Efficient Renewable Electricity Support: Designing an Incentive-compatible Support Scheme.

AuthorNewbety, David

    Faced with an economy-wide net-zero carbon target by 2050, the electricity industry will have to reach near zero emissions far sooner. That requires a massive increase in variable renewable electricity (VRE, primarily wind and solar PV). (1) The UK expects a doubling of renewable electricity (RE) between 2019 and 2030, requiring a volume of new contracts equal to all past support schemes (National Grid, 2020). Delivering that investment at least cost will require a drastic redesign of support schemes and contracts. This article proposes an efficient contract that addresses past market and policy distortions.

    Existing support schemes reflect past compromises to reconcile often conflicting objectives and to disentangle past unintended consequences of faulty policies (Bunn and Yusupov, 2015; Klobasa et al., 2013; Nock and Baker, 2017). Thus the EU Emissions Trading Scheme fixed a cap on emissions, but the subsequent 20-20-20 Renewables Directive increased renewable targets without reducing the cap commensurately. The unintended result was the additional renewables had zero impact on EU emissions. Reforming VRE support design yet again might worry policy makers concerned about investor confidence. In fact, it would increase confidence to offer more efficient new contracts while honouring existing RE contracts. Efficient policies are more credible as there is no need to change them, reducing investor risk and increasing their willingness to invest.

    There has been a tension between accelerating investment in RE and providing unnecessarily generous payments that risk excessive public cost. Price support schemes like Feed-in-Tariffs (FiTs) that set the price and allow all entrants to claim these FiTs have often led to excessive public cost and rapid closure, or in some cases to retrospective withdrawal (e.g. in Spain, see CEER, 2018). Quantity-based schemes, such as green certificates, can place excessive risk on developers, leading either to under-delivery or over-compensation (Finon, 2006). The solution is simple but took surprisingly long to rediscover, given that the first UK RE support schemes in the 1990s involved auctions (Mitchell, 2000). Well-designed auctions can dramatically reduce the cost of procuring RE compared to administratively fixing the strike price. Newbery (2016a) showed that the first GB auction after the 2011 Energy Act dramatically reduced contract prices, while successive auctions for off-shore wind in the North Sea more than halved prices (Grubb and Newbery, 2018; CEER, 2020). The auction can either fix the volume or the funds available to deliver the least cost solution that meets the capacity target or fits the budget.

    This article proposes an efficient contract that can be auctioned to deliver least cost decarbonisation while maintaining control over the amount of support. The Clean Energy Package requirements provide good principles to guide the design of Renewable Electricity Support Schemes (RESS), but avoids drawing out the design implications. It stresses the role of markets, but that requires policy makers to identify and address the market failures facing decarbonisation. The next section identifies the market failures that justify intervention, Section 3 sets out the criteria for efficient support schemes. Section 4 lists the types of support schemes, briefly reviews the relevant literature and provides evidence on their prevalence. Section 5 identifies the distortions of existing schemes to highlight the ways in which they can be overcome. Section 6 then proposes a contract design that avoids these distortions and addresses the market failures. Section 7 concludes.


    The two main arguments for supporting VRE are that their deployment drives down future costs (their learning benefit) and they face risks (particularly policy and market redesign risks) that are hard to hedge with existing futures and contract markets. Future investment in flexible fossil back-up generation and storage also face increased future risks that will also require careful market design and contracts but that is left to be dealt with elsewhere. Competing fossil investment will be over-subsidized unless it faces the right carbon price. World Bank (2019) argued that the 2020 Paris target-consistent price was at least US$40-80/t[CO.sub.2]. At least in the EU and UK, carbon prices facing electricity were over [euro]90 ($95)/tonne for December 2022 (in May 2022), above this range. Most other countries impose far lower carbon prices. If it is politically difficult to raise carbon prices, then a second-best policy might be to subsidize all technologies (and notably VRE) in proportion to the carbon they abate (Newbery, 2018a).

