Electricity Market Reform

A sensible commercial framework for nuclear power, Alex Henney, EEE Ltd

Summary

There has been much wheel spinning (and in the case of the Renewable Obligation substantial waste of customers’ money) considering how to meld expensive low carbon plant into a market. There should be a clear understanding that nuclear power plants (and windmills) are not being developed for economic reasons – they are manifestly not economic compared with gas plants – but to meet government environmental policy. Consequently the energy market as an investment mechanism is not a relevant consideration . T he financial risk of such plant should not be increased by exposing them to irrelevant risks ; r ather the aim should be to insulate their revenue from market price risk, and furthermore to minimise the cost of capital.

The British government has a long and lamentable history of incompetence in dealing with nuclear power, which has led to wasting many tens of £billions. It should learn from practice in other jurisdictions, notably Ontario and Georgia, where there are significant nuclear developments:-

· In Ontario the Bruce Power A nuclear plant is being refurbished at a cost of Can$5.25bn. The power is being paid Can$63/MWh indexed by the Consumer Price Index. The developer assumes the construction and availability risk. The post-tax weighted average cost of capital (WACC) is reported as being in a range 10.6% to 13.8% nominal. The contract is open book to the Ontario Power Authority

· Georgia Power Company (GPC) is the developer and has a 45.7% share in the ownership of the Westinghouse AP1000 Vogtle 3 and 4 units (the other owners are municipal entities). The construction cost of the contract is $9.8bn while including interest during construction the cost is estimated at $14bn. GPC’s share of the plant is being built against a price approved by the Public Service Commission of Georgia, and the cost will be included in the rate base and allowed a WACC of 7.8% nominal post-tax. The price is based on an Engineering Procurement Construction contract, and the amount allowed in the ratebase will be subject to a performance bonus/penalty. The contract is open book to the Commission, who have recruited an experienced nuclear engineer to assist with the formalized monitoring procedure. In December 2010 the Commission authorized expenditure of $1bn on GPC’s share of the work so far. As of the first week in March both the company and the Commission expressed satisfaction with the arrangements

Common features of both schemes is that they insulate the plants’ revenue from any market; they have open-book contracts; and the resulting output is "blended" with other electricity.

In the Energy Market Review DECC set its face against a regulated asset base approach for reasons that hold little water. Unfortunately DECC officials do not appear aware that GPC’s costs (and costs of thermal plant in California) are incorporated into the asset base on a regulated base. Mr. Atherton of Citigroup reported to the Committee that EDF will be seeking a post-tax nominal return of 10.5% (and other developers may seek more) .  According to Citi Investment Research’s model, using a construction cost of €3200/kW (which is in line with EDF’s claim of £9bn for 3300MW) and their other assumptions, the resulting cost of nuclear power would be £74/ MWh.  With the 7.8% return of the Georgia Power financial framework, the resulting cost of nuclear power would be £56/MWh , which is 24% lower and allows significant cost overrun yet still leaves the customers ahead.

It thus seems to me to make eminent sense to eschew talk of markets and their imagined disciplines (particularly as we have destroyed ours ), but rather to follow the approach adopted in Georgia . Plant should be developed to an agreed cost based on Engineering Procurement Construction c ontracts which are open book to the relevant authority (which may be the government or Ofgem or a special agency set up for the purpose) and subject to expert monitoring review , with a performance payment/penalty at the margin. The cost would be the regulatory asset base on which the company would earn an appropriate – but modest – return. The resulting electricity would be blended with other electricity.

A disastrous history with nuclear power

The British government has a lamentable track record of incompetence and wasting taxpayers money in dealing with nuclear power, which I set out in detail in "The Economic Failure of Nuclear Power in Great Britain" [1] ; "A study of the privatisation of the electricity supply industry in England & Wales" [2] ; and "The British electric industry 1990-2010: the rise and demise of competition" [3] . The economics of the Magnox reactors were fudged; most of the AGR programme was an economic disaster; the Sizewell Inquiry was intellectually fraudulent, if not downright dishonest; Sizewell B cost 40% more than estimate and was written down by £800m for the sale of British Energy. Following the debacle of the withdrawal from the privatisation of the nuclear power stations in November 1989 the House of Commons Energy Committee conducted an inquiry into "The Cost of Nuclear Power" [4] because:-

