Transmission expansion in Argentina 2: The Fourth Line revisited

Available online at www.sciencedirect.com Energy Economics 30 (2008) 1385 – 1419 www.elsevier.com/locate/eneco Transmission expansion in Argentina 2: The Fourth Line revisited Stephen C. Littlechild a,⁎, Carlos J. Skerk b a Judge Business School, Trumpington Street, Cambridge CB2 1AG, UK b Mercados Energéticos, Buenos Aires, Argentina Received 26 February 2005; received in revised form 20 December 2007; accepted 27 December 2007 Available online 9 January 2008 Abstract In 1992, Argentina introduced the Public Contest method, which required transmission users to propose, vote and pay for major expansions of the electricity transmission system. Economists, consultants, the regulator and others held the application of this method to be unsuccessful, mainly on the ground that “for many years” it delayed investment in a “much needed” Fourth Line to Buenos Aires. The commentators blamed externalities, free-rider problems, the Area of Influence method used to define voting rights, and transactions costs. This paper re-examines the history and economics of the Fourth Line. The delay to the Fourth Line was short (a year and a half rather than many years) and enabled a significant reduction in construction costs. Externalities, free-riding, the Area of Influence method and transactions costs were not problematic. The Fourth Line was initially rejected because it was unprofitable to key beneficiaries. Far from being much needed, the Fourth Line was not economic because the congestion it avoided was worth less than the cost of building it. It had become more economic to transport gas to generate electricity near Buenos Aires than to transmit electricity over long distances. The location of subsequent power stations reflected this. Even though the presence of a subsidy meant that the Public Contest method did not prevent the building of the Fourth Line, it nonetheless forced a much-needed reappraisal of traditional transmission investment policy. © 2008 Elsevier B.V. All rights reserved. JEL classification: L33; L51; L94; L98 Keywords: Argentina; Electricity; Transmission; Regulation ⁎ Corresponding author. E-mail addresses: sclittlechild@tanworth.mercianet.co.uk (S.C. Littlechild), cskerk@me-consultadores.com (C.J. Skerk). 0140-9883/$ - see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.eneco.2007.12.007 1386 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 1. Introduction In 1992, Argentina restructured and privatised its electricity sector. As part of the reform, a novel approach called the Public Contest method provided that major transmission expansions were to take place only where users proposed them and a majority voted in favour, confirming that they were prepared to pay. Financing, construction, operation and maintenance of the agreed expansions were normally to be put out to competitive tender. Accounts of electricity reform in Argentina suggest that in general it has been a remarkable success.1 However, regulation of transmission expansion is widely reported to have been deficient or unsuccessful. It is held responsible for preventing or delaying investment needed to meet increasing demand — specifically, the so-called “Fourth Line” from a main generation centre in Comahue to the main load centre in Buenos Aires. The blame is particularly placed on deficiencies of the Area of Influence method used to determine votes under the Public Contest method, inadequate transmission property rights, and conflicting interests coupled with transactions costs of negotiation between the market participants. Views to this effect were expressed by an independent consultant in 1994, the industry regulator ENRE from 1994/5 onwards, a consultancy report to the Ministry of Economics and a World Bank Policy Note in 1996, a consultant's report commissioned by the Secretary of Energy in 1998, and a widely-cited academic economic study in 2001.2 Generally citing one or more of these previous studies, economists and other subsequent authors have taken a similarly critical view about Argentine transmission regulation.3 Only a few commentators defend some aspects of the policy. Some, while critical of the detail of the Area of Influence method, are sympathetic to users determining transmission expansion.4 Others have recently argued that auctioning yields lower tariffs than regulation.5 The success or otherwise of this policy has implications beyond Argentina. Briefly, there has been much debate internationally as to how best to determine and regulate investment in a privately owned electricity transmission system — or, indeed, whether to regulate it at all.6 Because it enables users rather than regulators or transmission companies or merchant investors to choose the investment projects, the Argentine approach to transmission expansion represents a distinctive alternative to the approaches hitherto discussed. The implications also go well beyond electricity transmission systems to network infrastructure in other sectors. 1 For example Bastos and Abdala (1996), Estache and Rodríguez-Pardina (1996), Bouille et al. (2002), Gómez-Ibáñez (2003) and Pollitt (2008-this issue), plus Abdala and Chambouleyron (1999) on transmission, as well as more specialised articles cited herein. 2 Abdala (2008-this issue-a), ENRE Annual Reports 1994/5, 1996, 2002, Spiller and Torres (1996), Estache and Rodríguez-Pardina (1996), NERA (1998), Chisari et al. (2001). 3 For example Newbery (1999), Láutier (2001), Bouille et al. (2002), Woolf (2003), Gómez-Ibáñez (2003), Joskow and Tirole (2005). 4 For example Abdala (1994, 2008-this issue-a), Spiller and Torres (1996), Abdala and Chambouleyron (1999), Abdala and Spiller (2000). 5 Galetovic and Inostroza (2008-this issue). 6 For example Láutier (2000, 2001), Vogelsang (2001) on alternative methods of regulation; Joskow (2006) on incentive regulation; Hogan (1992, 2003), Bushnell and Stoft (1996, 1997) and Chao and Peck (1996) on merchant investment and contract networks; Joskow (2003), Joskow and Tirole (2005) on problems of merchant investment; Littlechild (2003, 2004) on the experience of merchant and regulated transmission in Australia; Rotger and Felder (2001) on a competitive solicitation process; Baldick (2006) on transmission property rights; and an extensive survey by Rossellón (2003). S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 1387 It is therefore of some importance for utility regulation generally to understand the Argentine approach and to assess whether the reported limitations and criticisms of Argentine transmission regulation and the Public Contest method are justified. We have elsewhere examined why the policy was adopted in the first place, and how it has developed since the Fourth Line and what has been the overall impact of the policy on performance. (Littlechild and Skerk, 2004a,b, 2008-this issue-a,b,c,d; also Littlechild and Ponzano, 2008-this issue) The present paper examines actual experience with the Fourth Line, with emphasis on such questions as: – Is there evidence that the policy unnecessarily and uneconomically delayed needed investment in the Fourth Line? – What were the magnitudes of the costs and benefits involved? – Why was the Fourth Line proposed in the first place? – Why did the market participants change their mind with respect to the Fourth Line? – What has happened on that transmission corridor since the Fourth Line was built? 2. Background 2.1. Regulation of expansions of transmission capacity The details of electricity sector privatisation and reform in Argentina, and the provisions for transmission expansion, have been set out elsewhere. (e.g. Littlechild and Skerk, 2004a, 2008-this issue-a) Briefly, an Extra High Voltage (EHV) transmission company called Transener was created out of the incumbent companies and privatised in 1993; five regional sub-transmission companies were created; generation and distribution companies were split up into individual units and many were privatised; and a system operator called CAMMESA and a regulatory body called ENRE were created. The new arrangements were put in place during 1991 to 1993. The Public Contest method was designed to deal with major transmission expansions for public use.7 In order to request an expansion of transmission capacity by Public Contest, the proponents notify the transmission company that holds the concession in the area of the expansion. Under the original scheme (which was shortly modified), the proponents were required to accompany the request by an initial bid for the work by a new or incumbent transmission company. The Dispatch Organisation (part of CAMMESA) carries out a technical study, using the so-called Area of Influence method, to identify the “beneficiaries” of the expansion and the proportion in which each beneficiary would have to share the costs of amortisation. The transmission company reports on the technical feasibility of the request. ENRE may consider a request only where the proponents represent at least 30% of the beneficiaries that the expansion would bring in its “Area of Influence”. ENRE then has to arrange for a public hearing. In the event of opposition by 30% or more of the beneficiaries of the expansion, ENRE must reject the expansion request. If there is no opposition, or not sufficient to warrant further investigation, ENRE must approve the request, and issue a Certificate of Convenience and Public Necessity. The proponents then arrange for a public tender to construct, operate and maintain the proposed expansion. Subject to some qualifications, the tender goes to the lowest bidder (or, under the original scheme, to the initial bidder if no bids were below that level). 7 There were other methods for less major transmission expansions, namely Contract Between Parties (expansions for one or a few users, such as connections), Minor Expansions (under $2m) and expansions for private use (Section 31). 1388 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 The method used to determine beneficiaries and the Area of Influence is evidently important and has attracted considerable attention. CAMMESA determines who are the beneficiaries of a new line, and its Area of Influence, by using a simulation model based on the model it uses to schedule plant and set nodal prices. A generator is a beneficiary if an increase in its output (with a corresponding increase in consumption at the ‘swing bus’ in Buenos Aires) would increase the flow along the new line. A distribution company or large user is a beneficiary if an increase in its consumption (with a corresponding reduction in consumption in Buenos Aires) would increase the flow along the new line. Both are simulated in normal conditions of operation. The Area of Influence of the new line is the set of these beneficiaries. CAMMESA then uses a more elaborate series of simulations to calculate the ‘participation’ of each beneficiary in the expansion. The voting share of each beneficiary is the weighted average of its expected participation over the first two years of the line's operation. Transmission expansions via the Public Contest method are financed by all those parties who are identified as beneficiaries in the Area of Influence of the expansion, in proportion to their shares as beneficiaries. This calculation is updated monthly during the amortisation period of the COM contract, so that actual users pay for the expansion, not only the users who voted for it. After the expiration of this amortisation period, the annual remuneration for operation follows the remuneration regime applicable to existing installations of the incumbent transmission company, which essentially covers operating and maintenance costs only. 2.2. Nodal pricing, congestion and the Salex Funds It was expected that generation and distribution companies and directly connected large users would propose expansions. They would have an interest in doing so because their revenues and costs depended on the level of congestion and quality of service in the network. This was because Argentina had adopted a nodal pricing scheme (together with obligations to supply on the distribution companies and penalties for failure to do so). In simple terms, price at Buenos Aires is set equal to marginal system cost. When the system is uncongested – that is, when there are no transmission constraints – the price at any other consumption node is equal to the price in Buenos Aires plus marginal transmission cost (including marginal transmission losses) from Buenos Aires. The price at any generation node is equal to the price in Buenos Aires less marginal transmission cost from that node to Buenos Aires. Where a line is congested – that is, there is a transmission constraint because the economic power flow from a generation node would exceed the transmission line's capacity – there are two consequences. First, the price in Buenos Aires increases because higher cost generation has to be secured instead and marginal system cost is now higher. Second, a ‘local price’ applies in the constrained generation node. This is set equal to the marginal cost of generation at that node, at the level of output equal to transmission capacity. This local price will be less than the price in Buenos Aires by more than the marginal transmission cost. If the transmission constraint is severe, so that a great deal of generation is precluded from being transmitted to Buenos Aires, then the difference in prices could be considerable. The local price could even be zero if enough generators were declaring zero marginal costs (e.g. as a result of hydro generators spilling water, as indeed happened in Comahue at one point). The nodal pricing scheme thus meant that, if the transmission lines were congested, generators lost revenue and distribution companies in Buenos Aires and other consumption nodes had to pay more for their supplies (and had a greater risk of non-supply). Both parties therefore had an interest in paying for new transmission lines to avoid or reduce congestion. Generators had a greater interest to the extent that the impact on them was typically heavier. (There was also some variation in how the provincial authorities regulated the distribution companies.) S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 1389 When congestion occurred and local prices applied, the difference between the prices paid by the distribution companies (and their consumers) and the revenues received by the generators were paid into so-called Apartamientos Accounts, along with the differences between estimated and actual transmission losses. These Accounts were initially used to stabilise the revenues of the transmission companies and any surplus was used to reduce other components of transmission charges. As congestion began to increase in the transmission corridor between Buenos Aires and the generation area of Comahue, the balances in the Accounts accumulated to such an extent that it seemed more sensible to make other provision for using them. (Littlechild and Skerk, 2004a, 2008-this issue-b). Rather than use the revenue for non-transmission purposes, in August 1994 the Secretary of Energy specified that, in future, revenues from nodal price differentials deriving from congestion (i.e. local prices) should be put into so-called Salex Funds, one for each of seven transmission corridors. These Salex Funds could be used to reduce the cost of transmission expansions using the Public Contest method. To be eligible for support, the expansions had to produce reductions in the transmission constraints that generated local prices in the corresponding corridor. This would also facilitate a competitive wholesale electricity market.8 Fig. 1 shows the build-up of the Salex Fund for the Comahue corridor. Its potential contribution soon became an issue in the Fourth Line debate. 