    The learning externality depends on the cumulative installed capacity, not current output, of VRE. The learning benefit derives from R&D, design and production economies of scale, all driven by demand for deployment, and not from the output the technology produces once installed. (2) (Investors will demand reliable and suitably durable plant, provided they face undistorted price signals.) Thus for each doubling of installation of solar PV units, future unit costs fall by about 20%, and have done for 40 years (ITRPV, 2016; Frauenhofer, 2016; Rubin et al., 2015). Similarly, doubling on-shore wind farm capacity appears to lower future unit costs by 12% (IRENA, 2019). Andor and Voss (2016), drawing on Newbery (2012), demonstrate that if the only externality facing renewables is a learning spill-over, there is no case for subsidizing output. Similarly, Ozdemir et al. (2020) find that capacity, not output, support is the least-cost route to future RE output and carbon targets.

    Previous EU policy has specified target shares of renewable energy for each Member State. That encouraged inefficient output support (Meus et al., 2021), without questioning the underlying reasons for intervention. Fortunately, the EU Clean Energy Package has dropped the Member State RE requirement, emphasising instead decarbonisation at "the lowest possible cost to consumers and taxpayers" using "(M)arket-based mechanisms, such as tendering procedures" (Directive (EU) 2018/2001 [section]19). As such the barrier to directing support on the source of the learning externality, installation rather than output, has now been removed.

    The second market/policy failure in an industry prone to unpredictable policy interventions is missing futures and insurance markets (Newbery, 2016b). Without suitable long-term risk hedging contracts, investors face risky future revenues that significantly impact the cost of capital. Newbery (2016a, pl325) showed that replacing Renewable Obligation Certificates that paid a market-determined premium on a volatile wholesale price by a guaranteed fixed price lowered the cost of capital by 3.3% real. The implied saving on projected generation investment of [pounds sterling]75 billion up to 2020 (DECC, 2011) would be [pounds sterling]2.5 billion per year by 2020, continuing for 15 years. CEER (2020) shows the remarkable improvements achieved by tendering in the EU in coverage and downward pressure on prices since their 2018 report.

    There are also specific problems in determining the capacity credit of VRE and addressing potentially excess entry that might distort free unsubsidized VRE entry (Newbery, 2020). Such distortions can be simultaneously overcome by auctions for the efficient volume of entry.


    Least system cost requires that new VRE is the right technology in the right location and is dispatched efficiently. Location decisions depend both on the form of support and the design of connection and use of network system charges, which will also have to be set correctly (as discussed below). The policy maker will set the design format of the efficient contract to give the right signals to locate and operate and which reduces risk to lower the cost of capital. Once the contract has been designed, the required revenue can be determined by a well-designed sequence of auctions (see del Rio, 2017 on lessons for good auction design). Auctions are the best way to deliver least-cost procurement, with the added advantage of allowing control over the volumes of RE or cost of the RESS.

    Operation or dispatch decisions require the generator to face and respond to the efficient locational spot price for electricity. Within a technology class (PV, wind) the right design choice depends on selling all its services (including ancillary services like ramping down) at their efficient value (Meus, 2021). Thus the choice of height, blade diameter and controllability of wind-turbines can be distorted by inefficient price signals, while the orientation of PV panels should maximise value, not output (Borenstein, 2005).

    Different technologies justify different levels of support (as they have different learning rates). Auctions for different technologies can be run in parallel--in Britain more mature technologies like on-shore wind and solar PV are allocated a separate "pot" (of money) to off-shore wind. The most immature technologies like wave and tidal stream have their own pot. For auctions to work well, bidders need clarity on the future policies that may impact their contract value such as changes in Grid Codes or balancing rules. They need reliable predictions of (or comparable duration contracts for) differential locational use-of-network charges over a reasonable fraction of the life of the investment, or at least 10 years.

    The main future sources of renewable electricity are wind and solar PV. They have high capital costs but low running costs. Variable running costs for PV are zero, while for wind they are modest at [euro]5-12/MWh (BEIS, 2020; NREL, 2018). It follows that the major cost of VRE is the cost of financing the investment--the weighted average cost of capital, WACC. The more predictable and certain are...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT