Market Mechanisms and Technology Transition in Australia’s National Electricity Market

To investigate the interaction between restructured electricity market designs, in particular Australia’s National Electricity Market (NEM), and variable renewable generators. In particular, to examine how the expansion of near-zero marginal cost, variable renewable generators may impact participant incentives and market power. The technology transition may significantly impact bidding behaviour in electricity markets such as the NEM. This paper identifies and explores key challenges including the impact of renewables on electricity market power, incentive structures for variable renewables, potential for exercise of market power in the Australia’s electricity market as well as contracting and bidding strategies for variable renewables.


Introduction
The recent implementation of electricity market mechanisms represents a step in a series of economic and technical transitions that have occurred over numerous decades across a wide number of regulatory environments.
Certainly for Australia, the move from a centralised, government-owned electricity system to a corporatised electricity market with more decentralised decision-making, was not undertaken out of necessity or crisis. It was part of a wider set of reforms designed to tackle what was seen as bureaucratic excess in some key state-owned monopoly infrastructure providers, and re-implement core aspects of Australia's economy around a market-based ideology [1,2].
Australia's electricity sector transitioned in the second half of the twentieth century from a wide assortment of privately and municipally owned piecemeal systems, to a small number of large state-based centrally managed power systems. This centralisation began in earnest in the postwar period, when state-based electricity commissions were tasked with the electrification of homes and industry [3]. These commissions constructed a web of electricity transmission and distribution systems that would eventually, when joined together, become arguably the longest electricity network in the world, owing to the vast distances travelled by rural power lines. The Snowy Mountains hydroelectric scheme became a key symbol of Australia's electrification and its associated economic progress, with the construction of 16 dams and 3.7 GW of generation capacity in the Australian alps [4]. This system was synchronised with regional thermal generators in the large states of NSW and Victoria, leveraging local black and brown coal reserves respectively. A centralised planning and decision-making system within each of the state-based electricity commissions oversaw both investment decisions and day-to-day operation of the power system [5]. Despite widespread, significantly crosssubsidised rural electrification, electricity prices in Australia 1 3 experienced a general downward trend from 1955 to 1995. Australian electricity was among the cheapest of the OECD countries [3].
There were however a number of issues with Australia's state-run electricity commissions that ultimately lead to their demise. The majority of them undertook enormous capital expansion in the 1980s, based on demand forecasts that did not eventuate; this construction-spree precipitated 4 years of electricity price rises and eroded public faith in the ability of the commissions to manage the responsibility of effective investment [3,6]. The electricity commissions accumulated significant levels of debt due to their expansion; by 1990 this debt represented more than that of all Australian private companies combined [7]. The Industry Commission [6] report into Energy Generation and Distribution, which recommended the urgent restructuring of Australia's electricity sector, summarised: "Poor investment decisions leading to excess capacity and gross overstaffing during the 1980s provide the most striking evidence that electricity and gas have not been supplied at least cost." The report estimated that inefficiencies in the sector were costing Australians $2.4 billion per year (in 1991 dollars).
This wave of poor bureaucratic decision-making coincided with an international ideological movement for greater competition and less government involvement in economic life. It is likely no coincidence that Australia's electricity restructuring coincided with a wave of similar efforts in other countries. Proponents of this flavour of reform prioritised 'efficiency', emphasising the benefits of a reduction in bureaucratic excess and lower reliance on a unionised public sector workforce. These ideologies are summarised neatly by Chester [7], "competition is a primary virtue; markets should determine the allocation of resources; industry sectors dominated by the state should be 'turned over' to private interests." Globally, economic liberalism in this period appears to have expanded to domains previously thought of as exclusive to government.
A stumbling, indebted bureaucracy and the emergence of a global neoliberal zeitgeist appear to have culminated in the political will for reform. The National Competition Policy [1] established the theoretical basis for the restructuring of large swathes of the Australian economy that were traditionally thought to be natural monopolies that should be administered by government agencies. The policy's theoretical basis was neatly summarised by Outhred [8]: "Competition is preferable to monopoly"; "Negotiation is preferred to direct regulatory intervention"; "When regulation is required, it should act as a proxy to competition at arm's length from government with the objectives of economic efficiency, regulatory efficiency and revenue sufficiency." The National Competition Policy paper sparked a number of significant industrial corporatisation efforts in a range of broad domains such as energy, telecommunications, post, air travel, rail, roads and ports, commonly known as the Hilmer reforms [9,10].