2.3. The Fourth Line By late 1994, the government as part-owner of the generation station Yacyretá had successfully put out to competitive tender three major high-voltage lines totalling 853 km, and several connections totalling 60 km had been made to the high-voltage system by agreement. The arrangements for Argentine transmission expansion seemed to be working. On 2 September 1994 two generators – the owners of the hydro plants El Chocón and Alicurá in Comahue – applied for an expansion of transmission capacity by Public Contest.9 They wished to construct a new 500 kV line, of length nearly 1300 km, from the Piedra del Águila hydro plant in Comahue to the edge of Buenos Aires. This was the famous Fourth Line, so-named because there were already three lines on that route (see Fig. 2). They offered a Construct, Operate and Maintain (COM) contract with an annual fee of $54.6m for the first three and a half years and $61.4m for the remainder of the 15 year period, to be provided by a proposed independent transmission operator called Tenasa.10 The two proponents represented just over 30% of the beneficiaries of the line.11 The public hearing was held on 17 February 1995. Two other hydro generators from Comahue (the owners of Piedra del Águila and Cerros Colorados) and a thermal generator (Central Térmica Alto Valle, in the same ownership as Cerros Colorados) opposed the project. Since these three generators represented 34% of the votes this sufficed to veto the project. Two other generators later joined them, bringing the opposition votes to over 50%. ENRE formally rejected the application on 28 March 1995. This vote was a surprise and disappointment to many parties, not least the regulator ENRE, which referred to the Fourth Line as “a work planned by the public undertaking Hidronor in the 8 “The practical effect of doing this is to give adequate direction to the funds originating from local pricing towards the expansion of transmission capacity. It is necessary to remove constraints on the free dispatch of generation, and necessary to give precision to the use of the funds derived from local prices that remain in the Apartamientos Accounts.” Resolution 274/1994. 9 Transcript of public hearing 17 February 1995, attached to ENRE Resolution 49/1995 of 28 March 1995. See also Galetovic and Inostroza (2008-this issue). 10 The Argentine peso and the US dollar are both denoted $. During the period covered by this paper, they exchanged at parity. 11 Later, on 24 November, they were joined by Turbine Power Co, which had about 2% of the votes. 1390 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 Fig. 1. Apartamientos Account and Salex Fund 1993–5. Source: Authors' calculations based on CAMMESA data. 1980s and ever since those days considered necessary by the industry”.12 With increasing demand in Buenos Aires and increasing generation in Comahue, and with increasing signs of congestion in that transmission corridor, there had been a long-standing expectation that a fourth line would and should be built after privatisation. Concerns were widely expressed. In 1996 the Fourth Line was proposed again, and this time it passed. It went into operation in 1999. It was a particularly large and important investment. It had about 2600 pylons, and employed over 3000 workers at its peak. It was 1300 km long, and increased the line length in the whole 500 kV transmission system by one fifth. It increased power transfer capacity on the Comahue – Buenos Aires link by nearly two fifths.13 2.4. Criticisms of the Public Contest and Area of Influence methods ENRE was disappointed by the initial rejection of the Fourth Line, and indeed concerned about the Public Contest method itself. It urged on beneficiaries the benefits of reinforcing weak or congested lines.14 It also noted the importance of correctly identifying the beneficiaries of an expansion, a remark that some interpreted as suggesting the need for change in the mechanism. After the Fourth Line had been approved, ENRE expanded on its concern about inadequacies in 12 ENRE Annual Report 2002, p. 49. Sources: Galetovic and Inostroza (2008-this issue); Argentina in the Third Millenium, Universidad Argentina de la Empresa (UADE), Julio Moyano Comunicaciones S A, Buenos Aires, 2000, p. 326. 14 It suggested that beneficiaries should address themselves to the advantage of any proposal, quantify its costs of implementation and accept their participation in paying for it. ENRE Annual Report 1994/5, ch. 5, esp p. 59. 13 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 Fig. 2. Map of Argentine transmission system. Source: Mercados Energéticos. 1391 1392 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 the Rules, not least the Area of Influence method. It explained why these numerous difficulties had obliged it to intervene repeatedly, albeit successfully.15 ENRE was not the only critic. An early economic appraisal of the transmission expansion method, predating the Fourth Line proposal, found the approach deficient.16 A report for the Ministry of Economics is cited as identifying a variety of concerns about institutional features that could delay investment.17 A World Bank Policy Note suggested that the Area of Influence method excluded the demand side and was likely to result in suboptimal decisions.18 ENRE commissioned academic research into incentives for transmission investment. The resulting study concluded “While the transformation of the Argentine electricity market in the last decade has positively affected generation and distribution performance, most agree that current regulation has failed to spur needed investments in high-tension transmission. The lack or delay of such investments arises from problems in the willingness-to-pay revelation under the Public Contest mechanism.”19 In 1999, consultants NERA, commissioned by the Secretary of Energy to review the entire industry, concluded “… the expansion of transmission has been a major problem in Argentina. The development of the fourth Comahue line was delayed for several years.”20 NERA diagnosed the main transmission problems as the absence of transmission rights and the use of the Area of Influence method. Even those critics sympathetic to the idea of users determining transmission investment identified the design of the payment mechanism as a limitation in the particular application of the Public Contest concept.21 Most subsequent authors based their critical conclusions on one or more of the above analyses. They focus in particular on the initial rejection of the Fourth Line.22 15 ENRE said that, in spite of these obstacles, it had been possible to create a strongly competitive atmosphere that had led to a fee of $24.5 millions against a maximum fee, fixed by the interested parties, of $43.67 millions. “If nothing else, this alone justified the active and firm participation of the Regulator in the process.” ENRE Annual Report 1996. 16 “The mechanism for capacity expansion misallocates the costs of financing new investment, as it implicitly ignores basic welfare effects, especially on the demand side. This, in combination with the ad hoc public hearing procedure adopted, will result in deviations from the optimal investment path in transmission.” Abdala (1994), p. 12. See also Abdala (2008-this issue-a at V.B). 17 “Much needed investments (construction of a fourth transmission line linking the main generation center to the main load center, the city of Buenos Aires) have been retarded by many years. Informed commentators (e.g. Spiller and Torres 1996) have attributed this delay to institutional features: difficulty in coordinations, free-rider problems, inappropriate measure of benefits, etc. — and have suggested remedial measures to foster transmission investments by groups of users.” Láutier (2001), p. 45. 18 Estache and Rodríguez-Pardina (1996), p. 4. The second author is a former senior economist and later adviser at ENRE. 19 Chisari et al. (2001), p. 713. 20 NERA (1998), p. 53. The report claimed that a veto provision in the Public Contest method “was used in the Comahue case to block development for four years.” This figure has been repeated elsewhere. 21 “Prolonged congestion in power transmission in Argentina indicates that the BOM and private contract procedures can lead to nonoptimal investment: … the BOM procedure … has conceptual flaws, and the veto safeguards are insufficient to prevent unfair and inefficient outcomes.” Abdala and Chambouleyron (1999), p. 3. “The rule [for allocating costs] is flawed as cost allocation is based on an elementary measure of power flows. Hence it does not take into account externalities or users preferences, and produces unfair results that distributors are not willing to accept. Veto safeguards to protect those receiving negative externalities or those who had to bear a share of investment costs larger than their willingness to pay are insufficient, and there are no provisions for compensation mechanisms.” Abdala and Spiller (2000), para 3.1.1. See also Abdala (1994, 2008-this issue-a,b), Spiller and Torres (1996). 22 “Evidence points to a failure to correct regulatory obstacles that prevented expansion of the transmission system.” Bouille et al. (2002) pp. 31–2, 46 citing Abdala and Chambouleyron (1999). “Some of the coordination methods developed in Argentina did not work well, particularly voting by beneficiaries.” Gómez-Ibáñez (2003), p. 324, citing and NERA (1998) and Chisari et al. (2001). “There is widespread acknowledgement of the shortcomings of the scheme” Woolf (2003), p. 272, also citing NERA (1998) and Chisari et al. (2001). “The Argentine system implemented an untried model of transmission expansion, which proved controversial” Pollitt (2008-this issue), p. 25 [check] citing most of these articles. See also Newbery (1999) p. 254 for concerns about transactions costs and quality of supply. S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 1393 The Fourth Line experience evidently continued to be a major regulatory concern. ENRE found it appropriate to open the transmission chapter of its annual report 2002 with some critical remarks about the inadequate incentive mechanism and opportunistic behaviour having delayed or impeded important expansion projects.23 Within the industry, Transener expressed concern about the apparent lack of coordinated planning.24 Other concerns believed to be shared by transmission companies included overloading of the transmission lines with adverse implications for reliability, capricious and unfair penalties for outages, inadequate incentives on distribution companies to participate in the transmission expansion process, inadequate arrangements as between Transener and CAMMESA, and undue restrictions on transmission companies.25 2.5. Support for the Public Contest approach There appear to be few public supporters of the Public Contest approach. Nonetheless, generators liked the control over their costs that the method provided, and the relative lack of dependence on uncertain government actions.26 A US executive with one of the generating companies says that the dependence of all market participants on the performance of the transmission system, and their responsibility for expansions, provided an incentive for them to discover new opportunities for improving this performance. This in turn stimulated improved information about the transmission system.27 Some saw the provision for putting expansions out to tender as advantageous in terms of flexibility and financing. The costs of constructing a project could be spread over anything from one to 15 years, depending on the convenience to the users and potential bidders.28 This also had implications for financing. The familiar method of price caps revised at five year intervals (for example as applied to Transener's existing system) implied uncertainty about the conditions that would follow the end of the cap. In contrast, the ability to tender for a project with a fixed 23 “… The signalling mechanism directed at incentivising those interested in expansions seems not to have completely met the demand and, furthermore, is not immune to certain opportunistic behaviours (of the free-riding type) by the agents. In the past these limitations have delayed and/or impeded the realisation of expansion projects of an important magnitude.” ENRE Annual Report 2002, p. 49. The comment then gives the Fourth Line as an example. 24 “In our brief experience of 30 months operating the system, we believe that a certain degree of planning for the future is essential so that there is no lack of coordination in the expansions and thus over-investment.” Statement of Jose Luis Antúnez of Transener, quoted in ENRE, International Seminar on Restructuring and Regulation of the Electric Power Sector. Buenos Aires, November 1995, p. 64. 25 Woolf (2003), pp. 262–276. She also expresses the view that “a backstop is needed in case the market fails to propose some expansion projects that are clearly needed, so that they can be implemented on a regulated basis, while still preserving the benefits of the competitive bidding process that has done so well to bring construction costs down.” (p. 276). 26 Mr E Badaracco, chairman Endesa and chairman of Association of Generators, personal communication, 1 December 2003. 27 “Capacity prices in the outlying regions were penalized if the connections to the market were not reliable, thereby adding a price signal to encourage participants to improve system reliability. Consultants would crawl the system looking for places to install things that would improve stability and eliminate constraints, or that would improve the unreliable links. You end up with a lot of people knowing quite a lot about the transmission system.” J D Roark (former Power Market Analyst, Southern Electric International), personal communication to Prof W Hogan. 28 The amortization period has indeed varied from 1 to 15 years. Not surprisingly, the period is generally longer for larger investments (say over $50m) although the periods for investments below this value show more dispersion. 1394 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 remuneration over the whole of its specified financing period facilitated the financing of large projects (perhaps similar to Public Finance Initiative projects in the UK).29 Galetovic and Inostroza (2008-this issue) have recently argued that the Argentine approach of putting the construction of transmission lines out to tender yields lower tariffs than conventional regulation would provide. In particular, the fee for the Fourth Line would have been 61% higher if the traditional regulatory approach had been adopted. As noted, some who identify problems with the specific implementation of the Public Contest approach nonetheless see coalitions of users – for example, via regional boards – as being part of the solution. (Abdala, 2008-this issue) And the Area of Influence method has been extended rather than replaced when it has been applied in Buenos Aires Province. Littlechild and Ponzano (2008-this issue). But supporters of the Public Contest approach have hitherto been very much in the minority. 3. Initial evaluation 3.1. The Fourth Line process and alleged delay Did the Public Contest method unduly delay a much-needed fourth line? Contrary to the alleged lack of coordination, opportunistic behaviour and transactions costs, the proponents and initial opponents of the project worked closely and effectively together after the first proposal was rejected. Eight generators formed the Group of Electrical Generators from Comahue Area (GEEAC). They secured two modifications of the Public Contest procedure (see below) and achieved consensus on the expansion. On 7 May 1996 the eight generators made a new application to construct the same line. This time the proponents accounted for 82% of the votes, so there was no possibility of 30% objecting. On 25 September 1996 ENRE held a second public hearing. Only two beneficiaries now opposed the project, accounting together for 9.5% of the votes.30 ENRE approved the requested transmission expansion and issued the Certificate of Convenience and Public Necessity on 24 October 1996,31 issued a call for tenders on 22 May 1997 (to be submitted on 27 October) and announced the winning bidder on 12 November 1997. Evidently the alleged delay was not “many years” or “four years”. It was just over a year and a half.32 In the context of transmission planning generally, a year and a half is not unduly long. One author comments that “The time taken to obtain the necessary siting, planning and environmental consents is inherently long. It is seldom shorter than two years and can take as long as 10 years for 29 “I have always admired the transmission enhancement feature of the Argentine market. It needs financial rights to make it complete, but it works as it is. Though it is facilitated by the relatively simple spider-radial nature of the Argentine system, there are some very important features of this procedure that modern-day proposals lack. In particular, when a line is accepted as a legitimate system procurement by CAMMESA and by (at least 70% of) the beneficiaries, it takes on an official stature. It will have the same revenue-collection status as any regulated line; its costs will be billed out over time, and they will be collected under the existing transmission tariff. The credit of the market stands behind the project, and this makes the project financeable. … In short, for me it stands out as a better thought-out idea than most of the modern day proposals.” J D Roark, personal communication, 23 May 2003. 30 The electricity company ESEBA owned by the province of Buenos Aires had a generating station located midway between Comahue and Buenos Aires, which would be subject to greater competition if the Fourth Line went ahead. ESEBA's distribution business (3.98%) voted with its generation business (5.51%) to oppose the expansion: although the two businesses were identified as separate beneficiaries, they acted as a single integrated company. These and most other beneficiaries had not recorded their votes on the first proposal since ENRE cancelled the proposal in the light of the opposition of more than 30% of the beneficiaries. 31 Resolution ENRE 0613/1996, 24 October 1996. For earlier stages see also Resolutions ENRE 0441/1996 and 0525/1996. 32 One year 8 months from first to second application, one year 7 months from first to second public hearing. S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 1395 a major project.”33 This is the case in the UK, for example.34 Other studies have made similar assumptions.35 Moreover, at least part of this delay – certainly before and after the second hearing – had nothing to do with the Public Contest approach or the controversial Area of Influence method or any lack of coordination between the beneficiaries. Rather, it had to do with the concept of introducing competition and new entry into the transmission sector. There were substantial transitional problems and uncertainties here, particularly concerning the role of the incumbent Transener.36 These problems took time to resolve.37 Indeed, the delay may have been necessary and beneficial in order to iron out these problems.38 3.2. Was the Fourth Line economic? Evidence from contemporary modelling Concern about delay to a ‘much needed investment’ presumes that the investment was in fact economic. However, evidence – or the lack of it – from contemporary modelling suggests legitimate room for doubt here. 1) The generators each did their own modelling, and the models gave different results. Modelling calculations at the time were very sensitive to assumptions made, not least with respect to future hydrological conditions and the appropriate discount rate or cost of capital in those early uncertain days. The representative of those Comahue generators initially voting against the project explained in some detail at the first hearing that they had no objections to expanding capacity, but their own modelling calculations, made with the help of international consultants from Madrid, showed that it was simply not a profitable investment for the generators as a whole, and that it had a negative net present value even for the generators who proposed it.39 33 Woolf (2003), p. 518. “The period in Norway is of the order of 7 years.” (fn 1 p. 581) “The process for a relatively short line outside Washington DC started in 1976 and was completed in 1992.” (p. 18). 34 Assuming no public enquiry is necessary it typically takes from one and a half to two and a half years to progress from beginning to identify system need and technical options to securing central and local consents. Exceptionally it can take much longer: the 75 km Second Yorkshire Line over the North York Moors took ten years to complete this process, plus another three and a half years to build. (National Grid Company, personal communication). 35 For example “Transmission lines do not take very long to build once they have obtained siting permits. However, for major new transmission corridors, the permitting process can be very lengthy.” Joskow and Tirole (2005) p. 56, fn. 30. They assume in one analysis that acquisition of transmission siting permits plus actual construction will take ten years. They also say that transmission investments “are particularly vulnerable to pre-emption strategies due to their long lead times”. (p. 57 fn. 31). 36 On the struggles between competing construction bidders, the resistance and conflicting roles of the incumbent Transener, and the important contribution of ENRE in enabling competition to take place, see Galetovic and Inostroza (2008-this issue). 37 For example, initial negotiations between GEEAC and Transener on the technical aspects of the project took about three months to resolve, before ENRE was able to call a public hearing. (Galetovic and Inostroza, 2008-this issue) Presumably the failure to agree these aspects before the first public hearing was the cause of this part of the delay. See Galetovic and Inostroza (2008-this issue) for further examples, ENRE Annual Report 1996 just cited, and various remarks in the ENRE resolutions on this case. 38 “The delay once the project had been approved was not the result of intervention by the beneficiaries, but a reflection of a struggle between transmission firms. Nonetheless, this delay made it possible to fine-tune the regulation, by satisfactorily solving the conflicts of interest raised by Transener participation. The final outcome of the auction shows that this delay was the price paid for refining the regulatory mechanism, and therefore should not be repeated in the future.” Galetovic and Inostroza (2008-this issue), p. [22]. 39 Transcript of public hearing 17 February 1995, attached to ENRE Resolution 0049/1995 of 28 March 1995, testimony of Mr Turri representing Piedra del Águila, Cerros Colorados and Central Térmica Alto Valle. A representative of another generating company has also confirmed to us that the calculations were very sensitive to assumptions made, particularly growth of demand, and in some cases the net benefit was negative. 1396 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 2) When the project was brought forward in 1995, did ENRE actually “determine that the benefits in lower electricity prices made the line in the public interest”?40 If ENRE ever took a formal view on this question, it does not seem to have published it. Perhaps, since the 1995 proposal was voted down, ENRE did not need to calculate whether it passed the Golden Rule.41 Whether ENRE actually made or saw any unpublished calculations at that time, and if so what they showed, is unknown. 3) ENRE did not in fact have the resources to do its own modelling of the net present value calculation. In 1996, when participants voted in favour of the Fourth Line and ENRE did need to do the calculation, it contracted out the modelling to a group at the University of San Juan. The University did not have a technical model designed for this purpose, and some questioned the adequacy of the resulting calculations. For example, it has been said that it included the increased consumer surplus from lower generation prices in Buenos Aires, but not the consequent lower producer surplus for the generators. ENRE did not circulate the study widely, and we have not been able to access and assess it.42 4) In 1996 senior staff members at CAMMESA made unofficial and unpublished calculations related to the new line.43 They used CAMMESA's own model, which was the most informed and authoritative then available, based on actual reported costs and up to date system data. They found that whether the net benefits were positive or negative was very sensitive to the assumptions made. The case for the expansion was at best borderline. 5) Several crucial factors changed from the time of the first hearing to the second, including the demand in Buenos Aires, the generation capacity and output in Comahue, and the load factors on the existing transmission lines. All these factors were greater in the year preceding the second hearing in September 1996 than in the year preceding the first hearing in February 1995. Hence the economic case for construction in late 1996 was stronger than it would have been in early 1995. If the case was borderline in 1996, and the net benefit in many scenarios negative, this casts even more doubt on the viability of the case in 1995. 6) In these circumstances, a deferral could have increased the net value of the investment, even at a given construction cost. In more formal economic parlance, the inclusion of an option value for waiting might well have shown an advantage in deferring the investment from 1995 to 1996.44 The evidence – or the lack of it – from these various modelling sources does not provide tangible reasons to assume that the Fourth Line was indeed economic from an aggregate (or 40 Gómez-Ibáñez (2003), p. 314. In order to apply the Golden Rule, ENRE has to check that the present value of the total costs of investment, operation and maintenance of the Electricity System as a result of the proposed expansion would be less than it would be without such expansion, where the costs of operation include the value of energy not supplied to the market. In making this evaluation the cost of investment, operation and maintenance of the expansion is taken to be as specified in the proposal. 42 Whether ENRE could realistically have published a study showing that the proposal did not pass the Golden Rule is an interesting question. For various reasons, there was widespread support for the line. (Government and politicians wished to respond to pressures from provinces, Comahue generators wished to avoid lower revenues caused by congestion, Transener and potential construction companies wished to increase their business, and distribution companies wished to improve their quality of service.) It would have been a bold regulatory agency that vetoed a decision by market participants to build and pay for the line. 43 These were unofficial and unpublished calculations. CAMMESA was required to calculate the participation percentages using its own model, but it was the role of ENRE, and not CAMMESA, to apply the Golden Rule and evaluate the economic case for a transmission investment. 44 We are indebted to Omar Chisari for this comment. The borderline nature of all these calculations also suggests, incidentally, the weakness of the case for building the Fourth Line a decade earlier, even though it was reportedly “considered necessary by the industry” ever since the 1980s. 41 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 1397 social) perspective when it was proposed in 1995. In fact, it casts considerable doubt on this assumption. 3.3. The study by Chisari, Dal-Bó and Romero (2001) Several commentators base their criticisms of Argentine transmission regulation on the calculations by Chisari et al. (2001), implying that this study has established that inadequacies of the Public Contest mechanism did indeed delay the economically beneficial Fourth Line project. This is an important and influential study, which certainly explains why the Public Contest mechanism might delay an economic project. But does it establish that deficiencies in that mechanism actually did delay the Fourth Line, and does it show that the Fourth Line actually was economic?45 The authors were commissioned by ENRE in 1996/97 to examine and explain why there was not more investment in transmission under the new arrangements. To this end, they built a simulation model of the national electricity system, and used it to analyse the voting behaviour of the market participants under the Public Contest mechanism. They used several examples from this model to identify flaws in that mechanism. In summary, these flaws were the exclusion of consumers from the mechanism, the exclusion of market participants in the ‘swing bus’, the assignment of votes and fees based on usage rather than profit, and the possibility of strategic vetoes on expansion. Their comment cited above, that “most agree that current regulation has failed to spur needed investments in high-tension transmission”, is to be understood as a description of the prevalent view that they were asked to explore, not as a summary of their own conclusions. However, the authors did conclude that “The lack or delay of such investments arises from problems in the willingness-to-pay revelation under the Public Contest mechanism.” This paper is an innovative, careful and valuable piece of research. The authors were among the first to analyse publicly, and from an economic perspective with detailed and realistic calculations, how the Public Contest mechanism might work.46 Their examples showed, among other things, that votes based on usage could be different from votes based on benefit. They suggested that there were circumstances such that “under the Public Contest mechanism, desirable expansions may not be constructed while undesirable ones may”.47 It was not simply, as some conjectured, that there was necessarily a divergence between private and social benefit. The specific design and application of the mechanism could distort the outcome in a way that was not previously appreciated or at least not fully understood. The simulations also revealed additional competition policy problems associated with joint ownership of generation and distribution companies in different areas. The authors calibrated their model against available data from the Argentine electricity system around the year 1997, and took the proposed Fourth Line as an example. This increased the potential relevance of the results, and raised important questions. However, for several reasons the results need to be treated with care. First, the authors themselves emphasised the limitations of their modelling. They also noted that their objective was not to appraise the actual situation, but to demonstrate the possibility of 45 This section draws on Chisari et al. (2001) and also on helpful clarifications by Omar Chisari and colleagues. Before that, CAMMESA had made some unpublished studies at the time of drawing up the rules for the mechanism. Abdala (1994, 2008-this issue-a) had pointed out, among other things, the importance of considering the demand-side in the cost allocation rule, and also proposed examples of rules based on welfare analysis, rather than on electricity flows, but these were more rudimentary and hypothetical examples. 47 Chisari et al. (2001), p. 714. 46 1398 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 situations where the results of the process were not optimal. These situations were “not necessarily realistic for the Argentina case”.48 Second, the limited representation of investment options in the model may have distorted the economic solution and the views attributed to beneficiaries and overemphasised the importance of the exclusion of users in the ‘swing bus’. For example, consumers and distribution companies in Buenos Aires were assumed to benefit from a new transmission line because it would reduce the price of electricity in that city, so it was assumed they would have voted for it if they had the chance. However, this implicitly assumed that the extent and location of generation capacity was given, and that the choice was between a new line and no line. In fact, however, an alternative way of meeting increasing demand in Buenos Aires was to transport gas there from Comahue, and to increase generation capacity in Buenos Aires. If this were more economic than transmitting electricity, prices in Buenos Aires would decrease even more if the new line were not built, so users in that city would be better advised to vote against the line, not for it. In other words, if the possibility that the Fourth Line was not the most economic investment had been included in the model, then the exclusion of users in the ‘swing bus’ might not have been critical at all.49 Third, any claim that deficiencies in the Public Contest mechanism were responsible for the failure of the 1995 vote implicitly rests on the assumption that the Fourth Line was economic, and would have attracted sufficient support in the absence of those deficiencies. However, the authors did not show – nor did they claim to show – that expanding the transmission system by the Fourth Line would have been economic in 1995. Since the foregoing discussion has cast doubt on this assumption, it is worth looking at what the numbers in this study imply about the economic nature of the expansion. 3.4. A first calculation Under the first proposal in September 1994, the proponents offered a Construct, Operate and Maintain (COM) contract with an annual fee of $54.6m for the first three and a half years and $61.4m for the remainder of the 15 year period. Under the second proposal in May 1996, the proponents specified a maximum annual fee of $43.67m (which assumed a contribution of $80m from the Salex Fund). The winning bid in November 1997 was an annual fee of $24.52m plus the $80m Salex contribution. Under very simplified assumptions Chisari et al. calculate that the benefits associated with the Fourth Line would have a present value of about $112m.50 This was calculated as the present value of an annual cost reduction of $6.1m, summed over 50 years at 5% discount rate. The authors continue “In this context it would be optimal to carry out the investment if the costs were below 112 million pesos and reject the proposal if its cost were above that figure.”51 They discuss the implications of the cost being above or below $112m. But they do not say what the cost of the investment actually was, so they do not reach a conclusion on whether the investment was actually economic. 48 The results here presented do not reflect in a precise way reality, or the results of the original model, and constitute only examples.” “A rigorous study of the social benefits of an extension of the lines between Comahue and Buenos Aires is not the aim here, instead, the objective is to present some situations, not necessarily realistic for the Argentina case, in which the decisions arising from the voting process are not optimal solutions.” Chisari et al. (2001), fn. 9 p. 704 and p. 709. 49 Chisari and Romero (2008-this issue) have now established this point quantitatively. 50 “Let's suppose, in an unrealistic way, that the electric system maintained the structure, costs and levels of demand of the second year until the end of time. If we assume the extension will work properly for 50 years and the discount rate is of 5%, then the social benefits of the investments in the line Comahue–Buenos Aires would be close to 112 million pesos.” Chisari et al. (2001), p. 709. 51 Chisari et al. (2001), p. 710. S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 1399 Fig. 3. Evaluation of benefit of additional transmission capacity. According to ENRE the eventual cost of the Fourth Line as a result of the competitive tender was about $250m. This is more than double the critical figure of $112m, above which the authors suggest it would be optimal to reject the proposal.52 The cost of the original proposed construction was of course much greater than $250m. It has been calculated at $370m.53 Similarly, the annual fee proposed in 1995 was $54.6m increasing to $61.4m. This is an order of magnitude greater than the annual benefit of $6.1m assumed in the study by Chisari et al. Thus, to the extent that weight can be placed on the assumptions in that study, they suggest that the Fourth Line expansion was not economic, certainly not at the time it was first proposed nor even later. 4. Cost–benefit analysis 4.1. A simple framework It is possible to make an alternative estimate of the benefit of the Fourth Line in terms of reducing congestion, using essentially the same analysis and diagram as in NERA (1998).54 Let Sm denote the supply curve from generators in Comahue and Sr the supply curve from the rest of the generators in the system. Assume the total demand is given (i.e. inelastic), and equal to Q, represented by the distance OQ between the two vertical axes in Fig. 3. In the absence of transmission constraints, and assuming cost-related pricing and no transmission losses (or assume the supply curves are net of transmission losses), then generation will be allocated between the two sets of generators so as to minimise total cost. This means that Comahue generators will produce quantity Qu (where u indicates unconstrained), and other generators will produce the remaining output (Q − Qu). There will be a uniform system price Pu (not shown on the figure) equal to generation cost at the margin, which will be the same in both sectors of the market. 52 This understates the differential because of the different time periods and discount rates used. The Fourth Line benefits were summed over 15 years and ENRE has typically used a discount rate of about 12%. If for comparability the assumed annual benefit of $6.1m is summed over 15 years at 10% as in other studies, the total benefit amounts to only $47m. This is less than one fifth of the eventual cost. 53 Galetovic and Inostroza (2008-this issue), p. [18], using 10% discount rate. 54 Joskow and Tirole (2005) use a similar diagram, with a net demand curve instead of a supply curve in the rest of the system. 1400 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 Table 1 Congestion revenues, Comahue corridor Year Salex revenue Energy Price difference Year $m GWh/yr $m/MWh 1994 (4 months) 1995 1996 1997 1998 1999 2000 2001 2002 2003 Total Average 29.95 41.59 39.77 18.07 15.78 8.66 51.44 47.62 56.42 22.79 332.09 35.59 6256 16548 13942 14260 11542 10517 16684 19707 17697 17487 144,640 15503 4.79 2.51 2.85 1.27 1.37 0.82 3.08 2.42 3.19 1.30 2.295 Source: CAMMESA annual and monthly reports; Mercados Energéticos calculations. Now suppose that there is a transmission constraint from Comahue at quantity Qo, assumed less than Qu. Output from Comahue will be limited to Qo, a reduction of (Qu − Qo) = ΔQ from the unconstrained level. Output by other generators will be increased by the same amount. Local prices will now apply, namely Pm in Comahue and Pr in the rest of the system. Let ΔP denote the price difference (Pr − Pm). This represents the value at the margin of an increase in transmission capacity between Comahue and Buenos Aires. At the margin, it is economic to build additional transmission capacity if and only if this value exceeds the cost of construction and operation. The marginal value decreases as capacity increases.55 In practice, of course, further adjustments need to be made for changes in demand and supply over time and for indivisibilities, and for discounting benefits and costs over the life of the investment. But a relatively conservative initial calculation, comparing value at the margin with the cost of construction and operation, will shed some light on the orders of magnitude involved. 4.2. Congestion revenues of the Fourth Line With local pricing, the congestion revenue equals (Pr − Pm) times quantity Qo. Table 1 shows the Salex congestion revenues obtaining in the Comahue corridor in each year since the Fund was set up 1994, together with the annual flow on that corridor and the implied average congestion price differential (excluding transmission losses) in each year. The Fourth Line increased existing peak capacity (after the installation of the capacitors) from 3375 MW (Qo) to 4600 MW (Qu). That is an increase of (ΔQ) = 1225 MW, so ΔQ / Q = 1225 / 3375 = 0.36, an expansion of just over one third. The average value of congestion during the whole period was nearly $36m/yr. On that basis, the average value of the additional transmission capacity provided by the Fourth Line was 0.36 × $36m = $13m/yr. The calculation is sensitive to the parameters involved, which evidently varied considerably over the period. During the early period September 1994 to December 1996, the congestion 55 With the specified demand and supply functions, the value of the additional capacity ΔQ needed to remove the transmission constraint entirely is the area of the ‘welfare triangle’ given by 1/2 × ΔQ × ΔP. Since capacity is costly to construct, it will generally not be worthwhile to remove congestion entirely. S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 1401 revenue to the Salex Fund averaged $48m/yr. However, this was mainly before new capacitors (see later) took effect in October 1966. During the middle period 1997 to 1999 the congestion revenue averaged only $14m/yr. Surprisingly, in the later period 2000 to 2002 after the Fourth Line came into effect in December 2000, congestion revenue then increased to an average of $52m/yr.56 Several factors explain this variation. Usage on this corridor was very sensitive to weather conditions, which determined the availability of hydro electricity. For example, 1995 to 1997 were relatively wet years, while 1998/99 was an exceptionally dry year (see Fig. 4 below). Demand in Buenos Aires was generally increasing over this period (though it fell after the crisis). Building the Fourth Line itself stimulated the building of more generation capacity in Comahue (discussed below). In fact, the corridor became as congested after the Fourth Line was built as it had been before.57 4.3. Costs and benefits of the Fourth Line To provide a range of values for the benefit of the Fourth Line capacity, we repeat the above calculation with the average value of congestion ranging from $14m (per the middle period 1997–99) to $52m/yr (per the later period 2000–02). Multiplying by 0.36 (the proportional expansion of capacity) puts the calculated benefit of the Fourth Line in the range $5m to $19m/yr. To put this in perspective, the benefit assumed by Chisari et al. (2001) was a cost reduction of $6.1m/yr, which is at the lower end of the above range. These are very rough calculations, but they suggest that the congestion benefits of the Fourth Line at the time it was built might average about $13m/yr, and lie in the range $5m to $19m/yr. This is very considerably less than the first proposed annual fee of nearly $60m. Following the second vote, that fee was eventually reduced to $24.5m after the bidding competition, to which must be added the annualised value of the $80m Salex contribution, say $11.2m/yr,58 making a total cost of nearly $36m/yr. Even so, the annual cost of the Fourth Line was still three times its estimated average annual congestion value, and double the top end of the estimated range of such benefits. It may be more familiar to express these benefits and costs in terms of $/MWh. The third column of Table 1 shows that the average congestion price differential (excluding transmission losses) ranged from $0.82/MWh in the dry year 1998/99 to $3.19/MWh in the recent year 2002 (and was somewhat higher in the initial part-year 1994). Taking the groups of years used earlier, it was $1.17/MWh in the middle period 1997–9, $2.87/MWh in the recent period 2000–02, and $3.03/MWh in the early period 1994–96. The broad range is thus $1 to $3/MWh.59 Over the whole period 1994 to 2003, the average congestion price differential was $2.295/MWh, say $2.30/MWh. Assume that on average the Fourth Line would increase the previous throughput of the line in direct proportion to the increase in capacity. During the five years 1995 to 1999 preceding the construction of the Fourth Line, the energy transmitted from Comahue to Buenos Aires averaged 56 From 2003 onwards, usage and especially revenues were significantly affected by the economic crisis and the distortions induced by the subsequent policy which froze prices in peso terms. 57 The average load factor was 51% during 1994–99 (the average of 62% 1994–96 and 39% 1997–99), and 47% during 2000–2003, per Fig. 4 below. 58 $80m recovered over 15 years at 12% discount rate is $11.15m/yr. 59 The averages in the early and recent periods are in line with assumed average congestion charges of about $3/MWh used by consultants in evaluating generation businesses in Comahue. Source: Mercados Energéticos. 1402 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 Fig. 4. Average load factors Comahue to Buenos Aires 1994–2003. Source: Authors' calculations based on CAMMESA data. 13,362 GWh per year. Assume the increase would be 0.36 × 13362 GWh = 4810 GWh.60 Dividing this into the annual fee yields an average cost of $58m/4810 GWh = $12/MWh for the first proposal, and $36m/4810 GWh = $7.5/MWh for the second proposal. It will be convenient, for later use in the Appendix A, to express the cost as a function of load factor, and to take a more conservative load factor than the average. In 2000, just after the Fourth Line was commissioned, the load factor was 44%. Take the unit cost of providing extra capacity as $36m/ (1225 MW × 8760 h/yr × 44% load factor) = $7.62/MWh. In other words, congestion benefits plausibly lie in the range $1 to $3/MWh with an average of about $2.30/MWh compared to an initially proposed cost of $12/MWh and an outturn cost of about $7.60/MWh. Once again, the costs are significantly higher than the benefits. 4.4. Possible long-run benefits of the Fourth Line The above calculations of the value of the Fourth Line reflect the degree of congestion on the line around the time of its construction. This in turn reflects the extent and location of generation and transmission as they happened to be at that time. They might be considered short-run benefits. Is it possible that the line would be more valuable, and even economic, if the transmission and generation system were fully adjusted, with generation in the most economic locations and an optimally sized transmission capacity? Such benefit might be considered long-run. This might mean a line more fully loaded (and therefore more congested) when the expansion took place than it actually was in 1999. 60 In fact the average flow over the subsequent three years 2000 to 2002 was 18,029 GWh, an increase of 4667 GWh. S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 1403 A way to approach this is to ask where it was most economic at that time to locate generation. Was it more economic to generate electricity in Comahue, where the gas was found, and then to transmit it to Buenos Aires, or to transport gas from Comahue to Buenos Aires and generate electricity there? The difference in cost between electricity generated in Buenos Aires, and electricity transmitted there, indicates the long-run value of transmission capacity, against which we can compare the cost of constructing and operating it.61 Table 2 sets out the costs of gas and electricity generation, based on conditions typically obtaining during 1997/98.62 On this basis, the long-run benefit of the Fourth Line, assumed working at 100% load factor, would be 1225 MW × 8760 h/yr × $2.58/MWh = $27.7m/yr. At a load factor of 88% (double the 44% load factor assumed in calculating short-run cost) the benefit would be $27.7m × 0.88 = $24.4m/yr. These estimates are about double the central estimate ($13m/yr) of the short-run value of additional transmission capacity calculated above. They are also above the upper end ($19m/yr) of the range of short-run benefits. But they still fall well short of the lowest cost ($36m/yr) of constructing the Fourth Line. The long-run unit cost would be $36m/(1225 MW × 8760) = $3.35/MWh assuming 100% load factor, or $3.81/MWh assuming the line were operated at 88% load factor. This exceeds the calculated long-run benefit of $2.58/MWh. 4.5. Possible quality of supply benefits Can it be argued that there was or is a plausible case for major transmission expansions such as the Fourth Line to improve quality and reliability of supply? Should a quality or reliability benefit – the value of less non-supplied energy – be added to the congestion benefits? Would this make the Fourth Line worthwhile? Table 3 shows that the worst year for non-supplied energy in the Argentine system as a whole was in 1999, when non-supplied energy rose to 14 GWh. In that year, about 7 GWh of the 14 GWh total was attributable to failures of generation and high-voltage (500 kV) transmission lines, mostly the latter.63 The standard “official value” of non-supplied energy for economic dispatch purposes is $1500/MWh.64 On this basis the value of non-supplied energy attributable to weaknesses in high-voltage transmission was at most about 7000 MWh × $1500/MWh = $10.5 million in 1999. About half the outages in 1999 were associated with the Comahue–Buenos Aires corridor.65 The most efficient investment to reduce this non-supplied energy would be to build a fifth (reliability only) line replicating the previous four lines in this corridor, but used only for reliability purposes.66 The reduction of outages by half might be valued at say 1/2 × $10.5m = $5m 61 The recent calculations of Chisari and Romero (2008-this issue) show the relevance to the Fourth Line decision of transporting gas to Buenos Aires and building gas-fired stations there. 62 Generation costs refer to an 800 MW CCGT plant running at 85% load factor, construction cost $420/kW with output priced to yield a 10% internal rate of return. Source: Mercados Energéticos. 63 CAMMESA Annual Report 1999, pp. 80, 85 (Table 2). 64 This value is used by CAMMESA for system operation and market administration activities. It is understood that on occasion ENRE may have invited CAMMESA to estimate benefits in order to provide an initial assessment of the Golden Rule for particular expansions. 65 Source: internal CAMMESA report. 66 This line would not avoid the cost of any double faults, nor would it relieve congestion. The 5 lines could each be run at 4/5 capacity but in order to preserve their reliability property in the event of one failing they could not exceed such loading. 1404 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 Table 2 Costs of gas and electricity generation 1997/98 Location Cost of gas Cost of electricity Buenos Aires Comahue Basin Cost differential Less transmission losses Net differential $1.90/MBTU $1.35/MBTU $27.88/MWh $23.91/MWh $3.97/MWh $1.39/MWh $2.58/MWh (5% BA price) in 1999. The lower loading would also reduce transmission losses by, say, $4m in 1999. This implies a maximum total annual benefit of about $(5 + 4 =) $9m in 1999. In other years around that time the benefit would have been much less, perhaps of the order of half that amount. In contrast, the annual cost of the Fourth Line was about $36m. On this basis, the value of the improved quality and reliability of supply provided by a fifth (reliability only) line from Comahue to Buenos Aires would be at most (in 1999) about one quarter of the cost of obtaining it, and in many years only about one eighth. It cannot be argued that such an investment would be economic. Nor can it be argued that it would have been more economic to reserve the Fourth Line exclusively for reliability purposes rather than for relieving congestion. It is possible that the Fourth Line may have improved reliability as well as reduced congestion, but only to the extent that the Comahue lines were less congested than they were before, which was not in fact the case (see below). 4.6. Costs and benefits of delay In the light of all these considerations, it is difficult to argue that the Fourth Line expansion was economic, in the conventional sense of creating benefits to consumers and producers that exceed the costs of achieving them. It was not economic when it was proposed in 1995 or 1996, or when it was constructed in 1999, or at any time previously or subsequently. This puts in a quite different light the claim that the Public Contest method failed because it delayed a much-needed economic investment. If the Fourth Line was not economic, then any delay would have been beneficial rather than costly. Galetovic and Inostroza (2008-this issue) reach a similar conclusion. Table 3 Non-supplied energy in Argentine system (GWh per year) Year Voltage reduction Shortages Total (MWh) 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 122 43 9 5 1 0 0 0 0 0 0 3 14 15 14 4 8 2 14 8 8 14 125 57 24 19 5 8 2 14 8 8 14 Source: CAMMESA. S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 1405 There were two main savings from delaying the investment. At the very least, there was value in deferring the final total cost of $250m by a year and seven months, say 19/12 × 10% × $250m = $40m. There was also a further saving to the extent that delay enabled the initial cost to be reduced to $250m. It has been calculated that the net present value of the first proposed fee (at 10% interest rate) was $370.1m.67 Some might argue that competition would have been just as effective under the first proposal as under the second, and that the same outcome would have resulted. Others might argue that competition was relatively ineffective at the time of the first vote, was unlikely to have brought the cost down very much, and was made more effective by the steps taken between the first and second proposals.68 On the latter basis, there was an additional value of delay equal in the limit to the difference in cost between the initial bid of $370m and the winning bid of $250m, namely $120m. Thus, depending on the view one takes of the competitive situation, the total value of the delay would be somewhere in the range $40m to $160m. In contrast, the present calculation suggests that the cost to users in terms of continued congestion for 19 extra months was of the order of 19/12 × $13m = $21m. The plausible range might be 19/12 × ($5m to $19m) = $8m to $30m.69 The net benefit of delaying the Fourth Line thus seems to have been at least $10m (= $40m − $30m) and conceivably as high as $152m (= $160m − $8 m). To have rejected the first and more expensive proposal for the Fourth Line is surely to the credit of the Public Contest method rather than an indication of its failure. There were additional less quantifiable benefits associated with delay. For example, it reduced uncertainty, or at least risks, about construction cost. It allowed time for further reflection. It also provided a clear demonstration that the new method required persuasive evidence to justify substantial transmission expansions. An unsubstantiated recommendation by an incumbent transmission company, or a regulator or minister, would no longer suffice. Even proposed expansions whose benefits were taken for granted would be scrutinised and could be rejected. Given the previous history of over-expansion, this was surely a significant merit in terms of encouraging more realistic appraisals of transmission projects in future. 5. The interests involved 5.1. Explaining the first proposal for the Fourth Line If the Fourth Line was fundamentally uneconomic, especially in 1995, why did some of the generators find it worthwhile to propose it? Three factors might be considered. 5.1.1. The influence of history There was a long-standing tradition of building transmission lines from Comahue to Buenos Aires. This was mainly to transmit hydro electricity for peaking purposes, even though these lines were not highly used on average.70 The Fourth Line had been planned consistent with this policy and practice. Since the growth of new gas-fired generation capacity in Comahue was accentuating the extent of transmission congestion, this seemed to indicate that the Fourth Line should go forward. 67 Galetovic and Inostroza (2008-this issue), p. [18]. See also Galetovic and Inostroza (2008-this issue), as discussed below. 69 In a fully adjusted system, the cost of delay might have been as high as 19/12× $28m= $44m, but in the mid-1990s the system was far from being fully adjusted: the load factor (for example) and the value of expansion were correspondingly lower. 70 Under typical rainfall conditions, the average load factor would have been 21% in 1983, 23% in 1987. See Littlechild and Skerk (2004a) pp. 51–2 for more detail. 68 1406 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 Although the concept of local prices was new, and the possible extent and implications of these generation developments and their interaction with the operation of the transmission system were as yet unclear, it was explained to the bidders for generation plant that if more transmission were needed to reduce such congestion, then a new line could be proposed and built, and indeed this was the expected development. In simple terms, this would reduce the number of times that the Comahue corridor would be constrained and that local prices would apply. The Fourth Line would therefore increase the revenues of the Comahue generators.71 By the same token it would reduce the prices paid by Buenos Aires distributors. 5.1.2. Concerns of the new generators Comahue generators that met to discuss building the Fourth Line nonetheless had substantial reservations.72 – First was a concern about free-riding. This was not in the conventional sense of other parties obtaining free use of the new facility paid for by the initiators.73 Rather, if existing generators did build the Fourth Line to accommodate present hydro and gas-fired generation, what was to stop new (or existing) generators building more generation capacity in the future in order to take advantage of the presence of the line, thereby increasing congestion again and causing lower local prices. Moreover, a newer and more efficient plant would have priority in the merit order for dispatch, and would thereby avoid some of the dispatch risks to which older and less efficient plant was exposed. – Second, some generators were concerned that the existing cost allocation rules were unfavourable to them and more favourable to others (including to the initial proponents Alicurá). There was also some suggestion that ENRE had reservations about the general approach to transmission investment, and might be sympathetic to a change in the rules. – Third, generators were pressing the Secretary of Energy to use congestion revenues for investment in this corridor, which he later did in creating the Salex Funds in August 1994. Before this, there was uncertainty as what government policy on congestion revenues might be. Even after this, there was initially some uncertainty as to what level of contribution could be expected from the Salex Fund to offset the costs of building the line. In parallel with these group discussions, each generating company was running dispatch models to try to work out what its contribution to an expansion would be under existing rules, and also trying to evaluate the possibility of alternative rules being introduced. Delaying the Fourth Line decision might or might not resolve some of these issues, and might work to the advantage or disadvantage of particular beneficiaries. But no general consensus was reached. 5.1.3. Search for more economic solutions At the same time, the Public Contest method caused the generators to challenge the need for a new transmission line and to look for more economic alternatives. To that end they hired 71 In calculating what to bid for the hydro stations, Southern Electric, for example, worked the Fourth Line into its projections and assumed (without analysis) that this would solve the congestion issues. 72 Manuel Abdala, then a consultant to the Comahue generators, personal communication, 9 May 2004. 73 After construction, the contributions of each beneficiary were re-calculated each month based on actual usage, and therefore any new users paid in proportion to their usage. Since actual usage was calculated as a rolling average of over the preceding 12 months, it might be argued that the arrangements moderated rather than eliminated ‘free riding’ in the conventional sense. S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 1407 engineering consultants to look for low cost ways of increasing the capacity of the existing system. This met with success. Stabilisation devices were installed in several power plants, which improved transmission capacity and reliability in several corridors including Comahue. The costs were shared between all users of the 500 kV system. A group of Comahue generators (including those who initially proposed and opposed the Fourth Line) also used the Public Contest method to propose the installation of capacitors on the existing lines. This, together with the new control equipment, could increase transmission capacity from about 2700 MW to 3300 MW.74 The proponents of the Fourth Line explained that it was only because these measures would not suffice to avoid increasing congestion that they were proposing the Fourth Line.75 5.2. Private profitability of the Fourth Line Given that the Fourth Line appeared to be uneconomic, how could at least some generators find it profitable? An important part of the answer is that private costs and benefits differed from social or aggregate costs and benefits. The Salex Fund was affected by investment decisions but the beneficiaries of the Fund did not have a vote in these decisions. So a proposal to expand capacity might have been supported by those who would pay lower congestion revenues, but did not take account of those who would suffer from lower congestion revenues.76 Beneficiaries would expect to pay the proposed cost of the line, about $58m annually with the first proposal. Competition for the tender might reduce this, but the extent of competition in construction was at that time largely unknown. The Salex Fund would be used to reduce this total cost. At the time of the first hearing in February 1995, this amounted to $25m.77 This might have been expected to reduce the annual fee by about $3.5m/yr. And distribution companies in Buenos Aires accounted for about 6% of the votes and fees. So the generators would in aggregate expect to pay about (1 − 0.06) × ($58m − $3.5m) = $51m/yr (less any uncertain reductions resulting from competition in construction and higher Salex Funds in the near future). On the benefit side, if congestion were completely removed the existing generators would no longer forfeit the annual congestion revenue (averaging $36m) on their existing output. They would also be able to increase output from existing Comahue plant by 36%. The value of this would be a triangle rather than a rectangle, hence might be valued at half the marginal congestion value, that is at 1/2 × 0.36 × $36m = $6.5m/yr. The total benefit might then be $36m + $6.5m = $42.5m/yr. This benefit is below the calculated cost of $51m/yr. However, some beneficiaries might have expected that congestion revenues would be around the level obtaining in the mid-1990s rather than at the average for the whole period 1994–2002 (which of course could not be known at the time). 74 This was the first completed application of the full Public Contest procedure. In fact, a public hearing was held on this expansion the day before the hearing on the Fourth Line. The proposal passed with no votes against. 75 Testimonies of C Inglesis and L Caruso, Transcript of Public Hearing accompanying ENRE 0049/1995. 