"After years of official assurances that nuclear power was (or could be) the cheapest form of electricity generation, Parliament and the public are entitled to know why it was only when faced with the commercial discipline of life in the private sector that nuclear power (from both existing and proposed reactors) suddenly became an expensive form of generation…we believe the Department of Energy, as the CEGB’s sponsoring department, must share the blame for this, since it apparently made no attempt to obtain realistic costings from the CEGB until it was seeking to privatise nuclear power…The manner in which the Department has supervised the CEGB over the years can only be described as inadequate."

By the early 1990s it had cost the British taxpayer and electricity customers more than £10bn (in 2010 prices) in research and development and about £50bn in capital expenditure, with another nearly £4½bn for Sizewell B [5] . Then came about £7.5bn for the nuclear levy in the 1990s, then the government devised an unwise approach for British Energy to pay for decommissioning, which contributed to its demise. On top of this is a bill of an undiscounted cost of discharging all future civil nuclear liabilities conservatively estimated at about £100bn [6] . This is an awesome bill for "electrical energy in homes that is too cheap to meter." [7]

If the British government is going to mess again with nuclear power it behoves it to get its act together properly. In particular it should learn from others.

Nuclear development and the concern about price risk

Over recent years there has been much discussion about how to facilitate financially the development of nuclear power plants because of the power price risk, see exhibit.

Source: Newbery [8] .

In addition to the obvious point that the lower the gas price which drives the electricity price the less favourable is the economics of nuclear, the graph also brings out the volatility of gas and electricity price which are closely linked. Although gas (and coal) plant are automatically hedged against the volatility, as Professor Newbery has pointed out, nuclear is not:-

"The price of electricity in the forward market moves very closely with the cost of generating using either gas or coal, allowing for the cos t of CO2 required for each. Although the prices of gas, coal, CO2 and electricity are separately highly volatile, (gas prices have fluctuated between 20p/th and 110p/th and coal has fluctuated from $50-200/ton between 2004-8) the forward clean spark spread and the forward dark green spread have remained far more stable. The reason is simple, the price of electricity is set by the cost of generating using the marginal fuel and the CO2 price moves to equate the marginal costs (including the EUA cost) of coal and gas. Companies with fossil generation are therefore naturally hedged against fluctuations in the input and output prices, while low-C electricity, whether renewables or nuclear, is exposed to the full volatility of the electricity price, as its variable costs are low, predictable and stable."

The potential price risk to nuclear plant in a gas price driven market was more than adequately shown by the demise of British Energy in September 2002 following the collapse of prices that began in the winter of 2000/01. The wish to mitigate price risk has led to proposals for a carbon floor price and now in the government’s Energy Market Review a contract-for-differences [9] . But in my view much of the discussion has been misplaced because it has started from the stance that the nuclear plants should operate within the framework of the market. Part of this storyline is that exposure t o the market price provide s the conventional i ncentives to construct and operate the plant efficiently.

In my opinion this approach is fundamentally flawed and results in a higher cost of capital than needs be, and consequently a higher cost of electricity . This has been illustrated by the Renewables Obligation , which was a case example of an ill judged, complex and expensive scheme based on ersatz market principles . The starting point should in my view be that nuclear power plants (and windmills) are not being developed for economic reasons – they are manifestly not economic compared with gas plants – but to meet government environmental policy. Consequently the energy market as an investment mechanism is not a relevant consideration . T he financial risk of such plant should not be increased by exposing them to irrelevant risks ; r ather the aim should be to insulate the ir revenue from market price risk, and furthermore to minimise the cost of capital. This approach has been adopted for both the renovation of the Bruce Power A plant in Ontario , and for the development of the Vogtle 3 and 4 units by Georgia Power Company in the US .

