Interaction between CO₂ emissions trading and renewable energy subsidies under uncertainty: feed-in tariffs as a safety net against over-allocation

ABSTRACT

We study the interactions between a CO₂ emissions trading system (ETS) and renewable energy subsidies under uncertainty over electricity demand and energy costs. We develop an analytical model and a numerical model applied to the European Union electricity market in which renewable energy subsidies are justified only by CO₂ abatement. We confirm that in this context, when uncertainty is small, renewable energy subsidies are not welfare-improving, but we show that when uncertainty is large enough, these subsidies increase expected welfare because they provide CO₂ abatement even in the case of over-allocation, i.e. when the cap is higher than the emissions which would have occurred without the ETS. The source of uncertainty is important when comparing the various types of renewable energy subsidies. Under uncertainty over electricity demand, renewable energy costs or gas prices, a feed-in tariff brings higher expected welfare than a feed-in premium because it provides a higher subsidy when it is actually needed i.e. when the electricity price is low. Under uncertainty over coal prices, the opposite result holds true.

Key policy insights

• Due to the possibility of over-allocation in an ETS, subsidies to renewable energies can increase expected welfare, even when climate change mitigation is the only benefit from renewables taken into account.

• In most cases studied, a feed-in tariff brings a higher expected welfare than a feed-in premium.

• The European Commission guidelines on State aid for energy, which incentivize member States to replace feed-in tariffs by feed-in premiums, should be reconsidered based on these results.

KEYWORDS:

• uncertainty
• ETS
• renewable energy
• feed-in tariff
• feed-in premium
• policy interaction
• uncertainty

Acknowledgements

For their useful comments, we thank Guy Meunier, as well as the participants at the EAERE, FAERE and AFSE conferences, CECILIA2050 workshop and CIRED and CERNA seminars, where earlier drafts of this paper were presented.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 We focus on subsidies to renewable energies rather than on other policies aimed at reducing emissions (e.g. in the transportation or industry sectors) for two reasons. First, electricity is covered by almost every GHG ETS in the world, unlike transportation or waste (Métivier et al., 2017, p. 3). Second, subsidies to renewable energies are much more widespread than climate policies in the industry sector which mostly consist of subsidies to R&D (Neuhoff et al., 2015). In the discussion section we briefly discuss whether our results would generalise to the latter policies.

2 A precision can be useful for our definition of over-allocation: several ETSs include a price floor taking the form of an auction reserve price (RGGI, California-Quebec). They can be seen as a hybrid price-quantity policy as analysed by Roberts and Spence (1976). When the auction clearing price equals the price floor, we consider that it is the price part of the policy which binds and not the quantity one, hence that there is over-allocation.

3 Cf. references in Fuss et al. (2018) and Branger et al. (2015).

4 Cf. Burtraw and Keyes (2018) for a longer discussion of the causes of over-allocation.

5 Goodkind and Coggins (2015) analyse another type of corner solution in Weitzman’s model: while they focus on complete abatement, we are interested in a nil CO₂ price. A complete abatement is irrelevant for greenhouse gases emissions before several decades at least.

6 Under a feed-in-tariff, a fixed price is guaranteed to producers of electricity generated by renewables for a period of typically 10–20 years.

7 A feed-in premium is a subsidy to electricity generated by renewables, which adds up to the market price at which electricity is sold. In this paper we consider a fixed premium, i.e. independent of the electricity price. A floating premium, under which a variable subsidy complements the market price to guarantee a fixed remuneration, would be equivalent to a feed-in-tariff (Dressler, 2016).

8 In Appendix A.5 we present the case of continuous uniform probability distributions. The qualitative conclusions remain the same.

9 Since we assume perfect competition it would make no difference to assume that some producers use fossils while others use renewables, rather than assuming a representative producer using both technologies.

10 Adding non-random linear terms to the cost function would make the resolution much more complicated without providing more insight. Hence we add them only in the numerical model below.

11 Admittedly, small hydro-electric plants are often subsidised but they are not separated from large ones in the Poles model and the potential for new capacity is limited in Europe so this is not of great importance.

12 Lignite is not separated from coal in the Poles model and the percentage of oil in electricity generation is so small (0.1% in 2025) that it is not worth keeping it as a separate technology.

13 We also ran the model without this tax. The welfare gain from subsidising renewables is higher with the tax because the latter reduces electricity consumption which, as we have seen in the analytical section, is too high.

14 We performed a sensitivity analysis with an elasticity of -0.5. Results are presented in Appendix A.7.

Additional information

Funding

This work was supported by the European Commission [FP7 project CECILIA2050, grant number 308680]; Seventh Framework Programme.