by Benjamin Tait

Liberalisation, Spanish style

1 March 2000

The liberalisation of the Spanish electricity market has forced nuclear operators to face increasing competition. The efforts made by Spain’s nuclear power industry to live with the market provide one of Europe’s most interesting test cases.

Until the early 1990s Spain’s nuclear power industry was part of state management of the entire electricity industry. But in the mid-1990s policies shifted in favour of liberal solutions. Top utility Endesa was privatised, regulatory scrutiny was increased and substantial utility efficiency gains were shared with customers through tariff cuts.

By January 1998, Spain was ready to launch a power pool and open 28% of its power market to competition by allowing its largest customers to choose their suppliers. This opening easily beat the February 1999 deadline for compliance with the European Union electricity directive. Today, 42% of the market is open, rising to 53% in July 2000. By 2007, or earlier if the government changes its mind, all Spanish customers will be free to choose their suppliers.

Controversy remains over how open the market really is, but Spain has undoubtedly earned a place in the European market reform camp, abandoning the region’s dwindling band of anti-market conservatives.


One thing that fares badly under competition is potentially market-sensitive information. Grid operator Red Eléctrica de España (REE) stopped publishing estimates of production costs by technology in 1998. These were part of the old Marco Legal y Estable (stable legal framework), which fixed utility cost recovery. The generators too are coy about the inner workings of their businesses. But there is enough information around to suggest that nuclear power can be an economic winner, if not yet a great money-spinner.

The REE 1997 estimates put nuclear fixed and variable costs at Pta 7.69/kWh

(US 4.6¢), in line with the average production costs of the power industry as a whole. Since then, operating costs have been squeezed. Utility personnel costs across the industry dropped 3.3% in 1998, according to data from utility association Unesa. Official reports for 1999 should show a similar fall. No specific data are available for the nuclear part of the payroll, but nuclear plants have not been spared from cost cutting. For example, in June 1998 Endesa and Iberdrola joined their Ascó and Vandellós plants under one management with unified technical and administrative support.

Fuel too is getting cheaper. Iberdrola, Spain’s second largest utility, has reported that its nuclear fuel costs were cut by 10% in 1999 to Pta 0.66/kWh (US 0.4¢), saving the group Pta 4 billion ($24 million). By contrast, REE 1997 estimates put nuclear fuel costs at Pta 0.82/kWh (US 0.5¢). Iberdrola has secured savings from state-owned fuel supplier Enusa, which reduced its prices after it lost its monopoly, and on fuel purchases from ABB. Iberdrola says it has also developed more efficient fuel placement methods.

All this adds up to average variable costs (fuel and O&M) of around Pta 2-3/kWh

(US 1.2¢ - 2.4¢), according to industry sources. And more cost savings are on the way. All utilities have ambitious restructuring plans for the next three to five years. Fuel cost reductions will continue too. Iberdrola expects to book a 23% reduction on 1998 per kWh fuel spending by 2003.

The old REE estimates also included a significant depreciation charge. This assumed a twenty-five year reactor life. But in 1998 Endesa and Iberdrola reviewed their assets to determine their “effective useful lives” and decided to extend nuclear plant depreciation periods to thirty years. This shaved Pta 12 billion ($72 million) off Endesa’s 1998 depreciation charge and Pta 7 billion

($42 million) off Iberdrola’s charge. If “useful life” proves in fact to be forty or even forty-five years, depreciation charges could be slashed again. One industry source says a forty-five year life would translate into a 35% reduction on the old regulated system cost estimates, even before accounting for savings achieved in operations.

Finally, utility finances have improved dramatically in recent years. Long term utility debt was Pta 2.2 trillion ($13.2 billion) by 1998, compared to a debt mountain of Pta 3.2 trillion ($19.3 billion) in 1990 and Pta 2.8 trillion ($17 billion) in 1995. Much of this debt had been accumulated during the nuclear building programme. The financial component of any estimate of nuclear fixed costs must account for this improved financial position.

Markets and policy makers have both boosted utility finances. Investors have amply rewarded Spain’s successful 1990s transition from high inflation and poor state finances to low inflation and qualification for European monetary union. Interest rates tumbled up to the launch of the euro in 1999, reducing the cost of utility debts. Utilities also secured funding for debt reduction from the 1996 securitisation of "compensation rights" for the nuclear projects stopped by the moratorium. This exercise raised Pta 715 billion ($4.3 billion).

Despite the increasing competitiveness of the nuclear sector, its output still appears to be too expensive. A fair assumption – confirmed by some if not all sources – is that a nuclear plant can deliver power profitably today for Pta 6.5 /kWh (US 3.9¢). This is not good enough – the average Spanish pool price over 1999 was just Pta 5.85/kWh

(US 3.5¢).

This is where policy comes in. Spanish utilities are entitled – their word – to stranded cost payments for generation assets that cannot compete in an open market and also recoup investments made before deregulation. These payments could total as much as Pta 1.7 trillion ($10 billion) for the power industry as a whole between 1998 and 2007. Around Pta 0.5 trillion ($3 billion) of this total is related to nuclear plants, according to industry sources.

