Thursday, 4 July 2019


Nalcor CEO Stan Marshall gave testimony this week at the Muskrat Falls Inquiry.  During his examination he referred to the need to see the “big picture” as to the economics of Muskrat Falls. He said that this can best be done by explaining the cost of Muskrat Falls in cents per Kilowatt Hour, which makes it easier to understand than by using large numbers in the millions and billions.

His theme was that the project is not really all bad, the costs not unreasonable. When you look closely a different picture emerges from the one Stan has presented. The “big picture” is not nearly as rosy. Was his presentation an attempt to explain away the “boondoggle” or to rationalize it? 

Muskrat Falls a Bargain!
In his February 15, 2018 presentation at Memorial University he said that the cost of generation at Muskrat Falls was a “bargain” at 7 cents per KWh for 2021. He said the problem was that transmission costs need to be added to calculate the 17.42 cents/KWh unit cost of Muskrat Falls. The transmission cost includes both the cost of the Labrador Island Link (LIL) and the cost of the transmission line from Churchill Falls to Muskrat Falls, known as Labrador Transmission Assets (LTA).  He blamed the transmission cost for the overall high unit energy cost, rather than generation. The data Marshall cited purport to show generation cost at a low 7 cents/KWh compared with 10.42 cents/KWh for transmission.
Stan Marshall, CEO Nalcor
According to Stan Marshall it is the high transmission costs which make the economics of the project questionable. These costs are shown in Marshall’s pie chart below to be 10.42 cents per KWh for transmission, including 8.98 cents for LIL and 1.44 cents for the LTA. This pie chart indeed shows that transmission costs exceed generation cost, 10.42 cents compared with 7.00 cents for a total of 17.42 cents/KWh. Does this tell the full story? How clear is the big picture? Let the buyer beware!

The Metrics
The funny thing is that these numbers do not compare with Nalcor’s construction numbers, which are shown in the bar chart below. There is a disconnection.  These numbers describe the direct construction costs and the financing costs of the project and its components. The direct construction costs for MF far exceed those of the LIL and the LTA combined, $5.5 billion compared with $4.8 billion.

So the mystery is why these numbers show a different picture than the pie chart with the cents per KWh.  When we begin to unravel this mystery we find not only that the generation costs per KWh exceed the transmission costs but also that the 17.42 cents per KWh for Muskrat Falls generation (MF) is too low. We also discover that the blended cost must be adjusted upward in order to cover all project costs including the cost of equity capital.

Readers of this esteemed blog will remember previous posts describing the two methodologies used for the project including Muskrat: The Biggest Gamble of All and Finance Department Left Out of Muskrat Risk Assessments. The “cost of service” approach expenses all costs annually over a period of years, including the cost of equity capital. The “escalating supply price” approach defers return on equity into the future.  The payback period for the project is more than 50 years.

Stan Marshall is quite right when he describes the cost recovery system for the LIL as the “public utility” model which applies to a regulated utility (like Newfoundland Power), while the other is a “non-regulated” approach. He explained that Emera wanted a defined return on its investment in the LIL with regular annual payment of dividends. The problem is that the two metrics cannot be combined to achieve meaningful results. They must be placed on a common standard if they are to be compared. We need make it no more complex than that.

A Common Metric
The differences in the annual costs used to calculate power rates relate to how the return on the equity invested by the province is treated by Nalcor. The regulated approach is used for the LIL while the non-regulated approach is used for generation (Muskrat Falls power plant) and for LTA.

Two alternative approaches exist to find a common standard. One would be to put the LIL on a non-regulated standard. The second would be to place the generation and transmission on a regulated standard. We have chosen the latter in order to ensure that the return on equity is not ignored. It is a real cost and to ignore the cost of capital borrowed by the province would be to understate the actual cost. There is yet again another alternative, similar to the regulated approach, which would reduce the rate of return on provincial equity to reflect the true cost of borrowed money to the province. For simplicity we have avoided this line of inquiry in order to make our numbers as comparable to Marshall’s numbers as possible.

What we have done is to take the regulated LIL return on equity numbers as the standard and to scale the other numbers to place them on the regulated standard. To do this we focus on the LIL return on equity numbers which are based on a return of 8.5% for LIL shareholders, both Emera Energy and the government of Newfoundland and Labrador (GNL). The return on equity set in the Power Purchase Agreement between Muskrat Falls Corporation and NL Hydro is 8.4%.

We have taken the LIL return on equity calculated by Nalcor for 2021 and adjusted it in two steps to calculate the ROE numbers for MF and LTA. For MF the ROE has been adjusted upward to reflect the higher direct construction costs of MF compared with LIL. The first step is simply to scale the dividend numbers for MF + LTA, the ROE, in relation to those for the LIL.  The loan guarantee requires a higher proportion of equity in MF and LTA than in LIL so the ROE has also been adjusted upward by 40% to reflect a minimum equity ratio of 35% compared with 25% for the LIL. This is the second step. The same approach has been followed with the LTA. The 2021 revenue requirements used in these calculations are shown in the Appendix. The total of the 2021 revenue requirements shown in the Appendix is $808 million and when divided by 4641 GWh we get 17.42 cents/KWh the number used by Mr. Marshall. These calculations then give us the new equity values to be inserted for each of MF and LTA in the table show in the Appendix. 

