The Beesley Lectures: Lectures on Regulation
My task this evening is to review the new electricity trading arrangements (NETA) which will come into force quite soon. This review cannot be that of an objective outsider, since my role has more that of an insider. Thus, following Stephen Littlechild's invitation, I first served as a special adviser to Offer on RETA as it was then called; then following our chairman's invitation as a member of the Ofgem Management Board; and then finally before the summer helping to see the Utilities Bill through the Lords, including putting forward with Gordon Borrie an amendment which the Government became minded to accept to place a limit on the level of fines that the new regulatory authority can impose. I sense that my role is more that of accomplice than objective observer. And the key ringleader sits as chairman of this evening's session, an example of collusion that some might think very nearly, but not quite, matches the excesses of the Pool arrangement that will soon be consigned to history.
Nonetheless, tonight I wear my academic hat. My aim, therefore, is to step back from the fray and provide an objective account of the new arrangements. How successful I am in that I leave to you to judge.
May I make one final point before plunging into the substance of what I have to say. This longstanding lecture series was the brainchild of my late colleague, Professor Michael Beesley, after whom the lecture series is now named. Michael was the grandfather of the British regulatory model, and a major influence on many aspects of its practical operation. He was also an important influence on the reform of electricity trading, serving as Economic Adviser to Ofgem on NETA, working closely with Eileen Marshall and her team. I well remember Michael just a few weeks before his death in animated discussion about the details of the reform. We shall miss his contribution this evening, but it is apposite that this first lecture in the new Beesley lecture series should be on the reform that was so close to his heart just before his sudden and unexpected death.
The structure of this paper is as follows. The first part addresses a key issue concerning the Pool arrangement: whether the problem of generator dominance in the Pool arises from the degree of concentration in generation, or from the market rules. The former would point to the need for divestment; the latter for changing the market rules. I argue that, although both matter, reform of the market rules is crucial to delivering a well-ordered market. I then go on to review some other areas of controversy, both past and continuing. Some of these are issues on which I think the answers are clear-cut; others are ones where we may well need to learn from experience and allow evolution. I also consider those issues that will need to be addressed in the future. Finally, I reflect on two questions: what are the main achievements of the new arrangements relative to the old; and why is it that reform has been pushed through successfully on this occasion, whereas in the past it has been ducked, despite the overwhelming evidence of the need for change?
Two basic ideas, both learnt from Michael Beesley, have underpinned my thinking on NETA since I got involved in electricity trading reform. The first is that, on the whole, market relationships work much better than any form of regulation or administered market, not least in promoting innovation and change. The second is that no system is perfect, so that any system of electricity trading is likely to need periodic adjustment and reform. The major virtues of NETA are twofold: that NETA moves the electricity market much closer to that of a normal market; and that it puts in place a governance structure that allows for relatively easy adjustment and change. In both respects, it represents a major advance on the Pool.
Industry Structure versus Market Rules
The deficiencies of the Pool arrangement are well known and require little rehearsing. The arrangement was put in place just prior to privatisation to provide a rudimentary market mechanism whereby a privatised and decentralised electricity industry could operate in a way consistent with the need to maintain balance on the network and subject to the physical laws governing electrical systems. Essentially it consisted of grafting part of the previous centralised control system onto the market system. Generators were required to bid into a centralised market (the Pool), and the system operator (NGC) used the old computer programme (GOAL) to schedule generation to match demand in a supposedly efficient manner given the information in the bids.
Those who put this arrangement in place generally saw it as a quick fix to ensure that a privatised and decentralised system could function. They thought it would quite quickly be replaced by a more sophisticated arrangement. However, such change was impeded by the structure for the governance of the Pool arrangement put in place at privatisation. This was a mutual structure, relying on near consensus to effect change. The obvious difficulty with such an arrangement was that reform in the general interest could be blocked by vested interests. In practice, although there was some evolution, as GOAL morphed into Super-GOAL, it was quite limited. This was a severe weakness of the way electricity privatisation was effected. In defence, one can perhaps note that few appreciated at the time just how strong would be the vested interests against change.