The Ontario approach to financing a large nuclear refurbishment

In 2005 the Ontario government announced that it has reached an agreement with the owners of Bruce Power A to refurbish the plant for a cost of Can$4.25bn. Subsequently in 2007 the agreement was extended to extend the refurbishment for an additional Can$1bn, resulting in a total investment of approximately Can$5.25bn. In return, the provincial government, through the Ontario Power Authority (which is guaranteed by the provincial government), agreed through a contract-for-differences to pay an initial price for electricity of Can$63/MWh as of the date the Refurbishment Ag reement was signed . This price is indexed by the Consumer Price Index. The Can$63/MWh is the only income the generator receives; consequently the plant owner bears both the construction risk and the availability risk.

According to the Office of the Auditor General of Ontario "Bruce Power and the Ministry agreed to an "open-book" process, and the Ministry was given access to a data room containing confidential documents provided by Bruce to support the refurbishment plans, supplemented by management presentations, facility site visits, and meetings with relevant government agencies" [10] . The weighted average cost of capital ( WACC ) was a post-tax return in the range o f 10.6% to 13.8% nominal.

The Vogtle 3 and 4 reactors in the US

The US Department of Energy is offering a loan guarantee to up to four nuclear plants in order to help pump prime development [11] . The loan guarantee reduces the cost of development by lowering the cost of debt, but is not by itself a critical factor in underpinning development.

The first scheme for which conditional [12] loan financing of $8.3bn covering 70% of the construction cost has been agreed for the Vogtle 3 and 4 Westinghouse AP1000 reactors with a total capacity of 2200MW in Georgia which is being developed by Georgia Power Company, an investor owned utility. It is taking a 45.7% share of the scheme and is the agent for development on behalf of the other owners, who are municipally supported generation production and supply entities.The construction cost of the scheme is $9.8bn ($4500/kW); the cost including financing is $14bn. The loan guarantee effectively reduces Georgia Power’s cost of borrowing by $15-20m p.a.

Georgia Power will incorporate its share of the construction cost of the plant in its rate base, where it will represent an increase of about 34% of the current rate base (the financing charges during development are paid off as incurred). When the plants are operating their costs will be blended with the other generation costs together with transmission and distribution costs to construct the tariffs in the traditional US manner for a vertically integrated utility – there is little or no market based input. The Public Service Commission has certified a sum of $6.11bn for Georgia Power’s share of the plant including the interest incurred during construction ($6080/kW) based on an Engineering Procurement Construction contract for the plant which is described as price defined turnkey contracts with Westinghouse and architect/engineer Stone & Wesbter. (Stone & Webster is a subsidiary of the Shaw Group, which also has a 20% interest in Westinghouse; Toshiba has a 65% interest). In December 2010 the Commission authorized expenditure of about $1bn for Georgia Power’s expenditure thus far (representing about half of total expenditure) on procurement of long lead time items; excavation and preparation of foundations, and construction of buildings. The contract has provisions that share cost overrun between the contractor and the company. Although this is the first AP1000 built in the US it is benefiting from the experience of the 3 AP1000s being built in China which have similar nuclear islands and are three years in advance. There is extensive interaction between people in Georgia and in China.

The staff of the Public Service Commission proposed an incentive scheme for the company consisting of a bandwidth of +$250m. If the in service cost of the project is less than the lower threshold, then the Company could earn an incentive return of 10 basis points in the return on common equity for every $100m the in-service cost is less than the lower threshold of the bandwidth. If the completed cost of the project is more than the upper threshold, then the return on common equity would be reduced by 10 basis points for every $100m the completed cost is over the upper threshold of the bandwidth. The company objected to this, and the Commission upheld it and asked the company to see if they could negotiate a mutually agreeable incentive scheme. The negotiation is still in progress, but is hoped it will be concluded by the end of March 2011.

The Public Service Commission has appointed an experienced ex-Westinghouse consultant to monitor performance with an allowed annual cost of $600,000. The Public Service Commission requires a Construction Monitoring Report every 6 months that includes actual expenditures for the January - June and July -December period s together with a narrative of activities and a forecast of expectation of final project cost.  I t must also include an analysis to show whether it is cost effective to move forward.  Costs are certified in the subsequent 180 day period if deemed prudent . The contract is open book to the Public Service Commission.

