The measurement of the overall impact of the single electricity market, on prices, costs, and wider GDP is difficult—as the attempts by the Commission to study this illustrate.
We can start by thinking about how the single market is supposed to have an effect, through bringing about pro-competitive structural change and improving the quality of sector regulation, before looking at the measured economic impact.
Pro-competitive Structural Change and the Quality of Sector Regulation
The OECD Product Market Regulation (PMR) index provides a relative measure of the degree of deregulation of a nation’s electricity sector.Footnote 31 The index has a range of 0–6, where 0 is most liberalised and 6 would be an unreformed integrated government monopoly. This index puts one-quarter weights each on: (a) entry regulation; (b) the percentage of public ownership; (c) the degree of vertical separation; and (d) the market share of the largest company in the sector.
For entry (a), a combination of free-entry regulation, regulated third-party access (TPA), a liberalised wholesale power market, and no minimum threshold on the size at which a customer can chose its supplier would give the lowest possible liberalisation score of 0. Under vertical integration (c), ownership separation gives the lowest possible liberalisation score of 0 (legal separation scores 3). Under market structure (d), a combined generation/import and supply share of less than 50% for the largest firm gives a score of zero. If 100% of the shares in the firms across generation/import, transmission, distribution, and supply were government-owned (b), this would score 6.
The implementation of the single electricity market directives should result in a 0 score for entry regulation (a) and for market share of largest firm in the sector (d) and a maximum of 3 for vertical integration (c) and less than 6 (due to some entry) for public ownership (b). This should imply that the maximum EU member state score under the PMR is 2.3. Indeed between 2008 and 2013 the PMR score for EU electricity did improve from 2.4 to 2.1, relative to 2.6 to 2.4 for the OECD.Footnote 32
Since 1999, there has been a significant increase in generation competition within MSs: The Czech Republic, France, Greece, and Luxembourg have all increased their number of main generators (a main generator is company that produces at least 5% of national production) from 1 in 2003 to an average of 2.75. Effective competition has also significantly increased in Germany due to increased renewables.Footnote 33 Estonia, France, Croatia, Latvia, and Portugal have increased their number of main (with 5% of national consumption) retailers from 1 in 2003 to an average of 2.6.Footnote 34
A developing theme of the single market in electricity project has been the need to improve the quality of both national and EU level regulation of the sector. This is discussed in Green et al. (2009). They point out the importance of improving the form, process, and outcomes of energy-sector regulation.
The form of regulation can be measured on the basis of the strength of the regulator in terms its separation from government and its ability to regulate the industry. Jamasb and Pollitt (2005) scored ex ante regulation—no ministerial involvement in decision making, network access conditions set by the regulator (rather than elsewhere), dispute settlement by the regulator, and strong information acquisition powers—at 5 out 5, for regulatory strength. In terms of the scope of the regulator’s functions, Larsen et al. (2005) found that 15 European regulators had from 1 to 7 different legislative objectives.Footnote 35 Overall, Pollitt (2009) notes improvements in the form of regulation at the national level since the beginning of the single market project.
The process of regulation focuses on: the degree of transparency that is exhibited by the regulator; the amount of stakeholder engagement; the procedural efficiency in coming to regulatory decisions; and the quality of the underlying techniques that are utilised in carrying out their functions. Regulators are thought to have better processes when they: are more transparent in their decision making; deliver their decisions in a timely way; undertake good levels of stakeholder engagement; and use best-practice techniques of regulation.Footnote 36 Transparency has increased significantly with respect to regulation across the single market area.Footnote 37
The outcomes of regulation are more difficult to measure. This is because regulators do operate under different rules and objectives. Process in a democracy matters for its own sake: Stakeholders do like and need to be consulted. Regulatory decisions can only be evaluated ex post, and it is not clear how to do this, given that there is very unlikely to be an unbiased, statistically identifiable measure of impact. Nillesen and Pollitt (2007) discuss a case study of the miscalculation of price controls in the Netherlands and how much this cost Dutch customers. It is possible to try to benchmark regulator costs across countries, but this would need to be corrected for the number of regulatory functions and economies of scale in regulation (see Domah and Pollitt 2001).
The single electricity market project has undoubtedly spread good practice in the form, process, and outcomes of regulation—especially initially, as noted by Jamasb and Pollitt (2005). The directives have promoted independent sector regulation of electricity—notably forcing Germany to have an electricity regulator, when it initially did not. CEER (the industry association of regulators) and ACER have been the fora for the spread of good regulatory practice, and many European countries have much better regulators then they would otherwise have had.
