Analysing the interactions between renewable energy promotion and energy efficiency support schemes: The impact of different instruments and design elements

ARTICLE IN PRESS Energy Policy 38 (2010) 4978–4989

Contents lists available at ScienceDirect

Energy Policy journal homepage: www.elsevier.com/locate/enpol

Analysing the interactions between renewable energy promotion and energy efficiency support schemes: The impact of different instruments and design elements Pablo del Rı´o n ´ blicos, Consejo Superior de Investigaciones Cientı´ficas (CSIC), C/Albasanz 26-28, 28037 Madrid, Spain Instituto de Polı´ticas y Bienes Pu

a r t i c l e in f o

a b s t r a c t

Article history: Received 20 September 2009 Accepted 2 April 2010 Available online 11 May 2010

CO2 emissions reduction, renewable energy deployment and energy efficiency are three main energy/ environmental goals, particularly in Europe. Their relevance has led to the implementation of support schemes in these realms. Their coexistence may lead to overlaps, synergies and conflicts between them. The aim of this paper is to analyse the interactions between energy efficiency measures and renewable energy promotion, whereas previous analyses have focused on the interactions between emissions trading schemes (ETS) and energy efficiency measures and ETS and renewable energy promotion schemes. Furthermore, the analysis in this paper transcends the ‘‘certificate’’ debate (i.e., tradable green and white certificates) and considers other instruments, particularly feed-in tariffs for renewable electricity. The goal is to identify positive and negative interactions between energy efficiency and renewable electricity promotion and to assess whether the choice of specific instruments and design elements within those instruments affects the results of the interactions. & 2010 Elsevier Ltd. All rights reserved.

Keywords: Energy efficiency Renewable energy Interactions

1. Introduction Climate change mitigation and security of energy supply have triggered the implementation of a wide array of policies in Europe, including emissions trading schemes (ETS), support schemes for electricity produced with renewable energy sources (RES-E) and measures to encourage energy efficiency (EE). Targets in these three realms have been set in the EU (and in MS) for 2020. However, whereas an EU-wide ETS has been functioning since 2005, RES-E and EE support will remain in the hands of Member States (MS). The combination of these instruments raises concerns about overlaps, conflicts and synergies in their interaction. Previous analyses have focused on the interactions between the EU ETS and RES-E promotion and between the EU ETS and EE policies. However, the analysis of the interaction between the three instruments (ETS, EE and RES-E promotion) and, especially, between the last two has been very scarce1 and limited to two

n

Tel.: +34 91 6022560. E-mail address: [email protected] 1 NERA (2005) focuses on the interaction between those instruments, but in a separate manner (i.e., for ETS and TWC and ETS and TGC) but the simultaneous interaction between TWC, ETS and TGC and even between TGCs and TWCs is disregarded. 0301-4215/$ - see front matter & 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.enpol.2010.04.003

specific RES-E and EE instruments: tradable green and white certificates (TGCs and TWCs, respectively).2 The aim of this paper is to analyse the interactions between EE and RES-E support schemes focusing on electricity. Two main gaps in the literature are covered. First, the analysis identifies whether the choice of specific instruments affects the results of the interactions, transcending the ‘‘certificate’’ debate (i.e., TGCs and TWCs) and considering other instruments, particularly feed-in tariffs for RES-E promotion.3 Second, this paper analyses the impacts of different design elements on those interactions, an issue which has received scant attention in the interactions literature. Accordingly, the paper is structured as follows. Section 2 describes the methodology. The instruments and assessment criteria are described in Section 3. A general analysis of the

2 See, among others, Bertoldi and Huld (2006), Oikonomou and Patel (2004), Farinelli et al. (2005), Klink and Langniss (2006) and Child et al. (2008). For example, Oikonomou and Patel (2004) observe that an integrated scheme of TWCs and TGCs could be feasible in terms of institutional setup. Bertoldi and Huld (2006) propose a combined tradable certificate scheme in which both RES and demandside EE measures could bid in real time through the Internet to meet a specific obligation. 3 This analysis is deemed highly necessary because only six EU-27 countries have adopted ‘‘certificates’’ to promote RES-E and EE, whereas the majority uses feed-in tariffs for RES-E promotion and non-certificate instruments to encourage EE (see European Commission, 2008). In addition, 50 countries in the world had implemented feed-in laws by the end of 2008, 23 of which were European (Mendonca and Jacobs, 2009).

ARTICLE IN PRESS P. del Rı´o / Energy Policy 38 (2010) 4978–4989

interactions between the instruments is provided in Section 4. Sections 5 and 6 analyse whether different support schemes and design elements lead to different interaction results. The paper closes with some concluding remarks.