The Results
The numbers we get are as follows:

LIL cost per KWh: 8.98 cents (unchanged)

LTA cost per KWh: 2.46 cents (compared with 1.44 cents)

MF cost per KWh: 12.78 cents (compared with 7 cents)

Transmission Costs: 11.44 cents (compared with 10.42 cents)

Cost of Muskrat Falls power 24.22 cents (compared with 17.42 cents)

Blended cost of Muskrat Falls per unit of total energy sales: 16.2 cents (compared with 11.66 cents)

Blended cost of power: 27.4 cents (compared with 22.89 cents, as explained below)

Revenue requirements for Muskrat Falls
The Nalcor presentation of February 15, 2018 is based on an output of 4,641 GWh at Soldier’s Pond, net of line losses. Multiplying this by 17.42 cents/KWh gives us $808 million as the revenue required to pay for Muskrat Falls in 2021. If we use the 24.22 cents and multiply it by the net output of 4,641 GWh we get a much higher number for the revenue requirement to be recovered in rates, namely $1,124 million compared with $808 million using 17.42 cents/KWh.

The Nalcor presentation includes the following table:
Revenue requirements for full system
This table enables us to calculate the annual costs of the existing power system to which the cost of Muskrat Falls must be added. The numbers in the table and chart above are the costs when blended over the projected sales volume of about 7 billion KWh. We have NL Hydro costs other than those related to Muskrat Falls (generation, transmission and distribution) of 6.80 cents per KWh and Newfoundland Power (generation and distribution) cost of 4.43 cents per KWh. This results in a system cost of $778 million, excluding Muskrat Falls. 

If we add this to the $1,124 million we get $1.9 billion as the total cost of the electrical power system in 2021. When we divide this by the 6.93 billion KWh projected in the load for 2021 we get a blended system cost of 27.4 cents per KWh instead of the 22.89 cents in the Nalcor presentation. So the result is to increase the unit cost of Muskrat Falls power but also to raise the overall blended system cost.

When we divide the adjusted revenue requirement for Muskrat Falls by the 6.93 billion KWh projected the result is a cost per blended unit of 16.2 cents KW/h compared with the 11.66 cents KWh in the table above.

The chart and table below show that only 1324 GWh of Muskrat Falls power will be used in 2021. Using the revised cost per KWh of 24.22 cents and the revenue requirements of $1,124 million we find that the cost per KWh is 84.9 cents. The 24.22 cents spreads the cost over all units of output, namely the net plant output of 4,641 GWh, while the 84.9 cents spreads the cost over only the amount of power we actually use. 

This illustrates that while Muskrat Falls saves fuel costs it adds large financing costs because of the huge amount of capital involved. The savings in fuel costs are far outweighed by the incremental capital cost. The problem with Muskrat Falls is that ratepayers have to bear the full cost even though they are using only 28% of the power. Holyrood with its fuel cost of 15 cents per KWh would have been a much cheaper option but not necessarily the best.

Without making these adjustments Stan Marshall is ignoring the impact on project costs of the large equity return on generations assets (MF + LTA), which is left out of costs in 2021, the first full year of Muskrat power. Stan's 7 c/kwh is simply not a real number without consideration of return on equity.  The deferred return on equity is an IOU of almost 5 cents/KWh and is a real cost that is going to be collected later in the 50 year payback period, compounded at 8.4%, making the future cost much higher.  Either way, the cost of generation is far from a bargain.

The notion that Muskrat Falls is a bargain cannot be defended. When disparate metrics are combined the result is nonsense. When we place both the regulated and non-regulated costs on a consistent basis we get totally different results. Generation costs exceed the cost of transmission and the overall costs far exceed the cost of alternatives.

What has to be kept in mind here and in the discussion of rate mitigation is that the magnitude of the cost is so high that it is likely impossible to recover all costs through rates. The blended costs of 27.4 cents per KWh cannot be recovered in rates, nor could the 22.89 cents. Marshall has failed to demonstrate how the project can be promoted as a bargain or indeed as comparable in cost with other projects across Canada, without ensuring that the cost recovery methodology is consistent among projects. Caveat emptor, let the buyer beware! 

The “non-regulated” model of cost recovery chosen by Nalcor was intended to avoid rate shock by keeping rates lower in the early years than they would otherwise have been. It was this model that produced blended rates of 22.89 cents/KWh instead of 27.4 cents/KWh. So it builds in about 5 cents of rate mitigation. However the reality is that the cost of provincial equity is a cost the taxpayer cannot avoid. The province can write off the ROE but that does not relieve the Treasury of the cost of paying interest on the debt it incurred to make its obligatory equity commitment.

The use of the hybrid cost recovery was an attempt at rate mitigation. The use of such complex arrangements should not blind us to the financial impact of the project. It is an attempt to defer costs into the future. We have to be very wary about rate mitigation which takes money away from other programs today or else defers costs and adds them to the legacy of costs we are bequeathing to future generations.

As we can see here the combination of different metrics can seriously mislead us in the interpretation of the problem we are facing, minimizing the real cost of Muskrat Falls and its impact. We are depending on the Muskrat Falls Inquiry and on the Public Utilities Board to present the unvarnished truth along with consistent metrics for measuring the problem and assessing potential solutions.

David Vardy

The data used to calculate the 2021 revenue requirements using the “regulated” approach come from the response to my ATIPPA request to Nalcor PB-651-2017 below. We have chosen to use these data rather than the later revenue requirements contained in PB-954-2018 because these are the data used in Stan Marshall’s February 15, 2018 presentation at Memorial.