The Principles of Electricity Restructuring
A number of significant energy policy initiatives were undertaken in the mid-1990s as part of the Hilmer reforms, often referred to as Australia's electricity 'restructuring.' Restructuring is a convenient blanket term that simultaneously encapsulates some measure of 'liberalisation', 'corporatisation' and 'privatisation' [11].
The stated intention of restructuring Australia's electricity sector was to provide a more efficient investment and operational decision-making mechanism than the state-owned vertically integrated monopolies that previously undertook these roles [12]. As Roarty [13] wrote during the implementation of these policies, "the commonly stated rationale for restructuring is to increase microeconomic efficiency through the introduction of competition in regulated industries, such as telecommunications, banking and energy. The main objectives are to ensure that customers receive: choice of supplier; lower cost of service; and increased range of services." The state electricity commissions were, until the 1990s, vertically integrated entities that managed all aspects of electricity generation and delivery. It was recognised that while transmission and distribution networks may be natural monopolies, generation and retail functions do not necessarily fit this definition [7]. Restructuring therefore involved a functional separation of the roles of transmission, distribution, generation and retail [8]. Generation and retail were 'unbundled' from natural monopoly networks and each entity given separate regulatory arrangements. This began a process of corporatisation and subsequent staged privatisation of electricity assets [8].
Transmission and distribution, still seen as natural monopolies, were corporatized, given internal profit incentives, and subjected to cost-plus regulation [13]. Networks were effectively re-imagined as 'platforms' or service providers upon which other entities may run private generation and electricity retail businesses.
State electricity commissions had historically coordinated the dispatch of generators to meet regional electricity demand at any given time. Instead, a decentralised wholesale market for electricity was created, wherein all output from large grid-connected generators must be sold [14]. It was expected that participants would be incentivised to scale up or down generation based on price signals from the market. Because all revenue would be based around (if not flowing directly from) the wholesale market, as the physical market underlying all contracts, it was also hoped that efficient pricing would uncover opportunities to meet future demand, encouraging entrepreneurship, innovation and new generation investment into the electricity system as it was needed. The wholesale market was therefore expected to provide efficient short and long-term decision-making and reduce barriers to entry [8].
A process of interconnecting previously isolated electricity systems was undertaken, both to increase system security and enable regional trade with the introduction of the National Electricity Market (NEM) [15]. Australia's energy reforms took place coincident with the development of these interconnectors; links between Queensland and NSW were put in place between 2000 and 2002, Murraylink, a second, crucial interconnector between Victoria and South Australia opened in 2002, and the Basslink connection between Victoria and Tasmania came online in 2006 [15].
Additionally, customer services were decentralised over time with the creation of a competitive retail market [8]. Newly created electricity retail businesses were expected to be the interface between consumers and the electricity industry, bundling electricity purchased from the wholesale market with a range of services delivered by transmission and distribution networks [16].
Together, these reforms: the corporatisation of electricity networks, the creation of a wholesale market for electricity and the emergence of a retail bundling market, represented a complete overhaul of Australia's electricity industry.

The National Electricity Objective
Private companies have private interests that are not necessarily in line with the wellbeing of the state or its inhabitants. It was therefore considered important that regulators and policymakers not lose sight of the goal of delivering electricity to citizens as an essential public good. To this end, accompanying Australia's electricity restructuring reforms is the National Electricity Objective (NEO), which is intended to set the goals of market designers, regulators and operators [17].
The National Electricity Objective as stated in the National Electricity Law [18] since 2004, is "to promote efficient investment in, and efficient operation and use of, electricity services for the long term interests of consumers of electricity with respect to…price, quality, safety, reliability and security of supply of electricity".
The key mechanism for implementing this objective is the NEM, in particular the wholesale market mechanism, in which the majority of operational and investment decisionmaking is expected to occur [8]. This highlights the incredible importance of market design in the overall ability of Australia's electricity system to meet its primary objective; without a functioning market, poor decision-making may lead to an electricity system that cannot reach these goals.
It is noted that the NEO addresses only two of the three key points of the commonly cited 'energy trilemma' of reliability, affordability and sustainability [19]. In the NEM, secondary mechanisms such as the mandatory renewable energy target [20] have been employed to address sustainability issues within Australia's electricity supply.