76 There was at that time a dispute as to who actually owned the Salex Funds. Generators claimed that they did, and are said to have put it as an asset in their accounts, although they could not actually access the money. So to have attempted to give a vote to beneficiaries of the Salex Fund would not have been straightforward. 77 The proponents calculated that if the Fund had been in operation throughout 1994 the total would have been $55m. They projected that it would grow to $45m by the end of 1998 and increase by a further $15m/yr during 1999 to 2001. But these were speculations. At the time, only $25m was known to be available. Looking at the actual income into the Salex Fund (Littlechild and Skerk, 2008-this issue-b, Fig. 2) and assuming that the process for the first proposal would have taken the same time (19 months) as the second proposal did, we estimate that, in the event, about $117m would have been available by the time that the funds came to be applied. 1408 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 The average for the period September 1994 to December 1996 was $48m (see above) and on that basis the private benefit might be calculated as $48m + (1/2 × 0.36 × $48m) = $57m. This is above the private cost of $51m. On this basis, and adding in something for capacity benefits (see below), at least some of the beneficiaries might have assessed the Fourth Line as a profitable venture. However, against this are certain other considerations. The extent of congestion over the next two years was not then known. As of early 1995, the two congestion spikes in mid- and late-1994 were exceptional (see Fig. 1 above). It could not be assumed that the Fourth Line would completely remove congestion, since many feared that other generators would be encouraged to build new plant in Comahue (as indeed proved to be the case). Such new generators might also replace the existing ones in the merit order, thereby reducing the benefits to existing beneficiaries. At least part of the benefits of congestion reduction would accrue to customers in the form of reduced prices system-wide, rather than to Comahue generators in the form of higher local prices. To estimate the extent of this, we ran some simple cases in a dispatch model. They indicated that, if the transmission expansion eliminated congestion, at most a quarter of the congestion revenue might accrue to customers, and the remaining three quarters to Comahue generators.78 What the beneficiaries expected here is unknown. But if customers or distribution companies took just 10% of the $57m private benefit just calculated, this would eliminate any surplus to generators over the private cost of $51m. This is quite apart from any less optimistic projections of congestion revenue and of the existing generators' share of this. These calculations suggest that the profitability of the Fourth Line was finely balanced. It is possible to envisage a set of parameters under which the Fourth Line would be a profitable proposal for the generators in aggregate, but it is easier to envisage scenarios with the opposite conclusion. It is therefore plausible that at least some generators would vote against the Fourth Line. 5.3. Differences of interest between generators The owners of El Chocón and Alicurá hydro stations proposed the Fourth Line, the owner of Piedra del Águila opposed it. So was it just a simple difference of opinion on future market prices and scenarios and the impact on market participants that led one group to take a different and more optimistic view than the other? Perhaps, but two additional factors seem likely to have had an impact: the rivalry between the owners of two of these companies, and the impact of their interests other than as the identified beneficiaries of this expansion.79 Endesa of Chile had taken a controlling stake in Costanera thermal station in Buenos Aires and El Chocón hydro station in Comahue.80 Southern Electric of the US had won the tender for Alicurá hydro station and was initially expecting to win Piedra too. Endesa and Southern were jointly developing a strategy that involved building the Fourth Line, which they envisaged would be advantageous to them as significant and expanding generators. To that end, they created the potential transmission operator called Tenasa with a view to developing a profitable independent transmission business. 78 With the difference between average system price and Comahue price about $2.10/MWh, eliminating congestion in the Comahue corridor would reduce average price in Buenos Aires by at most $0.5/MWh and increase Comahue price by about $1.6/MWh. This calculation was in a market where prices were lower than in long-run equilibrium. The proportion accruing to customers might be higher in the longer term. On the other hand, a study carried out in preparation for the second vote (see below) calculated that fewer than 5% of the aggregate benefits would accrue to customers. 79 We are indebted to Ruy Varela for these insights. 80 Details of privatisations in Argentina are at http://mepriv.mecon.gov.ar/. S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 1409 If Southern Electric had indeed acquired Piedra, then Endesa and Southern together would have owned the three largest Fourth Line beneficiaries, accounting for 58% of the votes. They would have needed only another 12% of the votes to secure the Fourth Line. In the event, however, Southern did not win the bidding for Piedra. Endesa and Southern nonetheless continued with their Fourth Line strategy. The leading partner in the consortium that won Piedra, namely Chilgener of Chile, had two concerns about this. First, Chilgener was a rival of Endesa in Chile. It was suspicious of the level of the Fourth Line fee offered by the proposed independent transmission company Tenasa, which was part-owned by its rival. It had no wish to pay an excessive transmission charge that would assist a rival company to expand its commercial operations. Second, Chilgener had a significant ownership stake in Puerto thermal station in Buenos Aires and in a forthcoming gas-fired wellhead station at Loma La Lata in Comahue. The profitability of a new line depended critically on the detail of the scenarios assumed, including on the interpretation and knowledge of the transmission system, and the extent of new entry by gas-fired generation in Comahue. In general the calculations suggested that hydro plants would mostly be winners and thermal plants would mostly be losers. Chilgener's prospective new gas-fired plant in Comahue had no votes as an existing beneficiary of the Fourth Line expansion; however, Chilgener naturally took into account the potential adverse impact on this future plant, as well as on Puerto, when exercising its vote as owner of Piedra.81 In summary, the profitability calculations were very sensitive to assumptions. It is possible that some beneficiaries could calculate a net advantage in the Fourth Line. However, it was not clear that Comahue generators as a whole would find the Line profitable at the time of the first vote. What may well have swung the decision of the main proponents of the Line was their interest in entering the newly-open transmission sector. This may have alienated the largest Comahue generator (Piedra), which also had other holdings that might have been disadvantaged by the Line. 5.4. What changed from 1995 to 1996? What changed from 1995 to 1996, to the extent that essentially all the Comahue generators now supported the proposal to build the Fourth Line? The Comahue generators who formerly opposed the expansion joined with the proponents of the extension to form the Group of Electrical Generators from Comahue Area (GEEAC) to explore how a line could be made economic and acceptable to all. Contrary to suggestions otherwise, it was not a problem to build such a coalition.82 Negotiations between the market participants were not unduly costly or problematic, nor did they preclude consensus.83 81 Chisari et al (2001) and especially Chisari and Romero (2008-this issue) emphasise the implications of such external holdings. “Agents can and do make side-payments to build coalitions to support investments, and the generators in Comahue have built such a coalition to support the ‘fourth line’ project, after examining the cost allocations proposed by CAMMESA. The worry is that such coalition building will raise transactions costs and may delay or even prevent worthwhile projects, lowering the quality of supply.” Newbery (1999) p. 254. We have no reason to believe that such side-payments were made, in relation either to the Fourth Line or to other transmission expansions; and neither the ‘cost allocations proposed by CAMMESA’ nor ‘transactions costs of coalition building’ were an issue. See also Galetovic and Inostroza (2008-this issue), and Littlechild and Skerk (2008-this issue-c) re other expansions. 83 As Juan Inostroza has pointed out (personal communication, 1 October 2004), time and costs spent modelling and negotiating with partners, suppliers and customers is the norm in commercial life. In that respect there was nothing exceptional about this negotiation. And it is not as if a regulated approach would obviate the need for this. The parties would still have to incur the time and costs of modelling and negotiating, but with the regulator, transmission company and government rather than with other market participants. 82 1410 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 Among other things, GEEAC sought, and achieved, two modifications to the Public Contest rules. The first modification was to enable proponents to specify a maximum acceptable fee instead of an initial offer.84 It has been argued that the original provision gave an undue advantage to the transmission bid accompanying the proposal.85 The new provision reduced the risk to the beneficiaries that there would be insufficient competition to build the line. In the particular context of the Fourth Line, it also meant that voting for the expansion did not mean supporting a rival company's potential transmission business. The second modification enabled the Salex Fund to be used to defray the up-front expenses of construction, not merely to reduce the payments in subsequent years. This considerably reduced the burden of financing the construction, and again facilitated competition.86 In May 1996 the generators group (GEEAC) presented a new expansion request, for essentially the same Fourth Line project as before. They specified a maximum fee based on what they calculated it would cost to build the line.87 The maximum fee was $43.67m/yr over 15 years (compared to the previous initial offer of $54.6m/yr followed by $61.4m/yr). Part of the reason why they could specify such a lower fee is that the construction cost would now be offset by $80m then available in the Salex account (payable up front) compared to the $25m available on the previous occasion (payable over time). Table 4 shows that, before deducting the Salex contributions, the net present value (NPV) of the initial bid was $370.1m and the NPV of the maximum fee was $346.5m, a reduction of $23.6m or 6.4% of the initial offer from Tenasa. When the new COM contract was put out to tender, a consortium called Atalaya Energy, comprising some of the GEEAC generators, put in a bid of $39.47m/yr.88 In terms of NPV, this was $26.6 m below the maximum fee. But there were three other and better bids, of $38.00 m, $24.99 m and $24.52 m/yr. In terms of NPV the winning bid (by Transener) was $120.7 m below the maximum fee, about 35% below the $346.5 m cost underlying that maximum fee (and $144.3 m or 39% below the initial offer). Competition was evidently very strong.89 In fact, the 84 Originally, the request had to be accompanied by an initial offer of a contract for Construction, Operation and Maintenance (COM) with that Transmission Company, or with an independent transmission company, with a proposed constant annual fee (called a canon) over an amortisation period approved by ENRE. If the lowest fee bid in the Public Contest is not below the fee in the initial offer, then that initial COM contract proceeds. If the lowest bid is less than 85% of the initial offer, ENRE will authorise confirmation of the COM contract with the lowest bidder in the Public Contest. If the lowest bid is between 85 and 100% of the initial offer, the lowest bidder and the initial bidder both have an opportunity to improve their bids within 72 h of the Public Contest (with the lower of the subsequent bids winning). 85 “The defect of an auction involving a BOM [COM] contract is that it gives a sort of first option to the transmission company that accompanies the request, since it will win the auction if no better bid is forthcoming. If this option proposes a toll that is too high, as some of the beneficiaries believed, the losers of the auction are forced to finance the project.” Galetovic and Inostroza (2008-this issue), p. [16]. 86 At the same time the modification limited to 70% the proportion of the expansion cost that could be defrayed by the Salex Fund. It was considered desirable to leave some cost or risk on the beneficiaries in order to leave them with an incentive to act efficiently. 87 “They simulated the financial results of an independent transmission firm whose aim was to build, operate and maintain this fourth line over a 30-year horizon. This exercise provided the value of the proposed maximum fee, which gave an internal rate of return of 12.55% (real) for shareholders of the future transmission firm.” Galetovic and Inostroza (2008-this issue) p. [16]. 88 The consortium comprised four generators from GEEAC (Capex, Central Puerto, Piedra del Águila and El Chocón) and one construction company (Inepar), each with a 20% share. El Chocón was one of the proponents of the first proposal, while Piedra del Águila had voted against. 89 The winning bid only just beat the bid submitted by one of the potential independent transmission firms LitsaCartelone ($24.99 m). The keenness of the winning bid presumably reflected, amongst other things, Transener's concern not to lose its pre-eminent position in transmission. In the run-up to the bidding, competition also took other forms: technical, procedural, regulatory, jurisdictional, legal, etc. Galetovic and Inostroza (2008-this issue) describe these aspects of “the battle between transmission firms”. Littlechild and Skerk (2008-this issue-c) calculate the effective reduction in construction costs exemplified by this line. 1411 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 Table 4 Costs and contributions related to the Fourth Line NPV a NPV reduction $m $m $m 54.6/61.4 370.1 Item Annual fee Initial offer August 1994 Prospective 1994 Salex contribution $25m Prospective 1996 Salex contribution $80m Initial bid net of Salex $80m Maximum fee May 1996 net of Salex $80m Maximum fee May 1996 incl. Salex $80m Reduction in cost compared to initial offer Maximum fee May 1996 net Highest bid (Atalaya Energy) net Reduction compared to max fee Lowest bid (Transener) net Reduction compared to max fee Additional Salex Fund $47.9m Final cost to users net of all Salex 22.4 71.8 43.67 298.3 274.7 346.5 43.67 39.47 274.7 248.1 24.52 154.0 23.6 26.6 120.7 38.7 115.3 a These NPV calculations follow Galetovic and Inostroza (2008-this issue) in assuming that the $80m Salex contribution is added to all except the initial proposal and is paid $68.98m and $11.04m at ends of years 1 and 2, respectively, and that the canon is paid at the ends of years 3 to 18, and using a 10% discount rate. final price paid by the beneficiaries was even lower because by the time the line was commissioned the Salex Fund was about $48 m larger than assumed at the time of the second vote.90 Net of Salex contributions, the final price paid by users was $115.3 m NPV, less than one third of the cost implicit in the initial 1994 proposal. The situation had also evolved in respect of other concerns (more precisely, there was a change in expectations about these factors.) The lower level of the maximum fee compared to the initial bid, the clearer expectation of competitive bids below that level, and the higher level of Salex contribution all served to reduce the cost of the transmission investment. So if it led to some freeriding (in the sense of encouraging further generation investment) less cost had been incurred. In fact generation capacity in Comahue had already increased faster than expected, leading to greater congestion and a larger Salex Fund. This not only reduced the cost of the Fourth Line to the users but also made more acute the consequences of not investing to avoid congestion and local prices.91 Finally, the users no doubt realised that there was now no immediate prospect of further changes to the rules that might warrant delaying a decision. How would the previous calculations of private profitability appear at the time of the second vote? The presumed knowledge that Atalaya Energy would be bidding around $40 m (net of Salex), less the 6% of the fee that Buenos Aires distribution companies might expect to pay, suggests that it might cost the generators about 0.94 × $40 m = $38 m/yr. This is somewhat below the $51m calculated above. The prospects for further reductions from higher Salex revenues by the time of construction, and stronger competition to construct the line, would also have seemed better than before. 