There is no target load factor in the contract, and so at first sight the ratepayers assume all of the load factor performance risk. But the company has a performance based ratemaking scheme with a target allowed return on equity of 11¼% pre-tax nominal and a deadband of +1% within which the company earns what it earns. If the company earns over 12¼% in a year it keeps 1/3 of the earnings and returns 2/3 to customers; if it earns less than 10¼% it can request a filing to raise tariffs. Thus since the plant will represent about a third of the company’s rate base its operating availability will have a noticeable impact on the company’s profitability – thus implicitly there is a link between plant availability and company profitability.

The company’s capital structure is approximately 43% debt; 10% preferred; and 47% equity; its cost of debt is about 5.8%; its cost of preferred stock is about 6.1%; its allowed pre-tax return on equity is 10¼-11¼% so its post-tax WACC is about 7.8% nominal.

As of the beginning of March 2011 discussion with officials of Georgia Power Company and the Public Service Commission found satisfaction both with the arrangements in place and the progress on the plant.

The foreign lessons for a commercial framework for nuclear

In both Ontario and Georgia the market price risk is removed from the developer. In Ontario the construction and availability risk remain entirely with the developer, while in Georgia the developer bears some of the construction risk at the margin through a profit incentive scheme, and the company has an incentive to achieve a good level of availability through its general performance based ratemaking scheme. The Georgia approach results in a lower WACC (8.1% cfr 10.6% to 13.8% for Bruce Power).

The Energy Market Reform Consultation paper argues against the Georgia approach, which is effectively a regulated asset base (RAB), commenting "It would represent the most fundamental change to the current arrangements of all the options; making such a radical change would be high risk. Moving to a RAB system would require the Government to sacrifice all market benefits and competitive pressures for greater efficiency, optimal operation and innovation that could be retained under other options considered as part of this project. The generation sector – where competition is viable and a key feature of the current market – is different to the natural monopoly market for the provision of transmission and distribution networks. As such, the Government does not consider this an attractive option for reform" (p66). The paper also argues that "the approach transfers construction risk, which generators are better able to manage, to customers."

In my opinion this rejection is not a soundly based regarding nuclear plants.

1. It is not at all obvious that this approach would in fact "sacrifice all market benefits" because those of optimal operation can be retained by a suitable contract structure

2. Innovation in nuclear design is a slow process pioneered by manufacturers working for an international market and subject to an elaborate licencing process; it is not influenced by the British power market.

3. I do no consider that "competition is viable and a key feature of the current market." With (i) the development of the complexities and economic distortions of NETA/BETTA; (ii) the vertical and horizontal consolidation of the industry into an oligopoly; (iii) the lack of liquidity in the contract market; (iv) the subsidies for renewables and quasi-planning for so much wind/4 CCS plants/possibly nuclear, the market long ago lost the semblance of competitiveness. It is overdue time we gave up the pretence that we have much of a generation market, let alone a competitive one, and recognize that flooding it with subsidized windmills, nuclear plants and CCS plants will destroy it entirely.

4. The claim that the approach transfers construction risk is a statement of hope, rather than of realization which will depend upon the eventual CfDs. It is perhaps noteworthy that not so long ago EDF Energy was claiming that it could bear the market price risk. Then it wanted a carbon floor price to mitigate part of the market price risk; now with a CfD it has probably got rid of the market price risk. Will it negotiate to the wire, then say that in the light of the cost overruns of the two EPRs being built [13] , it needs help with construction cost risk? Then as the government increased the ROCs for offshore windmills to get the London Array going, it requires little imagination to work out the government’s response.

5. The 7.8% of Georgia Power’s development of the Vogtle 3 and 4 plants compares with the 10.5% that Mr. Atherton reported EDF as seeking. (Other developers may seek more - DECC’s consultant (Redpoint) assumes an 11.2% hurdle rate for a nuclear plant with a CfD). According to Citi Investment Research’s model, using a construction cost of €3200/kW (which is in line with EDF’s claim of £9bn for 3300MW) and their other assumptions, the resulting cost of nuclear power would be £74/ MWh . With the 7.8% return of the Georgia Power financial framework, the resulting cost of nuclear power would be £56/MWh , which is 24% lower and allows significant cost overrun yet still leaves the customers ahead.