These vested interests came from the ability of the large portfolio generators to manipulate the Pool price. This has been well documented in a number of places, and requires little reiteration. (See, for example, Armstrong, Cowan and Vickers 1994 for an early discussion, as well as Von der Fehr and Harbord, 1993, Wolak and Patrick, 1996, Brealey and Lapuerta, 1997, and Wolfram, 1998.) The interesting question concerns the source of this market power. Two different sources are usually identified: dominant market positions; and the specifics of the Pool rules.
There is clearly a lot in both views. At privatisation, the generation market was dominated by the two big players, National Power and PowerGen. For most of the period since privatisation the market shares of these companies have been a source of concern. Now overall market shares are appreciably lower, but the issue remains a live one. That is partly because the big generators continue to have large market shares in the sub-market for mid-merit and peaking plant, which generally determines the uniform Pool price (System Marginal Price) that all generation is paid. Whether this sub-market dominance would be of such concern if it did not set the overall market price is less clear. Thus the issue of dominance may well be a product of the Pool rules, and this certainly helps to explain why in this market concerns about dominance arise at levels of market share that in other markets would not attract much interest.
This last point is a rather crucial one, and deserves more discussion. Some commentators have argued that the issue of dominance is the key in respect of generation, and that by focusing on trading arrangements NETA is missing the main target. On this view, it has been the gradual divestment by the large generators that has been key in bringing greater competition to the wholesale market, so that the new trading arrangements are a side-show. I disagree profoundly with this view, because it ignores the important influence that trading arrangements have on the scope for collusion, whether overt or implicit, in price setting. The reason why relatively low market shares in this sector go together with concerns over dominance is that the Pool arrangements have fostered collusion. NETA, by contrast, should discourage it.
Let me develop this point first in the context of the Pool. The essential feature of the Pool is (and I can still use the present tense) that in bidding generators are essentially dealing with a computer programme, GOAL or son of GOAL. Of course, the outcome of their transactions depends on how other bidders interact with the same programme; but nonetheless all transactions are mediated through GOAL. Not surprisingly, smart individuals playing repeatedly can hit on strategies (certain to be mixed) that systematically beat GOAL. The benefits of such strategies accrue to generators in the form either of a higher System Marginal Price or other payments (eg capacity payments), and the costs fall on customers.
Note that in developing this argument I have said nothing about the particular trading rules within the Pool. I could engage in a long digression on the many different and clever ways in which sophisticated bidders can take GOAL and customers to the cleaners, but I do not propose to do so. For my point is much more general: almost irrespective of the form of rules used by the computer programme at the heart of the Pool responsible for scheduling, smart people will learn the optimal way to outwit it. And since all generators are playing with the same programme, they will alight on similar types of strategy for winning. In effect, the programme acts to co-ordinate the different players, encouraging and sustaining a form of parallel behaviour to the detriment of customers. Such parallel behaviour is normally banned under the UK Fair Trading Act as a complex monopoly. It greatly amplifies the impact of strategic behaviour. What is certainly undesirable when carried out singly becomes totally unacceptable when generalised.
Now, of course, I have slightly overstated my case. Those with some knowledge of auction theory may argue that it is possible to devise a set of optimal rules for bidding that elicits the optimal price, and that a Pool that embodies such rules will not suffer from the defect that I have just been discussing. I will come to a key argument against this proposition in a moment. But first note an implication of the current line of argument. Any deviation from the optimal rule is likely to be exploited by bidders to the detriment of customers. In this respect, a compulsory Pool arrangement is unlikely to be robust in its performance.