Decision-Making and the Market Mechanism (Design of the NEM)
The restructuring of Australia's electricity industry represented a significant divestment of decision-making power from central planners. In today's market, centralised demand forecasting still takes place; it is no longer performed by state electricity commissions, but rather by the Australian Energy Market Operator (AEMO), which must aim to incentivise investment in new generation capacity, rather than directing its commencement [21]. These investment decisions are abstracted to vertically integrated gentailers and a futures market, based on medium-term (approximately 5-year) time horizon pricing [22]. Centralised network investment planning is also performed by the market operator via the Integrated System Plan [21], but in this case too, planners no longer have decision-making power; transmission and distribution investment usually relies on proposals from the existing regulated monopolies, with cost-benefit oversight from the regulator, and in some cases recently, government intervention to support investment in or approval of specific strategic infrastructure.
While the NEM is a complex set of arrangements, designed to manage the preferences of many different stakeholders, the central decision-making mechanism for Australia's generation fleet is the wholesale market [23]. This market is where the most important operational decisions are made, based on demand and available supply. Additionally, price signals from the spot market are expected to be the key drivers of investment in new plants and equipment. Effective functioning of the wholesale market mechanism is necessary for the NEM to satisfy the NEO.
The underlying principle of the wholesale market is that by operating an auction system, generators will be encouraged, through competition, to reveal details in their bids about the underlying costs of running their power plants [24]. Using this information, the least-cost set of generating units can be chosen at any given time. This would mean that the most 'efficient' set of operational choices are made in each trading period. By setting prices based on supply and demand, scarcity price signals from this market may incentivise investment in new plant and equipment to satisfy shortfalls. This means that when supply is constrained for extended periods of time, market prices could rise, signalling the need for new supply and incentivising its entry (either by altered operation or new capital build). The NEM's auction mechanism is an abstraction layer -an economic mechanism to solve a technical problem that was once solved by a central controller.
The NEM requires each generator, on a day-ahead basis, to submit a series of price-volume pairs for each trading period [25]. These price-volume pairs are expected to be a piecewise linear approximation of the generator's marginal cost curve [24]. For example, a gas-fired generator may have a narrow efficient operating window at which fuel costs are relatively low per unit of output; operation outside of this window may have higher associated costs. This generator would split up its output into a number of blocks that may have higher or lower cost. In the NEM, generators are given the opportunity to offer portions of their capacity in nine different price bands [23]. They are able to change the volume of energy submitted at each price, up until a given trading period is executed.
The simple conception of an electricity market is that once all bids are submitted, the operator conducts economic dispatch, selecting the cheapest possible set of bids that satisfy demand [26]. In reality, the situation is rather more complex; the market operator instead runs a simulation of all generators in the NEM, solving a vast number of constraint equations to ensure stability and security of supply by co-optimising frequency and wholesale markets [27]. This simulation finds the cheapest set of generators that are able to satisfy these constraints based on their bids, and these generators are chosen in what is commonly referred to as 'security-constrained dispatch'. The most expensive unit of generation dispatched is considered the 'marginal' unit, and all participants are paid this unit's bid price.
Subsidiary markets for ancillary services, such as frequency control, are also run on a similar basis [25].
Investment signals are expected to arise from long-term pricing trends in the wholesale market. For market based dispatch, pricing theory suggests that efficient operation requires both efficient spot prices (based on short run marginal costs) and future prices (based on opportunity costs associated with the impact of inter-temporal links between past, present and future decisions) [28]. An active market for electricity futures is in operation, and a number of financial derivatives based on the wholesale spot market are currently traded [29]. There are however a number of issues around the efficiency and effectiveness of these markets. The 'missing money' problem is one such issue, whereby spot market revenues have not been sufficient to adequately account for capital costs of new builds [30,31]. This issue, and how generators might achieve revenue sufficiency, is further discussed in the "Renewables, Incentives and Electricity Market Power" section.
These markets form the basis of investment and operational decision-making in the NEM, and the mechanism by which the NEM is expected to fulfil its obligation to the NEO. Not only do these markets need to provide appropriate signals for operation of and investment in expanding existing fleets and incorporate their preferences based on the economics of fossil fuelled generators, but they must also be able to support new technologies, which may not have been considered in the original market design.

Technology Transition and the NEM
Despite a fraught political landscape around energy and carbon emissions, Australia is emerging as a world leader in the rate of renewable energy deployment. At current growth rates, some researchers suggest that that the NEM may reach 50% renewable penetration (by energy) in the mid-2020s and up to 100% in the early 2030s [32]. This growth is expected to be comprised of a range of technologies such as rooftop PV, ground-mounted PV, wind, batteries and pumped hydro.
There is of course an additional urgency underlying rapid decarbonisation, based on scientific consensus around climate change [33] and its impacts; this may demand technology change regardless of domestic energy economics.
A number of high or 100% renewable electricity scenarios for Australia have been modelled in academia [34][35][36], as well as by the Australian Energy Market Operator [37], using a range of different technology cost assumptions and demand profiles. Even more ambitious scenarios have also been proposed; Australia's National Hydrogen Strategy, for example, suggests up to 700 GW of wind and solar could be built to support future industry; this is approximately 700% of Australia's current generation capacity [38]. High-voltage submarine DC interconnectors may also provide opportunities for expanding renewable capacity above the levels of Australia's domestic consumption [39].
Interestingly, some models have found that relatively high penetrations of renewable electricity can be supported without the need for significant levels of storage. One study has found that that renewable penetrations of over 50% could be integrated with relatively few additional storage and firming costs [36].
These examples are given to illustrate a key point: the transition to high penetration renewable electricity is a highly likely future outcome that therefore must be considered in detail in the context of the NEM. This transition will require critical investment decisions of enormous scale. The capital requirements of a transition to 100% renewable energy were recently estimated to have a price tag of over $100 billion AUD [36]; this is significantly less than AEMO's estimates from 2013, which ranged between $219 and $332 billion. The capital costs of renewable energy technologies are continuing to fall as they move across their respective learning curves, meaning that there is a chance these costs have been overestimated. Regardless, the scale of this investment is equivalent to a significant proportion (perhaps 5-10%) of Australia's GDP. These decisions will have significant additional flow-on effects on the Australian economy because electricity costs impact the net productivity of a large fraction of Australian industry.

Is the NEM up to the Task of Transition?
Many key decisions associated with this transition, with a scale of investment in the tens to hundreds of billions of dollars, will happen in the context of the NEM. If poor economic decisions are made over the coming decade, the transition to renewable electricity may be slowed, the cost of electricity may be unduly increased, and energy security may be weakened. The NEM's market mechanism will be the focal point of these decisions. It is the primary mechanism by which investment signals are expected to propagate to potential developers of generation assets and the key measure of benefits associated with regulated transmission network investments.
Additionally, the NEM's wholesale mechanism will continue to carry the weight of short-term operational decisions throughout this transition, choosing which generators are dispatched and which are not, and setting prices that are intended to incentivise efficient operation of generation assets. The NEM's market mechanism, which was designed at a time when effectively no renewable energy was present (or even expected) in the Australian grid, may shortly be required to support a grid comprised largely of variable renewable energy and storage. Troublingly, there are currently few precedents for the technical management of high-penetration renewable electricity, especially via market mechanisms, because such grids do not yet commonly exist, and there are as yet a myriad of unanswered questions around the adequacy of the mechanism to fulfil this task. The "Principles of a Functioning Market Mechanism" section provides an overview of the objectives and design principles of market mechanisms, while the "Emerging Challenges with Variable Renewable Integration" section reviews current evidence of their suitability to majority-renewable grids, and highlights key gaps in our understanding.

Principles of a Functioning Market Mechanism
The NEM's market mechanism can be viewed as having one primary goal: to keep electricity prices low for consumers. This simplification is somewhat valid because the rest of the NEM's goals, as set by the NEO [18]: quality, safety, reliability and security of supply, are handled largely by parallel ancillary services markets and the resulting technical constraints incorporated into security constrained economic dispatch, simulating the lowest-cost set of generators that are able to meet reliability and security requirements [40].
On the assumption that this can in fact be achieved, the goals of a market mechanism can be isolated to two key principles: competition must be available to erode margins in the short-term, and the market must be open to competition from new participants in the long-term. There may be a degree of trade-off between these principles, as short-term margins may be required to incentivise new market entrants [41]. But in the long-term, it appears that the optimal consumer outcome would be one where the market delivers efficient pricing, and simultaneously incentivises the timely entrance of new generation assets to meet future demand [26].
Efficient and transparent price signals are key. In the short-term, if prices do not respond adequately to demand so that generators can cover their operating costs, then generators will not be available for dispatch when needed [42]. In the long-term, investments will be made by private entities only if there is a reasonable expectation of risk-weighted profit; transparent and efficient pricing would seem to be a prerequisite of such an expectation. Ideally, consumers would also be able to dynamically express their willingness to pay for electricity and be able to see and respond to price signals, for example by curtailing load and operating DERs [43].
For price signals to work, effective competition is required. Without competition, participants are able to exercise market power, gaining the ability to extract rents from the marketplace. In the short-term, reducing the opportunity for exercise of market power means ensuring that the market mechanism is difficult to 'game' by participants [44]; firm size and diversity is considered crucial in this pursuit. In the long-term, it means lowering barriers to entry so that new investment can erode accumulated margins [45].

Emerging Challenges with Variable Renewable Integration
The primacy of competition and transparent pricing in the effective operation of an electricity market to minimise prices for consumers within reliability and security constraints provides a lens through which to examine the ongoing transition to renewable energy. There are a number of characteristics of variable renewable energy that are not shared by the fossil fuelled generators for which the electricity market was designed. These include near-zero marginal costs of generation, variable and stochastic supply, in the case of wind and PV a lack of ability to trade across inter-temporal links, and widespread third-party power purchase agreement (PPA) contracting. These characteristics may have a significant impact on the ability of a majorityrenewables market to continue to function as designed.
There is a significant ongoing debate in the literature around whether or not near-zero marginal cost renewables present significant challenges to the function of electricity market mechanisms. Some theorists, for example Hogan [46], Leslie et al. [47], Cramton [26] and Simshauser [48] argue that energy-only market design is fundamentally technology independent and that markets are able to function with low short-run marginal costs provided that adequate mechanisms for efficient scarcity pricing are provided. Others, such as Bunn, Yusupov [49,50] and Johnson, Oliver [51] argue that there may be fundamental issues in transitioning to near-zero marginal cost renewables while still attempting to utilise market mechanisms that were designed for dispatchable, high-marginal-cost generators, in particular relating to the stochastic nature of generator output, inability to price around inter-temporal links, negative pricing and market volatility.
Additionally, investment price signals have been introduced for renewable generators that sit outside of the NEM's central set of rule-based wholesale and ancillary service incentive mechanisms, the most significant of which is the mandatory renewable energy target [20] which provides a renewable energy certificate (REC) based side payment mechanism to incentivise investment in new generation projects. These certificates are based only on energy production, meaning that investment incentives may have been introduced into the system that do not fully account for the supply-demand balance in the long-term, or take into account energy security or reliability requirements. While certificate pricing within this system is now largely market-driven (via demand on the retail side) there are many possible inefficiencies and energy security or reliability issues that may emerge, including from the use of a national REC system versus zonal NEM pricing, as well as the obfuscation of energy security-based price signals in ancillary markets that may inform efficient capacity expansion decision-making.
There are also a range of fundamental market issues that may be raised in the transition to high-penetration, nearzero marginal cost renewables. One such issue is that price volatility may increase as scarcity pricing becomes more pronounced in the absence of dispatchable generation, and such volatility may be undesirable to a range of stakeholders, even if such price movements are 'economically efficient' from the perspective of market theorists. Joskow [52] highlights the importance of wholesale prices being reflective of marginal costs unless under scarcity pricing; the identification of true scarcity and therefore legitimate peak pricing, versus unjustified exercises of market power, may become simultaneously more pertinent and more difficult.
Another issue is whether the social choice function implemented by electricity markets can adequately dispatch from among a range of stochastic generators with high fixed costs and near-zero marginal costs.
Further concerns remain around whether the equality of outcomes when bidding near zero marginal cost (i.e. price indifference) and the lack of individual generator dispatchability (i.e. the inability to delay dispatch until times of high demand) may fundamentally alter the dynamics of electricity markets in ways that disadvantage both producers and consumers.
From the outset, this paper does not take a view on whether zero marginal cost renewables do or do not present significant challenges, but rather aims to examine this conundrum in the context of Australia's ongoing transition to renewable energy. As Hogan [46] writes, "the interesting questions are how can economically efficient renewables alter the fundamentals and how will this dictate changes in the basis of electricity market design?" This paper aims to provide some additional clarity with regard to these questions by exploring zero marginal cost renewables in the context of restructured electricity markets.

Impact of High-Penetration Renewables on Electricity Markets
Variable renewable generators, by nature of their respective technologies, tend to have very low or near-zero shortrun marginal costs; the Australian Energy Market Operator (AEMO) estimates new entrant short-run marginal costs for PV, batteries and pumped hydro at $0/MWh, solar thermal at a maximum of $7.22/MWh and wind at a maximum of $3.66/MWh [53]. This is in stark contrast to fossil-fuelled generators, which have substantial fuel and O&M costs that scale with the level of the generator's output at any given time. The NEM's market mechanism was designed with the expectation that a generator's output could be broken down into a monotonically increasing piecewise marginal benefit curve [24]; much of the theoretical basis for economically optimal market decision-making hinges on this expectation [26,42]. Variable renewable generators may not neatly fit this model because in addition to having near-zero marginal costs, they are likely to feature a flat (and inflexible) marginal benefit curve [54]. This might mean that these generators, by the nature of their technology and internal economic levers, are not incentivised to participate in the market mechanism in the same manner as other generators [55]. Additionally, markets with high proportions of variable renewable energy may be subject to frequent scarcity pricing and alternating periods of oversupply, due to variability of the underlying resources. This may require regulatory attention and the recalibration of ceiling prices to incentivise efficient investment decisions [47]. Understanding the impact of near-zero marginal costs and flat marginal benefit curves on the underlying theory of the NEM's market mechanism will be key to ensuring that the mechanism continues to incentivise competition and provide efficient price signals as more variable renewables are brought online. However, this concept appears to represent a significant gap in the literature, specifically with regard to how high-penetration renewable energy mixes might impact trends competition in restructured markets such as the NEM.