90 Littlechild and Skerk (2004a). The impact on nodal prices was exacerbated by separate developments that refined transmission price signals with the effect of increasing peak-offpeak differentials. Resolutions SE 105 of 20 March 1995 and SE 151 of17 April 1995. 91 1412 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 The prospect for benefits would also be better. Congestion had continued on the Comahue corridor. $48m a year (the average 1994 to 1996) would now seem quite plausible congestion revenue. If this congestion could be entirely removed, an aggregate annual benefit of $57m including from additional output would not seem out of reach. Even if only three quarters of the congestion could be removed, and if 10% of the benefits were lost to customers in the form of lower system prices, this would still leave the generators with benefits of $38m, sufficient to cover the estimated private cost. A study commissioned from consultants at the time sheds light on the situation as they perceived it.92 The study estimated lower benefits to generators in terms of energy prices in Comahue than conjectured above, but assumed that distribution companies would secure a lower proportion of total benefits (and negative benefits in 2003). On the other hand, the study estimated significant capacity benefits, of the same order as energy benefits. Other market participants are understood to have taken a more conservative view of capacity benefits. In retrospect, the estimated capacity benefits seem on the high side, especially seen from 2003. The total benefits to generators as estimated in this study were nonetheless sufficient to cover the estimated costs to them of building the line. 5.5. Effect of the Fourth Line on generation and transmission load factors After the agreement to build the Fourth Line, there was indeed a further increase of over 500 MW in Comahue generation, as the hydro generators feared, partly from a now-completed hydro scheme and partly from additional gas-fired capacity. This had implications for load factors. Fig. 4 shows that in the three years 1994–1996 the average load factor on the Comahue– Buenos Aires corridor was 60%. With the installation of the capacitors and with an exceptionally dry year,93 the average load factor fell to 40% during 1997–1999. But after the commissioning of the Fourth Line the average load factor actually increased to 48% during 2000–2003. Taking the post-privatisation period 1994–2003 as a whole, the load factor averaged 50%. This stands in contrast to the load factors of just under 25% in conditions typical of the 1980s. After privatisation, both before and after the Fourth Line was constructed, the load factor of the Comahue corridor was roughly double what it was before privatisation. This meant that the Fourth Line did not significantly reduce the load factor on that corridor, and certainly not to the level seen in earlier (pre-reform) days. Instead, as demand increased the transmission system was used more efficiently. 5.6. Alternative ways of meeting increased demand What do these developments say about the efficient use of resources? Given the available transmission capacity in the Comahue corridor, it made economic sense for the thermal generators in Comahue to build new plant to use surplus flared gas, to fill up the available offpeak capacity on the existing three lines, and to do the same again once the Fourth Line was to be built. But was it economic to build such transmission lines in the first place, and would it be economic to build more lines in the future? 92 For further details see Littlechild and Skerk (2004b). Dry year refers to the hydrological year, which in Comahue starts in June, with the first snow, and finishes in May/ June of the next year. 93 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 1413 Fig. 5. Comparative costs of electricity and gas transmission. Source: Authors' calculations based on CAMMESA data. The strategy before privatisation was essentially to build hydro stations in Comahue in order to serve peak demand in Buenos Aires, 1300 km away. This meant transmission lines operating at less than a quarter of capacity. As Roark (1997) pointed out, this was an expensive policy. When generating companies had to bear all the costs of production including transmission, strategy changed. Since the Fourth Line, no further generation has been built in Comahue, except for some minor projects for self-supplying oil fields. And no further new electricity transmission lines have been built from Comahue to Buenos Aires. In contrast, some 4000 MW of new combined cycle generation capacity has been built near Buenos Aires. Initially this took advantage of spare gas pipeline capacity. But there has also been significant investment in gas pipelines, especially from Comahue to Buenos Aires. This is in fact a more economic long-term solution. Fig. 5 compares the costs of the two main alternatives assuming a ‘greenfield situation’. For 2 GW generating capacity running at 60% load factor, the cost of transporting gas and converting to electricity in Buenos Aires would be about one sixth less than the cost of generating in Comahue and transmitting the electricity to Buenos Aires.94 In other words, at these costs it is unlikely to be economic to build more large-scale electricity transmission lines directly between Comahue and Buenos Aires.95 6. Conclusions In privatising its electricity sector in 1992, Argentina adopted innovative arrangements with respect to the regulation of transmission expansion. With the exception of minor investments, the incumbent transmission companies and the regulator were not allowed to initiate expansions in capacity. It was for users to propose, accept and finance major expansions, using a prescribed voting scheme called the Public Contest method. 94 This assumes a high voltage AC line. Since there are economies of scale for gas transmission but not for electricity, the cost of transporting gas equivalent to 8 GW of capacity is about half the cost of transmitting that electricity. 95 There may be scope for worthwhile local reinforcements of the transmission system given existing capacities, such as a Fifth Line via Comahue–Cuyo to take advantage of spare capacity between Cuyo and Buenos Aires. (In relation to Fig. 1, note that the main 500 kV node in Cuyo is at Gran Mendoza.) Littlechild and Skerk (2008-this issue-c) discuss the subsequent development of the National Transmission Plan. 1414 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 It was widely held (by economists, consultants and even the regulator) that this particular policy innovation was unsuccessful. That conclusion was based almost entirely on the view that the Public Contest method delayed by many years a “much-needed” Fourth Line into Buenos Aires. It led to the question: why did the Public Contest method fail to deliver economic transmission investments? Proffered and widely accepted explanations for this ‘failure’ include voting rights based on usage instead of benefits, exclusion of the important Buenos Aires demand node from the calculation, inadequate transmission property rights, and transactions costs associated with problems in negotiating and securing consensus among the parties involved. Accordingly, Argentina has been held up as an example of “how not to do it” with respect to transmission regulation. More generally, its experience has been used to suggest that conventional methods of regulation should not be replaced by methods that give a greater role to market participants. This paper has argued that in fact all these perceptions are incorrect. Evidence suggests that the Fourth Line – which was actually delayed by only a year and a half rather than by four or many years – was not “much-needed” at all. It was an uneconomic project in terms of aggregate net benefit. Deferring this investment was economically beneficial rather than costly, and in addition secured a substantial reduction in the cost of the Line. The reason for rejecting the initial proposal was not the pattern of voting rights, the exclusion of the Buenos Aires demand node, the inadequacy of property rights, the level of transactions costs or the inability of the parties to work together. Some of the elements of the Area of Influence method may not have been ideal, but in practice that method was not challenged. The reason for rejecting the initial proposal was the unprofitability of the first proposal to the main parties involved. None of the conjectured contributory factors proved an obstacle to approving the expansion once the underlying profitability conditions changed. In fact the generators that voted against the initial proposal worked actively with the proponents to develop a proposal that all could support, and this succeeded. This evidence prompts a quite different set of questions. Why was the Fourth Line proposed in the first place if it was uneconomic? Answer: two main proponent generators had an additional interest in using the Line to enter the transmission business. Why was the Fourth Line proposed and accepted on the second vote if it was still uneconomic? Answer: there was a divergence between private and social costs as a result of the application of the Salex Fund, and there was a significant change in conditions that made the expansion potentially profitable for sufficient market participants (albeit still uneconomic from an aggregate perspective). The main changes were the increased availability of Salex Funds and the lower cost of construction and operation. The latter reflected more competitive bidding conditions and the removal of the tie-in to a transmission company owned by the initial proponents. Why there was such a strong view that the Fourth Line was desirable, and that delay constituted a serious failure? Answer: as public choice theory (well documented elsewhere) explains, politicians, governments and regulators all have an interest in promoting new transmission lines. Transmission companies and contractors have an interest in building them. Distribution companies get improved quality of service and therefore lower penalty costs for non-performance. Some generators benefit from less restricted access to higher priced markets. For all these parties, the potential benefits of transmission expansion are significant. Under the conventional regulatory process, it costs many of them little or nothing. The costs of transmission expansion are spread among a large number of consumers in the system as a whole. Such consumers do not find it worthwhile to express a view, if indeed they are aware of what is happening. In consequence, transmission expansions find ready support and little opposition. In contrast, the Public Contest method focuses the costs of expansion on the potential beneficiaries, who have to think seriously about approving the investment. It is therefore not surprising that the beneficiaries took a more realistic view when they realised the cost to them. S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 1415 The initial adverse vote was disappointing to those who did not have to pay for the Fourth Line, and as a result the Public Contest method itself was readily held to have failed. It is perhaps disappointing that economists did not investigate the facts of this claim earlier.96 In the light of this new perspective, should the Public Contest method now be criticised because it failed to stop the Fourth Line being built, rather than because it delayed it? That would be harsh. This particular decision seems to have been swung by the size of the subsidy from the Salex Funds, which constituted a divergence between private and social costs. This was not intrinsic to the concept of the Public Contest or the Area of Influence method. Nor does it seem to have been led to uneconomic decisions on other expansions (see Appendix A). It was a significant achievement of the Public Contest method to force a much-needed reappraisal of pre-reform transmission investment policy. That policy was no longer appropriate for the future. The challenge laid the way for a subsequent change to a more economic transmission policy: for transmitting gas to Buenos Aires instead of electricity. Viewed over this longer-term perspective, far from demonstrating the failure of the Public Contest method, the Fourth Line experience taken as a whole is testament to its success. Acknowledgements We are grateful for comments and information from many people who have participated in the development of this industry, including Manual Abdala, Diego Bondorevski, Alejandro Capara, Luis Caruso, Omar Chisari, Roberto D'Addario, Guillermo del Georgio, Sally Hunt, Juan Ricardo Inostroza, Jorge Karacsonyi, Alfredo Mirkin, Ignacio Pérez-Arriaga, Martín Rodríguez-Pardina, Jeffrey Roark, José Sanz, Ramón Sanz, Luis Sbértoli, Pablo Spiller and Ruy Varela. None of these is responsible for the use we have made of their advice. The first author is grateful to the (then) Judge Institute of Management Studies and to the TSEC grant to the Electricity Policy Research Group at Cambridge University for supporting this research. Appendix A. Analysis of Public Contest expansions using Salex Funds We have argued that the availability of the Salex Funds deriving from local price differentials made it attractive for market participants to vote for the Fourth Line transmission investment that was uneconomic from an aggregate perspective. Was that also true for other expansions? Was the application of the Public Contest method seriously flawed as a result of the availability of the Salex Funds, or was the Fourth Line an exception? The potential ‘problem’ is limited in scope because Salex Funds can be applied only where there is congestion and therefore funds available, and where the expansion ameliorates this congestion. Littlechild and Skerk (2008-this issue-c Appendix) identify 36 transmission expansions that have been proposed under the Public Contest method. Of these, 11 expansions made use of Salex Funds. Six of these expansions were actually (or in one case effectively) proposed by the transmission company, which did not stand to gain from the reduced congestion or the use of the Funds insofar as it did not pay for the expansion. This leaves only five expansions proposed by users that reduced congestion and benefited from the Salex Funds. 96 “If it is believed that their theory tells us how people would behave in different circumstances, it will appear unnecessary to many to make a detailed study of how they did in fact act. This leads to a very casual attitude toward checking the facts. If it is believed that certain contractual arrangements will lead to opportunistic behaviour, it is not surprising that economists misinterpret the evidence and find what they expect to find.” Coase (2006) p. 275. 