In 1999, the authorities set the system stranded costs charge (excluding domestic coal support) at Pta 81 billion ($490 million), or Pta 0.66 (US 0.4¢) per kWh of system production. Again for argument’s sake, let’s assume the nuclear share of those payments is the same as the average charge of Pta 0.66/kWh, in line with the REE 1997 estimates putting nuclear costs on a par with system-wide costs. Adding this revenue to the well-run nuclear plant’s pool revenue shows the plant covers all of its costs, including return on investment.

In practice, however, stranded cost calculations and money flows are more complex. Stranded costs are accounted for at the industry and group levels, not on a plant-by-plant basis. An individual plant has no assurance of achieving the average market price used in this estimate – a nuclear plant’s average price is in fact likely to be slightly lower. Furthermore, the return on investment in the old regulated system’s estimates was based on annual policy-based calculations of investment that granted only low paybacks in a plant’s early years, not actual investment. This makes precise comparisons between the old and new systems artificial and misleading. But the existence and sheer scale of stranded costs provides another reason to suspect that efficient nuclear generators can book decent profits.

Published divisional financial results provide more concrete evidence. Generation division figures include all utility generation, not just nuclear power. But if nuclear power is an accounts black hole, one can only assume other assets are implausibly profitable. For example, Endesa’s generation division posted operating profits of 928 million euros on revenues of 2858 million euros for the first three quarters of 1999. This makes for an operating margin of 32.5%, the highest of Endesa’s utility divisions.

Some argue that the divisional figures reflect accounting abstractions required by regulators, not reality, but the utility bottom line is real. Net income at Spain’s four utilities rose between 9% and 17% over the first three quarters of 1999, in large part thanks to a good performance at home that balanced setbacks in Latin America.

So do stranded cost payments eliminate market discipline? The easy answer is no. Under Spain’s market opening legislation, stranded costs are collected and accounted for outside the market on a monthly basis. They are also subject to annual revisions. Day-to-day operations, by contrast, are market-based. Almost all utility output is channelled through the pool, where generators offer prices on an hourly basis. Put simply, these offers are stacked by price until the “system marginal price” for each hour is reached – the highest priced offer required to meet demand side bids for the next day. Generators that price too high will simply not be called on to generate.

In practice stranded costs distort what is happening in the market. When the market opened in 1998, the government set a reference pool price of Pta 6/kWh (US 3.6¢). If the average price a generator secures rises above this reference level over the course of a year, any stranded cost payments it is entitled to are reduced. This gives generators a price target and a trading interest that new entrants cannot share. If the Spanish system’s critics are right, pool trading tends, over the course of a full year, to be little more than careful adjustment of revenues and dissuasion of new entrants by Spain’s four utilities.

Even worse, from the critical point of view, is the government’s decision in September 1998 to allow utilities to ‘securitise’ around Pta 1 trillion ($6 billion) of their stranded cost ‘entitlements’. This move – which the European Commission has yet to approve – would be funded by a 4.5% levy on regulated tariffs in place of the monthly settlements and annual revision system. Securitisation might lessen pool trading distortions, but it would not eliminate them. Spain’s energy industry regulator, the CNE, has made its own doubts about the plan abundantly clear. It sees stranded cost calculations as ‘insurance’ to be renewed each year in light of market conditions, rather than entitlements to be funded through a guaranteed share of the tariff.

Market power is another favourite topic for critics (reportedly including the OECD). Endesa and Iberdrola collectively share around 70% of Spanish system generation and 80% of retail supply. This share was secured in part through 1990s consolidation and will be defended by a string of new gas projects that add up to more than the collective project efforts (so far) of new entrants. The top utilities' strong position is especially impressive when set against limited international interconnection capacity. As of January 2000, import capacity of 2050 MW was commercially available, compared to peak winter demand of over 30 000 MW. Much of this capacity is taken by the utilities or system operator REE under a long term contract with Electricité de France, leaving very little room for cross-border traders such as Atel and Electrabel.

There is a further market peculiarity for nuclear operators. They aim to run at baseload and cannot easily adjust output on an hourly basis, unlike more flexible price-setting capacity that covers mid-merit and peak demand. The best way to ensure smooth operation is to ‘bid zero’ – offer power ‘for free’ to the market to ensure that the plant is called on to generate.

But nuclear power is not free – all generators that are called on receive the system marginal price. This is arguably another Spanish market distortion. The UK has decided to reform its trading system for England and Wales in part to force its own big ‘zero bidders’ into the market, where they will have to secure bilateral deals at real prices to cover the bulk of their output.

Spanish utilities have little time for most critics. They insist the generation market is competitive. Competition for the industrial supply business is also real and gathering strength, as reflected in pool data. The policy reasoning behind stranded costs is sound and fair and removes damaging uncertainty. Utilities have slashed the amount they are due under stranded cost calculations by half. Regulated prices for customers who are still not free to choose their suppliers have dropped in real terms. More price cuts are on the way. Finally, utilities argue, Spain is swarming with new entrants. They are not limited to traditional trading partners turned rivals, such as Electricidade de Portugal and Electricité de France. New entrant ranks also include some of the biggest and wealthiest US and northern European energy groups, such as Enron and RWE. Within five years, the newcomers could have some 5000 MW of new gas-fired capacity in operation, backed by aggressive trading operations and independent fuel strategies.

The government backs up many utility arguments, most notably through the Ministry of Industry and Energy. Utilities and the government have also proved very difficult targets for the regulator, the CNE. It has launched stinging attacks on policy moves and industry contentions, but it has failed to force radical change. Furthermore, the market operator, Omel, has built a transparent and well-regarded platform for trading.

But scrutiny at the European level may yet force Spain to tackle its alleged distortions. The European Commission is currently reviewing the stranded cost policy and the securitisation plan, and is taking a suspiciously long time to make up its mind. Utilities are publicly confident, but other observers think Spain’s case for stranded costs is in deep trouble. If the Commission objects, one of the foundation stones of current Spanish power policy will be at risk. Pool prices could surge to make up for a policy ‘loss’.

The long term

Stranded costs cover the period up to 2007, but several Spanish nuclear plants should continue operations well beyond that date. Utilities assume thirty years of operation. Yet the Jose Cabrera plant has already passed the 30 year mark and is still in operation, with a permit valid to October 2002. Only one more of Spain’s reactors – Garoña – will pass the 30 year mark before 2011.

The following decade will be more interesting. By 2017, all reactors will have reached the 30 year mark. 30 year life assumptions mean plants’ fixed investment costs will have been covered by this time, reducing their production costs to fixed O&M and variable costs plus any required capital investments. This could make them clear economic winners, reaping golden years’ dividends for as long as utilities can keep them running. Prudent new entrants in Spain may want to factor reactor lives of 40 years or more into their project business plans.

Recent utility investments suggest operators are aiming for longer lives and better performance. In 1998 and 1999, capacity was increased at five reactors by a total of 173 MW.

Emissions arguments in favour of nuclear power may gain ground in the longer term. For now, ‘dash for gas’ is the big environmental issue. As much as

20 000 MW in gas-fired capacity may come on line over the next 10-15 years, knocking older conventional thermal plants out of the market. This should improve Spain’s emissions performance. But gas-fired capacity is only ‘clean’ relative to oil and coal. If combined cycle plants replace nuclear plants too, Spain’s emissions performance will deteriorate.

High level waste disposal is a wild card. Spain has postponed a decision on a disposal site and method until 2010, leaving nuclear plants to accumulate waste on site. But Spain cannot wait much beyond 2010 to take action. The Trillo plant has already reached the point where it must expand temporary on-site storage. All other reactors have increased storage capacity in recent years and will reach pool saturation between 2013 and 2022, according to the latest government waste plan. State waste company Enresa’s low and intermediate waste facility is currently 28% occupied and should reach saturation in 2016. Even the estimates these numbers are based on may prove optimistic. Spanish power demand is officially expected to rise by 3% per year, but it is currently surging at 7% per year. Nuclear plants are expected to operate – and produce waste – only 7000 hours per year, for a capacity factor of 80%. In 1999 the industry-wide nuclear capacity factor was 88%.

Money for addressing the waste issue is also a potential problem. A funding mechanism sets aside 0.8% of tariffs. But at the end of 1998, the fund for future spending stood at Pta 242 billion ($1.5 billion), compared to an estimated cost from 2000 to 2070, in today’s money, of Pta 1.3 trillion

($8 billion). This burden could mount further if solutions prove more expensive than expected, if electricity prices fall, if inflation runs at a higher level than the estimate of 2% in the waste plan, or if the discount rate falls below the planned 2.5%. Any financial policy burden would fall on the government, which bears ultimate responsibility for decommissioning and waste management. Utilities are only responsible for operating waste management at their plants.

Faced with so many short and long term uncertainties, any observer of the Spanish nuclear business must recognise that the conclusion of any analysis may be misleading, dangerous, or just plain wrong. But the long list of variables is not the most important story. Almost all of the deep changes in the late 1990s Spanish power business can be traced back to one policy shift – the market should decide, not the state.

That the market doesn’t decide all the time is a fact of political and commercial life. Markets will always have enemies. And successful reform takes time, as fallen reform leaders such as the UK know only too well. But the Spanish nuclear sector, like the rest of the Spanish power industry, has already delivered some of the competitive goods, benefiting customers as well as shareholders. If the market is allowed to do more of its useful work, the future will be even better. Perhaps the next step should be tackling reactor ownership swaps.

Privacy Policy
We have updated our privacy policy. In the latest update it explains what cookies are and how we use them on our site. To learn more about cookies and their benefits, please view our privacy policy. Please be aware that parts of this site will not function correctly if you disable cookies. By continuing to use this site, you consent to our use of cookies in accordance with our privacy policy unless you have disabled them.