Renewables, Incentives and Electricity Market Power
Market power is a significant barrier to the efficient operation of electricity markets [44]. One way for generators to increase revenues within the wholesale market is to choose not to bid their short-run marginal cost, but another value altogether based on their expectation that they may be able to set prices. Some theorists, for example Hogan [41] or Moran, Sood [56], argue that this as a necessary and efficient feature of electricity markets. Others such as Borenstein et al. [45] appear to view market power as problematic. From a theoretical perspective, market power may be a fundamental issue that challenges the very concept of an efficient market mechanism where market operators do not have access to accurate information about the underlying costs of generation, and inefficient price signals form the basis for dispatch and investment [44].
Most analyses concerning renewable energy and market power, such as Pham [57] investigating the French electricity market, Sirin, Yilmaz [58] in the Turkish market, Ciarreta et al. [59] in the Spanish market, and Bell et al. [60] in the NEM, focus on evidence of the merit order impacts of abundant variable renewables mitigating market power and placing downward pressure on spot prices. These works generally do not view renewable generators as strategic players; rather, they assume that by their presence they mitigate the ability of others to act in an anti-competitive manner.
One interesting result was the work of Acemoglu et al. [61] which theorised that merit order effects from wind power may be overcompensated by existing generator withholding, while if single firms owned both sets of assets, withholding incentives were mitigated and overall consumer outcomes improved; this produced the surprising result that market diversification resulted in a decrease in welfare. This work is indicative of the insights that may come from further analysis of oligopolistic behaviour in the presence of renewables.
Very little attention has however been paid to the concept of renewable generators themselves exercising market power, either unilaterally or in a collusive manner, within electricity markets. Earlier works on carbon policy interactions with market power such as Zhou et al. [62] flagged a future need for such work given forecast increases in renewable penetration, though the issue appears to have garnered little academic attention since. One notable exception is the work of [63] which explored the role of market concentration in high-renewable market simulations. Another is the work of Ekholm, Virasjoki [64] in which a 100% renewable energy and storage market was simulated and competitive equilibria modelled. The exercise of generator market power and its incentivisation in high-penetration renewable markets should be of serious concern of market operators because these generators may in the near future comprise the majority of capacity in many markets. There exists a great deal of scope for further research into incentives, competitive dynamics and market impacts of renewables, especially in the Australian context, in particular around the strategic incentives acting upon generators in these future systems.

Electricity Market Power in the Australian NEM
In the early years of Australia's National electricity market, the concept of market power and effective competition was a subject of some research and regulatory interest. The Industry Commission [65] appears to have been the first to attempt to model Australian electricity market power, using equilibrium models to determine that there existed incentives for generators to exert influence over market prices in South Australia. Game theoretic modelling by Brennan, Melanie [66] showed a similar result in NSW, concluding that "despite the high level of excess capacity held by NSW generators, they are likely to be in a position to practise noncompetitive pricing in the initial phases of the market, particularly in peak periods." Melanie, Weston [67] expressed concern around a lack of competition among Victorian generators. Outhred [8] noted also concerns from the Queensland Electricity Industry Structure Taskforce around generator competition.
Despite early concerns, there do not appear to have been many significant academic attempts to monitor market power in Australia's electricity market in the early 2000s. Wolak [68] applied a cost function estimation technique to early market data, but this appears to have been more of a demonstration of the technique and results seemed inconclusive. Short, Swan [69] reportedly applied the commonly utilised Lerner index to market data but the original report appears to be lost. Outhred [70] reviewed Australia's market structure and performance, concluding that "market power has not been identified as a major problem in the Australian electricity industry to date." However, in a single-participant focused study, Gans et al. [71] showed evidence of market power exercise after generator consolidation in the NEM.
In the last decade, there has been increased interest in the exercise of market power due to some significant price events. Hesamzadeh et al. [72] applied a transmission-constrained pivotal supplier index and found some instances of potential for market power that correlated with spot price increases. By analysing bids, capacities and supply costs, Mountain [73] showed some evidence of withholding in the South Australian electricity market. An empirical analysis of rebidding behaviour (adjustment of offers after initial prices and volumes are submitted) by Clements et al. [74] showed evidence of market power in late rebidding or 'sniping' 1 3 strategies by generators in Queensland; evidence of strategic rebidding was further uncovered by Dungey et al. [75] across a range of NEM trading nodes. Mountain, Percy [76] investigated price trends following the closure of Hazelwood coal-fired power station and found increases consistent with the exercise of market power resulting from an increase in concentration. Recently, McConnell, Sandiford [77] used a price-cost margin technique to explore the exercise of market power in South Australia, as well as calculating pivotal supply indices on a monthly bases over a ten-year period.
Electricity market power is increasingly being taken seriously by regulators in the Australian context. Yearly reporting on market concentration is included in the market operator's 'State of the Energy Market' report [78], and six monthly market power reports will be generated by the Electricity Market Monitoring inquiry [79].
The piecemeal nature of the majority of studies into competitiveness and market power in the NEM appears to neglect a crucial factor in market theory: that equilibria may be disrupted on the short-term, and long-term trends are important with reference to the function of efficient markets. There appears to be a significant lack of long-term analysis of electricity market competition in the NEM which might observe trends over many years to determine if competition is generally present in the market.