1416 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 The best known of the five expansions (#7 in that Appendix) was the Fourth Line in the Comahue transmission corridor. Another (#26a) involved the installation of capacitors at Choele Choel and Olavarría in order to expand the effective capacity of existing 500 kV lines in the same corridor. It was later discovered that these capacitors over-loaded the Third Line reactors, which had to be replaced (#26b). We have therefore also evaluated a combined expansion (#26a + b) rather than looked simply at the benefits and costs of #26b given that #26a was already in place. Another expansion (#6) was a 500/132 kV transformer at Chocón to enable generators in the 132 kV system to access the 500 kV lines in the same Comahue corridor. The fifth expansion (#9) was to introduce capacitors at Recreo in the North West transmission corridor. In line with the calculations for the Fourth Line in the present paper, for each of these five expansions we have tried to calculate plausible values for the short-run benefit and cost of the expansion, given the circumstances of each corridor at the time. Where relevant we have also calculated long-run estimates assuming the system is fully adjusted (and typically more congested than it actually was). Benefits (#7, 26a,b, 9) The same range of congestion benefits has been assumed for the capacitors (#26a) in the Comahue corridor, and for the combined capacitors and reactors (#26a + b), as for the Fourth Line (#7), though the timings differ by a few years. That is, we estimate $1 to $3/MWh, average $2.30, in the short run, based on the congestion charges actually paid in the Comahue corridor around that time. We again estimate $2.68/MWh in the long run, based on the lower cost of gas-fired generation in Comahue compared to gas-fired generation in Buenos Aires. For the Northwest corridor (#9) we have taken the range of short-run benefits as the average congestion value before ($0.62/MWh) and after ($0.35/MWh) the introduction of the capacitors, with an overall average of $0.49. The long-run benefit is $2.32/MWh, based again on natural gas prices but slightly below that for #7 in the Comahue corridor because the distance from Buenos Aires is less. Costs (#7, 26a,b, 9) We have calculated the average cost of expansion on a common basis, by converting the published cost or fee (before application of Salex Funds reductions) to an equivalent annual fee over 15 years, then dividing by 8760 capacity-hours per year at an assumed load factor (44% in the short run, 88% in the long run).97 On this basis, long-run costs are half the level of short-run costs. Benefits and costs for #6 The transformer (#6) presents a more complex evaluation. It reduces congestion between the 132 kV and 500 kV systems. This has two benefits: first, it enables the 132 kV generators to 97 Specifically, the short run costs are as follows (long run costs are half these levels): #7 Fourth Line $36m/(1225 MW × 8760 × 44%) = $7.62/MWh; #26a CC–O capacitors $2.0 m/(200 MW × 8760 × 50%) = $2.59/MWh; #26a + b CC–O capacitors plus reactors $2.2m/(200 MW × 8760 × 50%) = $2.85/MWh; #9 Recreo capacitors $1.4m/(200 MW × 8760 × 44%) = $1.82/MWh. For the first Fourth Line proposal (#3) where the anticipated cost was $58m/yr rather than $36m/yr, the short-run cost would have been $58m/(1225 MW × 8760 × 44%) = $12.28/MWh. S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 1417 access the prices available at the relevant 500 kV node before the installation of the Fourth Line (#7), which are nonetheless higher than the prices at the relevant 132 kV node; second, it enables these generators to benefit from the higher prices available at the 500 kV node after the installation of the Fourth Line. However, unlike the other cases there is no overall increase in output, and that part of these benefits that refer to the existing 132 kV output are simply a transfer of income from Salex to existing generators. The social benefits derive from the ability to convey more hydro power from the 132 kV to the 500 kV system in the more valuable peak period instead of in the less valuable offpeak period. We therefore calculate the benefits and costs per MWh transferred from peak to offpeak. Congestion cost for the 132/500 kV link totalled $751k in the two years 1995 and 1996 before #6 was proposed, an average of $751k/2 = $375.5k per year. Assume that congestion occurred entirely in peak hours, defined by the market rules as those 5 h with highest load each day. This means that energy affected by congestion prior to the expansion was 200 MW × 8760 × 5.24 = 365 GWh per year. Hence the average value of this congestion is $375.5k/365 GWh = $1.03/MWh(peak). This is the value before #7 is in place, which increases the value of 500 kV generation by between $1 and $3, on average $2.30/MWh. We therefore take the short-run benefit of the transformer as $1.03 + $2.30 = $3.33 per 132 kV MWh transferred from offpeak to peak, with a range of $2 to $4. The concept of a long-run evaluation is inapplicable for #6 because the hydro generators in the 132 kV system that are enabled to access the 500 kV system were built under the previous State-owned regime and would not be constructed in an economic long-run scenario. Assume that #6 expansion capacity (120 MW) is fully used at peak with output previously provided offpeak. This implies a maximum transfer of 120 MW × 8760 × 5/24 = 219 GWh from offpeak to peak. The annualised cost of the transformer is $382k per year, so the annual cost per MWh transferred is $382k/219GWh = $1.74/MWh(peak). Table 5 sets out the resulting values of short-run and long-run benefits and costs. As calculated in the main text, the range of short-run benefits of the Fourth Line (#7) lies considerably below the lowest short-run cost of the Line, and the long-run benefit of that line also lies below the long-run cost. In contrast, for the CC–O capacitors (#26a), even including the cost of replacing the reactors (#26a+b), the average short-run benefit is only a little below the short-run cost, and the long-run benefit is about double the long-run level of cost. For the Chocón transformer (#6) the short-run benefit is about double the short-run cost. For the Recreo capacitors (#9), the short-run benefits are below the short-run cost, but the long-run benefit is more than double the level of long-run cost. In other words, the Comahue capacitors seem economic even including the unanticipated cost of replacing the reactors. The Chocón transformer is economic given the Fourth Line (though it would not have been otherwise). The Recreo capacitors seem premature but potentially economic, Table 5 Benefits and costs ($/MWh) for user-proposed expansions with Salex Funds Expansion SR benefit SR cost LR benefit LR cost #6 Chocón transformer #7 Fourth Line #26a CC–O capacitors #26a + b CC–O capacitors + Replacement of reactors #9 Recreo capacitors $2–$4, ave. $3.33 (peak) $1–$3, ave. $2.30 $1–$3, ave. $2.30 $1.74 (peak) $7.62 $2.59 n.a. $2.68 $2.68 n.a. $3.81 $1.30 $1–$3, ave. $2.30 $0.35–$0.62, ave. $0.49 $2.85 $1.82 $2.68 $2.32 $1.43 $0.91 1418 S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 at least under growth expectations obtaining at the time.98 In contrast, the Fourth Line seems uneconomic under any plausible short-run or long-run circumstances. The Fourth Line thus seems an exception in terms of economic value, and not characteristic of the operation of the Public Contest method generally. References Abdala, Manuel A., 1994. Transmission pricing in private-owned electricity grids: an illustration from the Argentine electricity pool. XXIX Reunión Anual de la Asociación Argentina de Economía Política. LaPlata: November. Abdala, Manuel A., 2008a. Transmission pricing in privately-owned electricity grids: an illustration from the Argentine electricity pool. Energy Economics 30, 1284–1305 (this issue). doi:10.1016/j.eneco.2007.12.010. Abdala, Manuel A., 2008b. Governance of competitive transmission investment in weak institutional systems. Energy Economics 30, 1306–1320 (this issue). doi:10.1016/j.eneco.2007.12.009. Abdala, Manuel Ángel, Chambouleyron, Andres, 1999. Transmission investment in competitive power systems: decentralising decisions in Argentina. Public Policy for the Private Sector, No. 192. The World Bank, Washington. September. Abdala, Manuel A., Spiller, Pablo T., 2000. Decentralised Investment and Quality Decisions in Common Pool Networks. AAEP - Córdoba. 13 November. Baldick, Ross, 2006. Border Flow Rights and Contracts for Differences of Differences Technical Reference. http://www. ece.utexas.edu/~baldick/papers/CFDD_technical.pdf. Bastos, Carlos M., Abdala, Manuel A., 1996. Reform of the Electric Power Sector in Argentina. Buenos Aires. (originally published in Spanish in 1993). Bouille, Daniel, Dubrovsky, Hilda, Maurer, Crescencia, 2002. Argentina: market drive reform of the electricity sector. chapter 3 In: Navroz, N.K. (Ed.), WRI Power Politics. World Resources Institute, Washington DC, pp. 31–50. Bushnell, J., Stoft, S., 1996. Electric grid investment under a contract network regime. Journal of Regulatory Economics 10, 61–79. Bushnell, J., Stoft, S., 1997. Improving private incentives for electric grid investment. Resource and Energy Economics 19, 85–108. Chao, H.P., Peck, S., 1996. A market mechanism for electric power transmission. Journal of Regulatory Economics 10 (1), 25–59. Chisari, O.O., Dal-Bó, P., Romero, C.A., 2001. High-tension electricity network expansion in Argentina: decision mechanisms and willingness-to-pay revelation. Energy Economics 23, 697–715. Chisari, Omar O., Romero, Carlos A., 2008. Investment decisions in transmission of electricity: the case of a club with imperfect representation. Energy Economics 30, 1321–1333 (this issue). doi:10.1016/j.eneco.2007.12.014. Coase, Ronald, 2006. The conduct of economics: the example of Fisher Body and General Motors. Journal of Economics & Management Strategy 15 (2), 255–278 Summer. Estache, Antonio, Rodríguez-Pardina, Martin, 1996. Regulatory lessons from Argentina's power concessions. Public Policy for the Private Sector, Note No. 92. World Bank, Washington DC. Galetovic, Alexander, Inostroza, Juan Ricardo, 2008. Electric power transmission and the Ley Corta [‘Short Law’]: why auctioning is (much) better than regulating. Energy Economics 30, 1334–1366 (this issue). Gómez-Ibáñez, Jose A., 2003. Designing capacity markets. (with Martin Rodríguez-Pardina). Chapter 12 Regulating Infrastructure: Monopoly, Contracts, and Discretion. Harvard University Press, Cambridge, Mass. Hogan, W., 1992. Contract networks for electric power transmission. Journal of Regulatory Economics 4, 211–242. Hogan, W., 2003. Electricity deregulation: where to from here? Paper Presented to Conference at Bush Presidential Conference Center. Texas A&M University. 4 April 2003. Joskow, Paul L., 2003. Comments of Professor Paul L Joskow. Before the Federal Regulatory Commission, Docket No. RM01-12-000. 10 January. 98 The Northwest corridor has been characterised by changing expectations. “The 500 MW line from Almafuerte up to El Bracho [through Recreo] was originally installed in 1987 to provide a load flow to meet demand in the northwest. But after the market developed, it began to appear attractive to install generation in the northwest, and to use the line in the opposite direction, south from El Bracho through Almafuerte to supply Buenos Aires.” (Littlechild and Skerk 2004b, section 1.2) The Recreo capacitors were proposed in 1997 as an economic way of meeting this increasing generation and demand. But at about the same time as the capacitors came into service in 2000, a new Australian mining project increased local northwest demand by about 200 MW. The load factor of southbound transmission was significantly reduced, and in some hours load flow was in the opposite direction. S.C. Littlechild, C.J. Skerk / Energy Economics 30 (2008) 1385–1419 1419 Joskow, Paul L., 2006. Incentive regulation in theory and practice: electricity distribution and transmission networks. National Bureau of Economic Research Conference on Economic Regulation, September 9–10, 2005. revised version January 21, 2006. Joskow, Paul L., Tirole, Jean, 2005. Merchant transmission investment. Journal of Industrial Economics 53 (2), 233–264 June. Léautier, T.-O., 2000. Regulation of an electric power transmission company. The Energy Journal 24 (1), 61–92. Léautier, T.-O., 2001. Transmission constraints and imperfect markets for power. Journal of Regulatory Economics 19 (1), 27–54 January. Littlechild, Stephen C. 2003. Transmission regulation, merchant investment, and the experience of SNI and Murraylink in the Australian National Electricity Market. Mimeo 12 June, available on Harvard Electricity Policy Group website and on EPRG Cambridge website. Littlechild, Stephen C., 2004. Regulated and merchant interconnectors in Australia: SNI and Murraylink revisited. Cambridge MIT Institute Electricity Project, CMI Working Paper EP 37. The Cambridge-MIT Institute. (16 December 2003). Littlechild, Stephen C., Ponsano, Eduardo A., 2008. Transmission expansion in Argentina 5: the Regional Electricity Forum of Buenos Aires province. Energy Economics 30, 1491–1526 (this issue). doi:10.1016/j.eneco.2007.12.005. Littlechild, Stephen C., Skerk, Carlos J., 2004a. Regulation of transmission expansion in Argentina: part I — state ownership, reform and the Fourth Line. Cambridge - MIT Institute Electricity Project, CMI Working Paper 61, CWPE 0464. The Cambridge-MIT Institute. 15 November. Littlechild, Stephen C., Skerk, Carlos J., 2004b. Regulation of transmission expansion in Argentina: part II — developments since the Fourth Line. Cambridge - MIT Institute Electricity Project, CMI Working Paper 62 CWPE 0465. The Cambridge-MIT Institute. November. Littlechild, Stephen C., Skerk, Carlos J., 2008a. Transmission expansion in Argentina 1: the origins of policy. Energy Economics 30, 1367–1384 (this issue). doi:10.1016/j.eneco.2007.12.003. Littlechild, Stephen C., Skerk, Carlos J., 2008b. Transmission expansion in Argentina 3: the evolution of policy. Energy Economics 30, 1420–1461 (this issue). doi:10.1016/j.eneco.2007.12.011. Littlechild, Stephen C., Skerk, Carlos J., 2008c. Transmission expansion in Argentina 4: a review of performance. Energy Economics 30, 1462–1490 (this issue). doi:10.1016/j.eneco.2007.12.006. Littlechild, Stephen C., Skerk, Carlos J., 2008d. Transmission expansion in Argentina 6: distribution companies, regulation and the Public Contest method. Energy Economics 30, 1527–1535 (this issue). doi:10.1016/j.eneco.2007.12.008. NERA, 1998. Analysis of the Reform of the Argentine Power Sector: Final Report. New York, January. Newbery, David M., 1999. Privatization, Restructuring, and Regulation of Network Utilities. The MIT Press, Cambridge, Mass. Pollitt, Michael, Blignaut, J.N., 2008. Electricity Reform in Argentina: Lessons for Developing Countries. Energy Economics 30, 1536–1567 (this issue). doi:10.1016/j.eneco.2007.12.012. Roark, Jeffrey D., 1997. Argentina's wholesale electricity market. Potencia, ITP Editorial. February (translation of Spanish original). Rossellón, Juan, 2003. Different approaches towards electricity transmission expansion. Review of Network Economics 2 (3), 238–269 September. Rotger, J.A., Felder, F.A., 2001. Reconciling market-based transmission and transmission planning. The Electricity Journal 31–43 November. Spiller, Pablo T., Torres, Clemencia, 1996. Argentina's electricity regulation: its performance, credibility and options for the future. Report to the Ministry of Economics, Argentina. 19 November (confidential). Vogelsang, I., 2001. Price regulation for independent transmission companies. Journal of Regulatory Economics 20 (2), 141–165. Woolf, Fiona, 2003. Global Transmission Expansion: Recipes for Success. Pennwell, Tulsa, OK.