5.822% interest rate that will earn them billions to further reduce their rates
Equity Contribution Section of the MOU
There is another section of the MOU that isn’t quite what it appears.
The MOU states that HQ is giving cash to NL Hydro to use as NL Hydro’s equity in the new-build projects. Schedule H of the MOU identifies a total of $4.84 Billion in transfers from HQ to NL Hydro during the period of 2025-2035. These transfers appear to be a gratuitous gift from HQ but this is not the case.
As HQ owns 34% of CFLCo, it will receive dividends on CFLCo’s profits. Of the $18 Billion HQ will supply as incremental new revenue to CFLCo, HQ will get back very close to a full 34% of it as a dividend. HQ’s dividend is then passed along to NL Hydro to be used as equity.
Using net present value calculations, it was confirmed all of the cash flows related to NL Hydro equity and HQ dividend add up equally. This is another thing that just wouldn’t happen by coincidence.
This technique ultimately puts the entire $18 Billion revenue paid under the MOU into the books of NL Hydro. The equity money transfer therefore represents approximately one-third of the entire repayable credit scheme and NL Hydro will pay it back with interest as reductions to its earnings from CFLCo.
Concluding Remarks
Only one possible positive finding has emerged within this analysis. It does seem the MOU for CF renewal may be based on “replacement cost” of comparable hydropower projects. It’s so deeply buried as to be invisible but the actual design of the rates and cash flows within the MOU strongly suggest it is in there as the benchmark behind the financial design. That’s a good thing.
Beyond that, it’s hard to find anything for which NL Hydro and GNL deserve a pat on the back.
There remains much to be wary of in terms of what the MOU and the entire secret process leading up to the MOU are not telling us. Far more must be disclosed about why the renewal with HQ is the best and only option put forward. The burden of evidence is entirely on NL Hydro and GNL to prove that.
What is most concerning here, however, is that NL Hydro and GNL have acted to make HQ the Province’s high-rate lender of choice.
This is a scandalous development that should never proceed.
GNL always has the option to go direct to financial markets to obtain bonds. There is no need to create a facsimile of a lending program through HQ with interest rates that may be 2-3% above market.
The interest costs of direct financing might be as little as half of the $24 Billion estimated to be hidden within this MOU. The potential to save $10 Billion or more in payouts after 2041 represents a terrible intergenerational cost burden. HQ will earn that residual benefit instead.
The MOU could have completely foregone the hidden loan scheme. It would be very easy to delete it and revise the MOU. But then the deal would be boring and a long time off from offering returns – it would have far less political value. It’s by far the smarter long run decision but completely useless to today’s political agenda. Premier Furey is willing to sacrifice $10 Billion plus that belongs to a future generation for political ambitions.
GNL should borrow money as needed for deficits and for equity in the new hydro projects up to 2041 as necessary rather than take on the high hidden costs with this MOU. Worries over bond availability are nonsense as a post-2041 MOU based on replacement cost rates would give bond ratings agencies confidence that debts can be repaid.
Before re-engaging on this MOU, there needs to be a pause and a commitment to a more exhaustive and open process for review. There’s a lot to fix. Transparency and fullness and rigorous independent analysis are crucial.
The leaders who have foisted this MOU on the public in the way they have cannot be part of the process going forward. Their boisterous commitment to deceit and misrepresentation, demands their immediate resignations. The hidden cost of this MOU rivals the reprehensible conduct of the leading GNL and NLH figures who notoriously schemed their way to sanction of Muskrat Falls through deliberate fraudulent misrepresentation.
There’s time to make this right but the people in charge today are clearly not representing the public’s best long-term interests, to put it mildly. The public should not be fooled twice.