Renewable Generator Contracting and Bidding Behaviour
To fully understand the interaction between renewable energy and market mechanisms, it is important to understand the impact of generator contracting on market behaviour. Renewable generators are often seen to enter into power purchase agreements with third parties [80]; these agreements are administered as a financial derivative on top of the NEM, and may significantly impact bidding behaviour [22,81]. This could mean that bidders are not exposed to price signals and therefore incentivised to bid in ways that are not consistent with the efficient operation of the market, or potentially their own efficient operation. If a large majority of generators in the NEM are effectively de-risked through power purchase agreements, the impact on market prices may be significant. While third-party contracting and bidding to maximise dispatch are not necessarily damaging, they should be understood given their potential for disruption. There does not appear to have been a significant review of renewable contracting behaviour in the NEM (nor the majority of other restructured electricity markets) despite its potential significance.
Variable renewable bidders are also unable to participate in the market in the way that was envisaged by market designers because they cannot delay dispatch to later trading periods with higher prices. Market theory suggests that participants in wholesale mechanisms should be able to take advantage of inter-temporal links between trading periods, delaying output to times of higher price (and thus demand) for efficient market outcomes [28]. Variable renewable generators by their very nature cannot do this. This means that time-dependent decision-making around available capacity for variable renewable generators is largely made at the investment stage (through location selection relative to wind or solar resources and demand centres). This seems to increase the likelihood of inefficient investment decisions and the construction of assets that may become effectively stranded. Operationally, the lack of dispatchability increases uncertainty in near-term availability and cost forecasting for all generators, and this uncertainty may be priced into generator offers across the market. Variability can make coordination with other generators, especially those with low ramp rates, more difficult. A lack of dispatchability can therefore make balancing of security, reliability and pricing constraints more pertinent, moving the market further away from economic dispatch toward more complex pricing outcomes that may lead to additional inefficiencies. This may mean that the promise of efficient or optimal outcomes for wholesale market mechanisms may not be as certain with high levels of variable renewable generation.
An understanding of bidding behaviour resulting from contractual, market-based and external policy incentives will therefore be crucial in determining the future performance of renewable energy in the NEM. Some early analysis of fossilfuelled generator bidding patterns has been undertaken in the NEM, for example by Mielczarski et al. [82] and Hu et al. [83], however such analysis does not appear to have been replicated for renewable generators.

Conclusion
This review has highlighted a number of characteristics of renewable energy generators that may mean that markets with high penetrations of renewable electricity do not function as designed. The paper began by examining the history of Australia's electricity industry and its decision-making structures, highlighting the fact that current mechanisms were put in place in the absence of any considerable renewable generation. It raised the questions of whether electricity markets would be fundamentally different if conceived with a forward-looking energy mix in mind, as well as whether existing mechanisms are up to the task of facilitating the expected transition in line with energy system goals such as security, reliability and cost of electricity. Additionally, this paper highlighted some fundamental gaps in the literature around the impact of renewables on existing market mechanisms, whether there may be a need for mechanism reform, the possibility of 1 3 the exercise of market power in the presence of variable renewables, and the real-world behaviour of renewable generators in restructured market mechanisms.
It is critical that policymakers understand whether the NEM and restructured electricity markets more generally can continue to facilitate efficient competition, as well as the ways in which they may fail to do so. Additionally, given the scale of the expected technology shift in the coming years, it is important to understand how the transition to high penetrations of renewable energy may impact the function of electricity markets, and the ways in which renewable generators may be incentivised to participate in these mechanisms.
Funding Open Access funding enabled and organized by CAUL and its Member Institutions.

Human and Animal Rights and Informed Consent Not applicable.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.