This is a serious deficiency once one notes the complexity of devising an optimal set of auction rules for electricity. The theoretical models assume, for understandable reasons of theoretical tractability, that electricity is a homogeneous commodity, at least within each period of trade. Yet it is not. Electricity supplied at different points of the grid entails different transmission losses, and more importantly has quite different value because of transmission constraints. Unless the trading period shrinks to an instant, electricity supplied at different moments within the trading period has different value, depending on the demands on the system. Acceptance of generation from one plant for one period of trade may well necessarily entail acceptance of supply for adjacent periods because of inflexibilities in turning generation on and off, and these characteristics differ from plant to plant. I am no auction theorist, but I guess that the resulting interdependencies between auction periods greatly complicate the design of optimal auction rules. Resulting sub-optimalities in design will then give rise to the problems that I have been discussing.
However, there is another line of argument that reinforces these concerns. The UK's leading auction theorist, Paul Klemperer (who after designing the UK's auction of UMTS licences deserves to be one of the richest economists in the country), notes the greater vulnerability of uniform-price auctions to collusive behaviour. (Klemperer 2000a,b. See also Fabra, 2000.) This is especially so in repeated auctions, such as in electricity. The reason for this greater vulnerability is that, with uniform price auctions, bidders can more freely use their bids to signal their intentions to other players. Thus the knowledge that bidders will receive the same price as everyone else irrespective of how they bid facilitates the use of signalling both co-operative behaviour and threats. This makes covert collusion much easier to establish and enforce. In discriminatory price auctions, where bidders get what they bid (pay-as-bid), the use of bids for signalling purposes will carry a much higher cost, because the bid may well be accepted. Pay-as-bid removes the guarantee that all will get the same, and thereby increases the cost of signalling and collusive behaviour. The fact that it does not eliminate the scope for such behaviour is a point to which I return.
Though the formal literature followed, it is this line of reasoning that partly motivated the NETA design. The first key element in the design is to eliminate the compulsion to trade through the Pool. This facilitates bilateral, pay-as-bid, trading of the kind normal in most other markets. The intent is that the bulk of electricity will be traded in this way, probably through the power exchanges that are being established by private sector players. The design of such exchanges is for the market to provide, and is therefore not part of the NETA. But the fact that NETA does not specify these exchanges and Ofgem has played no role in getting them established should not obscure their central importance to the new trading arrangements.
In most other markets (eg financial markets), such bilateral trading suffices. However, the special physical characteristics of an electricity grid necessarily entail a balancing mechanism to ensure the physical integrity of the system. As I have said, the design intent is that this will be a residual balancing mechanism, used for achieving balance close to real time, and that the vast bulk of trading will happen through forward trading. But one way of subverting this intent is by making this market susceptible to manipulation and collusion. The decision, controversial in the view of some commentators, to adopt pay-as-bid in the balancing market was motivated by the advantages discussed above of pay-as-bid in reducing the susceptibility to market manipulation and collusion.
There is another, much more down to earth, argument for pay-as-bid in the balancing market, which concerns the heterogeneity of electricity supply1. It is common to think of electricity as a homogeneous product - after all, all electrons are identical, a key fact that makes possible retail competition in electricity. But in the balancing mechanism, it is not electrons that are traded, but rather bundles of electricity supply, all with very different characteristics. From the perspective of the system operator, there is a world of difference between the plant that needs three and half hours notification to supply, and may then need to be ramped down slowly, and capacity that can be turned on ten minutes before (such as pumped storage or partly loaded plant on spinning reserve). Speed of response and different ramp rates all require different rewards, so that there is no common price that can be applied to them all. This heterogeneity undermines the definition of, let alone the case for, a common system marginal price.
Dual Cash Out Prices
One aspect of the design of the balancing market that provoked much debate has been the adoption of dual cash-out prices, with the price varying depending on whether the bidder is seeking to rectify a short or long position. It has been suggested that this arrangement of dual cash-out prices is arbitrary and penal. The renewables lobby argued strongly, in debate during the passage through Parliament of the Utilities Bill, for netting, so that supply into and demand from the grid are equally priced.