The Measured Economic Impact of the Single Electricity Market
We examined this question in earlier papers (Jamasb and Pollitt 2005; Pollitt 2009). In what follows we review more recent work. We examine economic impact under a number of headings: prices, costs, return on capital, and fuel-related income distribution; quality of service; impact on the environment (via renewables and decarbonisation); and the impact on innovation. The first three can be summarized as: affordability, security of supply and the environment (the so called ‘energy trilemma’). The fourth is about dynamic efficiency. We conclude this section by revisiting overall assessments.
Prices, Costs, Rate of Return on Capital, and Fuel-Related Income Distribution
Has the single market reduced average prices (relative to business as usual) and price dispersion as we would expect from theory—especially in the wholesale power market?
Measuring the average price effect is hard because of the confounding impact of fossil fuel price fluctuations, renewables support policy, and network expenditure. There is some limited evidence of liberalization’s reducing household electricity prices, based on data for 23 EU countries from 2000 to 2014 (da Silva and Cerqueira 2017). This is in line with earlier studies that suggest limited price effects due to the introduction of a wholesale power market and the unbundling of transmission and generation.
There is more recent support for price convergence. The EU does measure the relative standard deviation of retail electricity prices in the EU member states. Industrial prices do show a degree of convergence over the period 2014–2018 (see European Commission 2017b, 2018, p. 29). Market coupling does appear to be associated with strong price correlations. Ouraichi and Spataru (2015) find high price correlations between national markets within four regional markets (CWE, CEE, MIBEL, and NordPool)Footnote 38 and high price correlations between countries in each of two different regional markets. They conclude that the flow-based market coupling policy of the EU has had promising—if incomplete—results; for instance, CWE (Belgium, France, Germany, Luxembourg, and the Netherlands) countries exhibit only 74% correlation within the region. Using a different measure (fractional cointegration analysis) of price convergence, de Menezes and Houllier (2016) also find evidence of price convergence between European electricity markets.
We would expect that prices and costs would move in parallel: decreases in prices likely imply decreases in costs. We can observe prices; but costs are more difficult to observe. They can be observed only indirectly: by examining returns on equity as well as prices. If returns stay constant, costs and prices move together; if returns fall but prices stay the same, then costs must be rising—though the cost of capital may be falling—and vice versa. We would expect liberalization to reduce overall returns to capital in the electricity sector; but it is possible that the profit-maximizing firms are good at price discriminating across their customer base, or at bargaining with their suppliers.
Jamasb and Pollitt (2005) produced some early evidence of falling returns in the electricity sector following liberalization, as well as falling costs. More recent studies that examine returns find mixed evidence of any change in underlying profitability.
Mergers in the electricity sector have created more value for European acquirers over the period 1998–2013 (Kishimoto et al. 2017). This would seem to indicate that the benefits of the single market are being brought about by the takeover of inefficient domestic firms by more efficient firms, and is consistent with the ability of mergers to create larger firms with greater economies of scale and scope. OECD (2016, chapter 5) shows a declining trend in clean energy returns on equity (relative to the cost of equity) in Europe over the period 2004–2015; this might be indicative of tougher conditions for renewables procurement—though this is not necessarily directly related to liberalization.
Tulloch et al. (2018) find declining returns for European electricity and gas utilities over the period 1996–2013, with a larger effect from the 2003 directives. They suggest that their results are consistent with evidence that the markets are becoming more competitive, albeit with some offsetting increase in risk. CEER (2017) find continuing wide variation in the target rates of return (WACC) used in electricity distribution and transmission, indicating a lack of standardization of network investment risks across Europe.
Since the beginning of the reform period fuel-related income distribution (‘fuel poverty’) has emerged as a big issue in some EU countries: notably, the UK. Conventionally defined as involving those spending 10% or more of their household expenditure on electricity and gas, this issue has become a focus of government policy on energy pricing. There is little evidence that this has systematically increased since 1996; but there is no doubt that energy remains a significant expenditure item for many households across the EU.Footnote 39
Quality of Service
The willingness to pay for continuity of service—avoiding outages—in electricity is very high: typically, of the order of 50–100 times more than the value of the retailed price of electricity.Footnote 40 This suggests that small improvements in the quality of service can be very valuable. For instance, the avoidance of an outage of just half an hour in customer minutes could be worth as much as 0.5% of the total value of electricity sold in the year.Footnote 41 Most interruptions to customer supply are local, due to problems at local substations. Regulators are also concerned about transmission system availability as this may be an indication of more serious failings in the transmission grid and the risk of wide-area blackouts.