DECC’s lac k of understanding of the issue was perhaps shown by a presentation on 3 March at the Policy Exchange which included a slide which stated that nowhere was plant investment based on a regulated asset base – this view is wrong (see my submission "Capacity markets and reliability options") [14] .

EDF definitely , and perhaps other nuclear developers , have the government over a barrel if it seriously wants the development of nuclear plants . I t thus seems to me to make eminent sense to follow the approach adopted in Georgia . Plant should be developed to an agreed cost based on Engine ering Procurement Construction c ontract s which are open book to the relevant authority (which may be the government or Ofgem or a special agency set up for the purpose) and subject to expert monitoring review , with a performance payment/penalty at the margin. The cost would be the regulatory asset base on which the company would earn an appropriate – but modest – return. The resulting electricity would be blended with other electricity.


[1] The Economic Failure of Nuclear Power in Great Britain , Alex Henney, Greenpeace, 1989.

[2] The privatisation of the electricity supply industry in England & Wales , Alex Henney, published by EEE Limited, 1994.

[3] The British electric industry 1990-2010: the rise and demise of competition, Alex Henney, published by EEE Limited , 2011.

[4] The Cost of Nuclear Power, Volume I, Energy Committee, Fourth Report, Session 1989-90, HC205-II, HMSO, 7 June 1990.

[5] In addition to the shambles over “conventional” reactors , in current prices we wasted several tens of £bns on the fast breeder reactor and on the Thorp reprocessing plant.

[6] The assessment was provided by Professor Gordon Mac K erron, former cha irman of the Committee on Radioa ctive Waste Management, in an e-mail to Alex Henney dated 16/12/2010 .

[6]

[7] Admiral Lewis L. Strauss, first chairman of the US Atomic Energy Commission, 16 September 1954 .

[8] A Nuclear Future? UK government policy and the role of the market, David Newbery, paper presented to the Beesley Lectures on Regulation in London , 22 October 2009 .

[9] This is misleadingly referred to as a FIT CfD – presumably the “FIT” part of the term is supposed be a semantic sop to the terminology in the Conservative election manifesto.

[10] Special Review for the Minister of Energy, Office of the Auditor General,

[10] http://www.auditor.on.ca/en/reports_en/brucespecial_en.pdf .

[11] The Energy Policy Act of 2005 provides a number of incentives for “Innovative Technologies” which apply to “advanced nuclear energy facilities, including a loan guarantee for up to 80% of eligible project costs, http://www.ne.doe.gov/energypolicyact2005/neepact2a.html .

[12] The guarantee is conditional upon the Nuclear Regulatory Commission licensing the plant, which is expected late this year.

[13] The performance of the two European schemes being built highlight construction cost risk:-

[13]

[13] The 1600MW Finnish EPR (European Pressurised Water Reactor) being built by Areva at Olkiluoto is 4 years behind its original target commissioning in 2009 and the construction cost has increased from €3bn in 2003 money to €5.7bn (€3,600 or $4,600/kW). Areva NP is claiming compensation of about €1bn for alleged failures of Teollisuuden Voima Oy (TVO). TVO, in a January 2009 counterclaim, is demanding €2.4bn in compensation from Areva NP for delays in the project ( Agence France Presse , “Setbacks Plague Finland ’s French-built Reactor,” January 30, 2009 )

[13]

[13] In May 2006 EDF estimated that the construction cost of its EPR at Flamanville would be €3.5bn. Two years later An Areva official suggested that the cost will be at least €4.5bn, although it was not specified whether this was an overnight cost ( Nucleonics Week , “Areva Official Says Costs for New EPR Rising, Exceeding $6.5 billion,” September 4, 2008, p. 1)

[13]

[14] I provided DECC in August 2010 with a paper wh ich clearly explained in some detail that in the California energy market thermal generators are being built to a return on a regulated asset base, and is Georgia Power investment in Vogtle 3 and 4, “ A multiclient project on developing trading arrangements for a windy electric industry”, Alex Henney, EEE Ltd, July 2010.

[14]