What is proposed is that cash-out prices will be determined by the prices that the system operator has to pay for buying or selling electricity in the balancing market. Thus in the period from gate closure to real time, the system operator will intervene in the balancing market to secure additional generation for the half-hour period in question. The average of the prices that are paid for additional generation will be the price for cashing-out those whose actual supply fall short of contract notification. Similarly the system operator will enter the balancing market to stand-down generation. The average of the prices for these transactions is the price for cashing-out those whose actual supply exceeds contract notification.2
It is fair to say that this particular formula for determining the dual cash-out prices is somewhat arbitrary, and can be criticised and probably improved in the light of experience. However, the principle that there should be a dual price is entirely defensible. Generators whose actual supply differs from that contracted impose additional costs on the system operator who is required to balance the system in real time. It is important that these costs are charged to the generator concerned, to act as a deterrent to such divergences of supply from plan. This will ensure appropriate incentives for ensuring predictability of supply, and will encourage the predictable and reliable plant and deter the unpredictable and unreliable. It may be difficult to judge the actual cost imposed by such unpredictable supply, but to assume that it is zero by adopting a single cash-out price would be even more arbitrary. In all markets, the supply and demand prices diverge by a margin to reflect the costs of market making, and this should be no different in the balancing market. Indeed, the margin is likely to be larger in this market where the costs of balancing are high, as compared with financial markets where the market-making process is typically lower cost. It may be possible to improve on the formula over time to get it to reflect better the true costs that are imposed on the system, and this development should be encouraged. It is not the intention for a mountain of cash to arise from the spread between the two prices (a fear sometimes referred to in the industry as "beer money"); and if there were indications that the spread was too great in practice then the formula should be adjusted. But the principle of the spread is appropriate.
One objection to this line of argument is that the cash-out price penalises divergences between the contracted position and actual metered supply, not the informational deficit, which is the difference between final physical notification and the metered position. If actual metered supply were to match exactly the contracted position, but the generator did not notify, then the cash-out arrangement would impose no penalty. But the generator would then be in breach of the Grid Code, which requires generators to give the system operator accurate notification of their contractual position. Breaches of the Grid Code can attract penalties under the generator's licence. In designing the new trading arrangements, it was decided that a potential double penalty - fines for a licence breach coupled with a cash-out penalty - would be unduly onerous, so the cash-out penalty for the informational deficit was dropped. However, I think this is one of the design features that may be worth reconsidering in the light of experience. Clearly systematic and large-scale discrepancies between the contract position and the notification position can be dealt with via the Grid Code, but it may well be sensible to have a cash-out penalty to discourage both minor and major discrepancies. Since the software functionality for this is in place, this is something that can easily be revisited.
More generally, ensuring that costs fall where they should is a key design feature, ensuring the correct incentives for reliability. This is a feature that has been absent from the current Pool arrangement. Capacity that is unable to perform on the day is not penalised under current arrangements, despite the fact that its unreliability requires the system operator to bring on additional, more expensive plant. Correct incentives should ensure greater reliability, and therefore enhance security of supply. This is a very important feature of the reform that is often ignored, perhaps because of a failure to appreciate the importance having the right incentives in a market-based system.
This argument about predictability and reliability is one that many agree with except when it comes to renewables. In the debates on the Utilities Bill in the Lords, there were persistent calls for netting for renewables, so that supplies to and demand from the Grid would be priced similarly. One can see the attraction of that for renewables, since much existing plant, whether wind, solar or other, is highly variable and unpredictable. Despite my personal interest in renewables, I argued very strongly against netting. If the future is to be renewables, it is very important that we encourage more reliable, more predictable renewables plant. We must establish a premium for reliable renewables, as for all other plant, so that as to give an incentive for innovation that improves on reliability or develops effective methods of storage (such as the advances in regenerative fuel cells). It is therefore important that the dual cash-out arrangement applies to all forms of generation.
To encourage renewables in general is a task that lies beyond NETA, with the climate change levy and other forms of incentives for green energy. (I could digress into the limitations of what is currently proposed, the weaknesses of the climate change levy, and what should be done instead, but that would be to stray too far from tonight's topic.) The right approach is to combine a sharpening of those incentives with the dual cash out system in NETA, so as to promote renewables in general and reliable and predictable renewables in particular.
Ex Ante Versus Ex Post Notification
Another issue that generated a certain amount of heat concerned the question of whether traded positions should be notified in advance, or whether ex post notification should be permitted. I found this a very peculiar debate and I must confess that I found it hard to take it very seriously. If notification happens later, then there is nothing to prevent contracts being settled after the event. Now while there are many examples where contracts are set in advance of exchange and others where contracting and exchange happen at the same moment, I can think of no markets where contracts are settled after exchange. To allow ex post notification would have encouraged people not to contract forward. It would also have allowed the complete circumvention of the dual cash out system, the case for which I have developed earlier.
A much more substantive issue concerns aggregation, where there are serious arguments to be weighed on both sides. Let me put the issue at its simplest and most stark. Imagine a user with variable and unpredictable demand. It will be difficult for that user to contract forward for the unpredictable component of its demand, and she may therefore be forced into the balancing market and thereby suffer the penalty of the dual cash out arrangement. Since she thereby imposes costs on the system operator, that penalty is appropriate.
Suppose now that there is a flexible generator nearby, who can readily match the peaks and troughs in the user's demand. It would be very simple to write a requirements contract, whereby the two agree that the generator will meet the user's fluctuations in demand. The consequence would be that the user places no unpredictable demand on the grid, so that the system operator is unaware and need not be concerned with the user's fluctuations in demand. It would seem inappropriate to impose any penalty on either the user or generator. In effect, they are making life easier for the system operator. The proposed contract enhances economic efficiency.
However, such a requirements contract is not permitted under NETA, which rules out the aggregation implicit in the contract just described, except in the very special case where the generator and user lie behind a common meter, so that the generator becomes, in effect, embedded generation. Ruling out aggregation means that the variable demand of the user will be priced in the balancing market at one cash out price, while the variable supply of the generator will be priced at the other, lower cash price. In effect, the efficient trade is subject to a tax. The trading rules precluding aggregation introduce an inefficiency.
What justification can be given for such an apparently inefficient system? The answer is that we are in the world of second best. (I will return later to the question of how we move closer to first best.) There are three important arguments against allowing aggregation of the type just described: one concerned with market liquidity, the second with market dominance, and the third with transmission constraints.
I see the market liquidity argument as a transitional one. In creating new trading arrangements, it is clearly important that the new markets develop substantial depth, facilitating liquid trade and encouraging the growth of secondary forward and derivative markets. Once they are established, liquidity is unlikely to be a problem. But in the early stages it is probably wise to avoid policies that may drain liquidity from the market. The difficulty with aggregation, if permitted on a large scale, is that it could very easily take a large amount of trade out of markets, inhibiting the development of deep, liquid markets. Vertical integration in the industry adds to this concern, since it facilitates aggregation within a single company.
The second argument concerns the fact that, though there has been considerable divestment, some companies retain appreciable market share. For the reasons that I have argued before, I believe that the new trading arrangements will appreciably reduce concern about these market shares, moving the electricity market closer to a much more normal market structure in it is possible to be relaxed about market shares of around 20% or so. However, as the new trading arrangements bed down, there must be concerns that large players may find opportunities for gaming. In particular, permitting large-scale aggregation would undoubtedly give the larger players greater scope to tie up important parts of the market, and possibly impede the development of a robust and liquid market.
The third argument points to the most obvious type of gaming, namely exploitation of transmission constraints. Let me return to my previous example, and now imagine that a significant transmission constraint lies between user and the generator. Then permitting the requirements contract through aggregation could well result in an outcome that is highly inefficient. The user and generator face a common price; but the system operator may have to enter the balancing market on either side of the transmission mechanism, taking off generation on the generator's side and bringing on generation on the user's side. In this case, the aggregated requirements contract yields no efficiencies, only inefficiencies. Since transmission constraints are commonplace, this represents a powerful argument against aggregation.
These arguments reinforce each other. The scope for gaming is enormously increased by the presence of transmission constraints. Thus a large portfolio generator could enter into a requirements contract across a transmission constraint in the expectation that some of its plant will benefit from the actions of the system operator and bidding in the balancing market accordingly. The extreme would be where the presence of the constraint requires the system operator to take off the generator's plant on one side of the constraint and call another of the generator's plant on the other side. The scope for gaming would therefore be appreciable, and it would be the large players who have the most scope to take advantage of it. Ruling out aggregation reduces the scope for this, though by no means eliminates it.
The other key point is that widespread gaming of this type would inhibit the development of broad and liquid markets and of secondary and derivative markets. That is why it is especially important to inhibit such behaviour in the early development stages of the new markets. I return to this point later when I discuss the market abuse condition. But first let me address one key direction in which I see the new trading arrangements developing, namely towards locational pricing and the explicit pricing of transmission capacity.
I have just set out what I trust is a convincing second best argument for ruling out aggregation. A key part of the argument concerned the limited extent of pricing of transmission capacity and therefore of transmission constraints3. An important question for the future is whether and how we will see the emergence of the separate pricing of transmission capacity, resulting in locational pricing.
I have no doubt that this will be a necessary evolution of the trading arrangements. The move to a more market-based structure will expose the difficulties that the lack of an explicit market in transmission capacity entails. It will also allow the easing of the constraints on aggregation, since a well-designed system for pricing transmission capacity should eliminate the problems discussed earlier. Indeed one possible evolution towards transmission pricing would be to allow greater aggregation behind key transmission constraints, so that different area prices start to emerge from the resulting market trades.
The advantages of pricing transmission capacity are considerable. It would allow the market to reveal the value of additional capacity, thereby providing robust signals for investment in future capacity in the right parts of the network. It would allow the various interconnectors to be treated within the same framework as the rest of the system, facilitating the creation of both a more integrated UK- wide network and integration within the European market. And it could facilitate the development of a more competitive market in the provision of transmission capacity, moving away from the standard view of the grid as something that can be provided only by the monopoly network operator.
I would not be surprised to see such developments starting within a year or two of the launch of NETA - our Chairman will continue to have a very full agenda! But this development needs to be handled with certain care. There are inevitable political sensitivities to the idea that people in different parts of the country will face different prices for electricity, and these sensitivities need to be managed. There is the related transitional problem that, while the pricing of transmission capacity will incentivise the investment to relieve transmission constraints, the presence of acute transmission constraints could easily throw up very big initial price differences that cause acute discomfort to customers and government: an ill-judged and poorly managed move to a secondary market in transmission rights could risk the problems of California, and promote the spread of ochlocratic tendencies from other parts of the energy market to electricity4. Moreover, the technical design of the appropriate market rules for a market in transmission capacity is no simple task, and will need very careful consideration.
Market Abuse Condition
One other area of appreciable controversy, currently before the Competition Commission, has been Ofgem's introduction of the market abuse condition. I think that I have probably said enough about the scope for large players in the market to engage in a variety of forms of gaming to the detriment of the other market participants not to have to dwell on this point at length. The scope for gaming was very considerable in the Pool, and I am confident that it will reduce with the new trading arrangements for reasons that I have discussed. But electricity markets give considerable market power to players, even small ones, close to real time, both because of transmission constraints and because of inflexibility of generation. The need for a fair trading or good behaviour condition is therefore relatively uncontroversial.
What is more controversial is the form that this condition should take. In particular, the new Competition Act gives considerable concurrent powers to the Office of Fair Trading and Ofgem to deal with anti-competitive behaviour. Why are these powers not sufficient?
The key to the application of competition policy is the definition of the appropriate market. From the economist's perspective, the relevant definition of market can be defined in many dimensions, as the Arrow-Debreu framework demonstrates, and in particular is likely to be both time and state contingent. But in practice, the adopted definition of market is driven by legal precedent, especially in the European context. Whereas in electricity, small players at particular times and in particular places may have appreciable market power, competition law may not in practice apply. While one generator may have appreciable market power, not to say monopoly, in the market for flexible plant behind a binding transmission constraint ten minutes ahead of real time for delivery of electricity between at 19.30 and 20.00 on October 10th 2000, it is highly unlikely that the courts would uphold such a narrow definition of market, relevant though it may be from the perspective of the economics of the case. Certainly there are no precedents in European case law for such an interpretation of market.
It is for these reasons that Ofgem decided that it could not rely on the Competition Act and needed the market abuse condition. Of course, an alternative and more heroic course would have been to test the application of the Competition Act by seeking to extend (or more accurately narrow) the received definition of market. But even if that met with initial success, it would have been challenged all the way through the courts, and the issue would have taken years to resolve. The resulting uncertainty could have been damaging for electricity markets, and could have weakened, or even jeopardised, the launch of the new trading arrangements. Not surprisingly and rightly in my view, Ofgem chose to safeguard the new arrangements by introducing the market abuse condition.
There is a further and rather different objection to the market abuse condition; namely that its incorporation into licence conditions allows for no formal appeal mechanism against the judgement of the regulator, and this is against natural justice. Although few are willing overtly to cast our chairman tonight as the enemy of justice, the spectre of a future regulatory ogre is often invoked, with passing reference to one or other current regulator.
It is, of course, true that the regulator cannot establish a formal, de jure, appeal mechanism to his or her judgement concerning the breach of a licence condition, and so can establish no appeal route against judgements made under the market abuse condition. This is because the Director General (and the new authority when it is established shortly) cannot delegate or give away its statutory powers given to him, her or it by Parliament. But Ofgem has made every effort to establish a de facto review procedure that almost certainly achieves the same result as a formal appeals procedure5. Callum McCarthy has established a high-powered Advisory Board, made up of individuals of considerable reputation and independence, to give a judgement on any disputed case resulting from decisions under the market abuse condition. That judgement will be in the public domain. For the regulator or authority to disregard the reasoned judgement of this Advisory Board is very likely to lead to a successful challenge under judicial review. Even it did not, for the regulator or authority to disregard the Board would result almost certainly in a sharp drop in confidence and reputation. In every conceivable situation, the judgement of the Board will hold sway.
So the de facto appeal procedure is as good as a de jure one. That leaves the concern that the authority may, at some point in the future, decide to sweep away the appeal process. I think that is an extremely remote possibility.
The Process of Reform
It will be clear from what I have said that I think that the new electricity trading arrangements represent a major step advance in the direction of a more competitive and efficient electricity market. They put in place a better structure of governance that will allow the arrangements to evolve organically in the light of experience; and the technical design of the new arrangements are, in my view, well thought through. Although they will continue to be improved in the light of experience, even in their initial form they will deliver major benefits. Current market signals give the lie to those, including some of my colleagues at London Business School (Bower and Bunn, 1999), who argued that the reforms would drive prices higher. Forward prices some four or five years out are some 20/25% below levels prior to the reform process. And each whisper of the possibility of delay in introducing NETA pushes forward prices higher.
Let me conclude by reflecting on the process of reform itself, and why it succeeded now and not before. I would identify four factors, three fundamental, and one contingent.
First, there was simply the increasing indefensibility of the old trading arrangements. What was acceptable around privatisation as a way of making privatisation work and obtaining the other benefits of privatisation became increasingly hard to justify with the passage of time. The pressures for reform grew and became increasingly hard to resist.
Second, there was the increasingly diverse nature of the industry, in contrast to the relatively homogeneous nature of the industry immediately after privatisation. This came from changes of industry structure, new entry, and the rise of retail competition. This heterogeneity meant that reform proposals met with a much more differentiated response than before. The opponents of reform were still there, but Ofgem and the Government found many friends of reform arguing for change. The politics of reform in a diverse market were more manageable.
Third, there were the examples of reform elsewhere. Whereas the UK was at the cutting edge of reform at the start of the decade, the electricity industry had fallen behind change in other countries, and in other sectors here in the UK. The growing contrast with the gas industry was increasingly hard to justify.