CEER has been monitoring the quality of service at the EU level (see CEER 2018). There does seem to be a general decrease in customer minutes that were lost over the period 2002–2016, based on data that were reported to them. Measuring this is not straightforward, as it depends on how interruptions are monitored and reported—with different national definitions of what constitutes an interruption and different qualities of national reporting systems—and on exogenous drivers of interruption, such as extreme weather events.
Transmission system reliability also seems to have improved for several countries. One measure is the average interruption time (AIT). However, this measure started from very low levels: Spain improved from 2.006 min in 2002 to 0.14 min in 2016 (CEER 2016, p. 46; 2018 p. 36).
One quality issue is whether increased power flows between countries and system operator areas increases the risk of wider area blackouts. This is because one system operator has full visibility on the scheduling of generators and loads only within its control area, while interconnectors in an alternating current (AC) power system are subject to unscheduled loop flows as power that is transacted between areas A and B, may actually travel via a third area (from A to C to B, without area C having sight of the transaction between A and B).Footnote 42 This situation is especially difficult to manage if there is an unplanned transmission line outage that affects flows through C.
This problem is examined in Bialek (2004), who reviews the 2003 blackouts that involved six wide area blackouts in 6 weeks that affected 112 million people: five of the blackouts were in Europe; the sixth covered a large part of the Northeast of the US, including New York City. These included a blackout involving large parts of Denmark/Sweden and the whole of Italy—which was the largest blackout in Europe since 1945.
Bialek reviews the New York, Denmark/Sweden, and Italy incidents and concludes that while increased cross-border trading did not directly cause the blackouts it did necessitate an upgrading of inter-system operator communication: In both the Italy case and the similar case in New York, outdated communications infrastructure contributed to the slow communication of evolving problems in one system, which ultimately had serious consequences for interconnected systems. There have been other internationally power outage incidents, such as on the 4th November 2006 incident when a power line in Northern Germany had to be switched off to let a ship pass underneath it and caused a blackout in several European countries (see UCTE 2007).
Europe has made significant efforts to improve inter-transmission system operator (TSO) coordination to reduce these sorts of problems. The association of European TSOs (ENTSO-E) has overseen the establishment of regional security coordinators (RSCs), which are made up of groups of national TSOs (see ENTSO-E 2017a, b). These RSCs undertake regional operational security coordination, regional outage coordination, coordinated capacity calculation, and capacity adequacy assessments and provide dynamic asset information. The largest of these RSCs is CORESO, which began in 2008 and covers a population of 279 m [which is a private joint venture of REE (Spain), Elia (Belgium), RTE (France), National Grid (UK), Terna (Italy), 50 Hertz (Germany), and REN (Portugal)]. These RSCs are designed to reduce the risks that led to the international incidents in 2003 and 2006.
The European electricity system has experienced a profound transition with respect to its environmental impact. Between 2004 and 2016, the share of electricity generation from renewable sources has doubled from 14.3 to 29.6%. Hydro-electricity has barely changed in absolute quantity over this period. The increase is almost entirely due to an increase in wind, solar, biofuels, and other renewables; total generation has fallen by around 1%.Footnote 43
The emissions intensity and energy efficiency of EU electricity has also improved significantly.Footnote 44 The carbon emissions intensity was 431 g of CO2/kW h in 1990; 371 g/kW h in 1999; and 276 g/kW h in 2014—a decline of 36% between 1990 and 2014. The increased share of renewables and the increased use of more efficient gas-fired power generation and reduced use of coal in electricity generation has contributed to average power plant efficiency’s increasing from 36% in 1990 to 44% in 2014.Footnote 45
Jamasb and Pollitt (2005) noted impressive improvements in emissions of sulphur and nitrous oxides from the power sector in the period from 1990 up to 2003. Between 2004 and 2015, emissions have fallen by 77% for sulphur dioxide; 49% for nitrous oxides; and 81% for particulates.Footnote 46
The econometric evidence—presented in Asane-Otoo (2016)—suggests that for individual countries CO2, NOX, and SO2 intensity all decline with more liberalization (as measured by PMR—Product Market Regulation—indicators for private ownership, vertical integration and market entry), though not always significantly, over the period 1990–2012. Other econometric evidence from Vona and Nicolli (2014) suggests that countries with greater amounts of liberalization in their electricity sector have higher penetration of wind and solar power. Two channels for these effects might be: First, greater reliance on market forces to guide generation investment resulted in faster coal to gas substitution in the EU (and across the non-Europe OECD); and, second, that the initial efficiency gains from liberalization (discussed in Pollitt 2009) were partly spent on supporting renewables.