The contingent factor came from politics. Although I was and continue to be a firm supporter of this New Labour Government, I was surprised how readily ministers embraced the reform of electricity trading, an arcane subject devoid of obvious popular appeal. It is a quirk of history, of the kind that makes politics so fascinating, that the Geoffrey Robinson's misguided attempt to rescue the coal industry and the debacle of the ban on new gas fired power stations created a real determination to push through the reform of electricity trading.
I think that in all of this the industry as whole missed a real chance to initiate reform itself. Mired in the rigidities, politics and sectional interests of the Pool, the industry only saw the need for reform very late in the day and only after the regulator and the Government were determined to act. Those interested in virtual history may speculate what the reformed arrangements might have looked like had the industry been able to take the lead. As it is, we can all watch with keen interest over the next year how the new arrangements are working. I am confident both that they will work well and that they will need to evolve. I am sure that future Beesley lectures will return to this fascinating subject.
1 I owe this point to Julian Bagwell.
2 This is a simplified statement of the rules. Additional rules specify what happens if the system operator does not enter one side of the balancing market in the period from gate closure to real time, as well as an algorithm which removes 'arbitrageable' offers and bids and removes offers and bids that are deemed to be due to transmission constraints.
3 Currently NGC transmission charges reflect overall transmission constraints, so and therefore vary depending upon where generation is located. A key question is how much such charges will affect final consumer bills; transmission is typically 5% to 10% of the total bill, and in general transmission charges will probably not vary enormously, except when the system is under stress.
4 Ochlocracy: government by the mob (OED)
5 This is similar to the arrangement adopted by Don Cruickshank at Oftel when he placed a similar fair trading requirement on BT prior to the introduction of the Competition Act, with a view to more quickly reducing the scope of the price cap imposed on BT.
Armstrong, Mark, Cowan, Simon, and Vickers, John (1994): Regulatory Reform, MIT Press, Cambridge Mass.
Brealey, Richard and Lapuerta, Carlos (1997): "A Report on Generator Market Power in the Electricity Market of England and Wales"; The Brattle Group. Revised 1998.
Bower, John and Bunn, Derek (1999): "A Model-Based Comparison of Pool and Bilateral Market Mechanisms for Electricity Trading"; London Business School, May 1999.
Fabra, Natalia (2000): "Uniform Pricing Facilitates Collusion: The Case of Electricity Markets"; European University Institute, January, mimeo.
Von der Fehr, N. and Harbord, D. (1993): "Spot Market Competition in the UK Electricity Market"; Economic Journal, 103, 531-546.
Harbord, D. and McCoy, C. (2000): "Mis-designing the Electricity Market"; European Competition Law Review, May.
Klemperer, Paul (1999): "Auction Theory: A Guide to the Literature"; Journal of Economic Surveys, 13, 2, 227-286.
Klemperer, Paul (2000a): "Why Every Economist Should Learn Some Auction Theory"; Nuffield College Oxford, mimeo.
Klemperer, Paul (2000b): "What Really Matters in Auction Design"; Nuffield College, Oxford, mimeo.
Newbery, David (1998): "Pool Reform and Competition in Electricity"; in Michael Beesley (ed.), Regulating Utilities: Understanding the Issues, IEA.
Wolak, F. and Patrick, R. (1996): "The Impact of Market Rules and Market Structure on the Price Determination Process in the England and Wales Electricity Market"; POWER Working Paper PWP-047, University of California.
Wolfram, C. (1998): "Strategic Bidding in a Multi-Unit Auction: An Empirical Analysis of Bids to Supply Electricity in England and Wales"; Rand Journal of Economics, 29, 4, 703-725.
Wolfram, C. (1999a): "Measuring Duopoly Power in the British Spot Market"" American Economic Review, 89, September, 805-826.
Wolfram, C (1999b): "Electricity Markets: Should the Rest of the World Adopt the UK Reforms?" Regulation, forthcoming.