What is certainly true is that a combination of market liberalization (which initially favoured gas over coal), carbon pricing (which also favours gas over coal), the Large Combustion Plant DirectiveFootnote 47 (which forced higher emissions standards on existing and new coal fired power plants and raised their costs), and renewables support policies (which have led to renewables’ increasingly reducing wholesale electricity prices for both coal- and gas-fired power plants) have largely ended the building of new coal-fired power plants in the EU (see Caldecott et al. 2017).
Impact on Innovation
It is possible that over such a long period (1999–2018) the EU single market in electricity may have produced some positive short-run effects that mask negative long-run effects: Static gains may be offset by dynamic losses. One channel for this might be the impact of liberalization on the quantity of investment in research and development (R&D), or indeed on the productivity of R&D.
This is very difficult to measure because the rate of technological progress in electricity is slow and significantly affected by environmental regulation (and other types of regulation, such as nuclear safety).Footnote 48 However Jamasb and Pollitt (2008) discuss how one would expect liberalization to reduce R&D expenditure by electricity companies in theory: e.g., due to reductions in size and incentives to invest in R&D; they subsequently showed how this effect has played out in the UK (Jamasb and Pollitt 2011, 2015) and evidence of similar effects across the world—notably in the US.
One significant issue is the extent to which declines in electricity company R&D actually reduce aggregate energy R&D and the productivity of energy R&D expenditure. R&D may shift to the supply chain and away from former monopoly incumbent utilities. Marino et al. (2017) find that the deregulation of electricity initially increases but then reduces electricity patents across 31 OECD countries over the period 1985 to 2010. They conclude that this is evidence for an inverted-U relationship between the strength of liberalization and patenting.
The general equilibrium effect of the release of R&D resources to other sectors and the specific impact of higher expenditure on renewables linked to reform makes it very difficult to say what the aggregate dynamic effect of European electricity liberalization might be—even if aggregate energy R&D has declined.
Overall Assessments of the Impact of the Single Electricity Market
It is striking how little large-scale evidence has been produced for the overall impacts of the single market in electricity. It is even more striking that the most convincing evidence is simulated. Simulation studies can show the benefits of market coupling and changing the pricing behavior of incumbent players. These studies find that reducing market power in pricing can have substantial benefits: e.g., Hobbs et al. (2005) who examine Netherlands-Belgium market coupling. A move from strategic interaction to perfect competition can be shown to yield big impacts, reducing the profits of incumbents significantly—Lise et al. (2006) show the profits of EdF and Tractabel falling by a third—if there is enough interconnection between countries. However, these theoretical gains from more perfect competition may not be realizable in a second-best word.
As to how actual overall reform effects can be assessed: There are a number of ways to assess the overall impact of a liberalization package, such as that represented by the single electricity market project (as noted in Pollitt 2012). These include: performance metric regressions on panel data (e.g. Steiner 2001); simple statistical tests of before and after performance using t tests (e.g., following D’Souza and Megginson 1999); social cost–benefit analyses of reform (following Jones et al. 1990) with reform as the public investment project; and macro studies of reform that attempt to find effects with the use of general equilibrium models of the economy (e.g. Chisari et al. 1999 and Copenhagen Economics 2005).
Remarkably, not all of these methods have been applied clearly to the EU single market project, as opposed to electricity reform in the OECD more generally. Fiorio et al. (2007) attempts panel data regression on reform variables to show evidence of modest productivity improvements—but ambiguous price impacts—for 15 EU countries.
Reform evidence—both for samples of countries and for specific case studies—seems to support the view (in Pollitt 2012) that market liberalisation reduces costs somewhat but may not affect prices. This positive view on productivity has recently been disputed by Polemis and Stengos (2017), who suggest that for their sample of OECD countries over the period 1975–2013 electricity liberalisation does not increase their measures of productivity—which include generation per capita and labour productivity—in already liberalised countries. However a standard problem with panel data analysis is that reform is a package and identification of its effects is a problem, given that multiple significant policy impacts happen simultaneously: for instance, cost-increasing environmental policy and cost-reducing market reform.
What is more clear is that public dissatisfaction with liberalised energy markets in Europe remains strong. Fiorio and Florio (2011) showed that in Europe private ownership of electricity assets was correlated with increased public dissatisfaction with the industry. This dissatisfaction explains the remaining significant public ownership in European electricity—in spite of the single electricity market—and the fact that in at least half of EU countries residential price controls still exist for some household customers: something that ACER continues to condemn as limiting the impact of the single market project.Footnote 49