2. Methodology and main assumptions The analysis of interactions performed in this paper uses a qualitative method in a partial equilibrium framework, where the electricity sector is represented. The impacts of the interactions between those instruments are assessed by investigating how they affect a number of criteria and variables. This theoretical/methodological framework has been used in the past by, among others, del Rı´o et al. (2005), del Rı´o (2007), Jensen and Skytte (2003) and Sorrell et al. (2009). This paper complements other studies, which either analyse the interactions between TGC and ETS (Jensen and Skytte, 2002, 2003; Morthorst, 2003; NERA, 2005), between TWC and ETS (Sorrell et al. 2009; NERA, 2005) or between different EE policies (Boonekamp, 2006).4 Given the qualitative aspects involved in different design elements, the graphical approach (often used in this literature) is not deemed suitable to deal with the analysis of interactions proposed in this paper. In addition, only the potential direction of the effect but not its magnitude is considered. The analysis is based on the following assumptions:

 Separation between the three instruments, i.e., the achievements  



 

(certificates) from EE measures and RES-E promotion schemes are not translated into a CO2 emissions reduction value. A competitive electricity market. The costs of EE and RES-E support measures (the so-called ‘‘add-on’’) are fully passed to electricity consumers. The current situation in Europe is assumed: an EU-wide common ETS but domestic RES-E and EE support schemes. No country is large enough to significantly influence the ETS through their RES-E and EE support schemes. EE instruments reduce electricity demand by end-use sectors, which are covered by an ETS. Electricity suppliers are the obligated parties regarding the implementation of EE measures and RES-E deployment (under a TGC scheme). No EE measures and RES-E technologies simultaneously fall under EE and RES-E support schemes. Secondary effects are disregarded. Only the increase in retail electricity prices due to the implementation of EE and RES-E support instruments is considered, given its impact on EE investments.

3. Linking ETS, RES-E and EE support schemes: instruments, assessment criteria and variables 3.1. Instruments for the promotion of ETS, RES-E and EE Several instruments are available to reduce CO2 emissions, promote EE improvements and encourage RES-E deployment, of which market-based are usually favoured by economists (Table 1). Within this later category, an important distinction 4 See Del Rı´o (2007) for an overview of the interaction literature, mostly focused on the interaction between ETS and RES-E support schemes. The empirical literature on the interactions between RES-E and ETS has been based on qualitative, country-specific case studies (see Sorrell, 2003; Sijm, 2003) and simulation models (see Unger and Ahlgren, 2005; Linares et al., submitted for publication; De Jonghe et al., 2009, among others).

4979

which may have some impact on the results of the interactions is between quantity-based and price-based instruments.5 Within the EE realm, mandatory saving targets with TWCs have recently received much attention, although feed-in tariffs (FITs) for energy savings are also theoretically possible (see Bertoldi et al., 2009). In a TWC there is a requirement for electricity suppliers or distributors to meet a certain energysaving target through the promotion of EE measures among enduser sectors. The energy savings are measured, verified and certified (in a TWC), allowing them to be traded in a market. The ‘‘value’’ of the tradable commodity (e.g., kWh energy savings) is calculated with respect to a counterfactual baseline (NERA, 2005). In theory, this allows the obligated parties to meet their targets cost-effectively. In turn, RES-E promotion in the EU has mostly been based on two main mechanisms: FITs and TGCs. FITs are subsidies per kWh generated paid in the form of a total quantity (tariff) or as an amount on top of the wholesale electricity price (premium) fed into the grid and combined with a purchase obligation by the utilities. TGCs are certificates issued for every MWh of RES-E, allowing generators to obtain additional revenue to the sale of electricity (i.e., two streams of revenue). Demand for TGCs generally originates from an obligation on electricity distributors to surrender a number of TGCs as a share of their annual consumption (quota). Otherwise, they would pay a penalty. The TGC price strongly depends on several factors, including the level of the quota and the penalty and the duration of the obligation (IEA, 2008). In both cases (TGCs and FITs), the costs are borne by consumers. 3.2. Assessment criteria and variables The coexistence and interactions between instruments can be assessed according to several criteria (see, among others, Boonekamp, 2006; Oikonomou and Jepma, 2008; Konidari and Mavrakis, 2007). However, the focus in this paper is on the following three criteria: Effectiveness refers to the achievement of a certain target (i.e., emissions reductions, RES-E and EE investments). Cost-effectiveness refers to the achievement of the target at the lowest possible cost. Cost-effectiveness is attained when an instrument encourages proportionally greater emissions reductions, RES-E or EE investments by firms with lower costs and lower emissions reductions RES-E or EE investments by companies with higher costs, leading to an equalisation of marginal costs across firms/plants/agents (equimarginality). The administrative/ transaction costs of the instruments should also be considered. Dynamic efficiency refers to the ability of an instrument to generate a continuous incentive for technical improvements and costs reductions in technologies. Currently expensive technologies with a significant cost reduction potential need to be supported today in order to have them available in the future to reach new targets at moderate costs (Ragwitz et al., 2007). We thus analyse the impact of the interactions on several key variables relevant to the aforementioned three criteria:

 Regarding the effectiveness criteria, CO2 emissions, electricity demand and RES-E generation and investments are considered. 5 A detailed overview of instruments in these three realms is provided by IPCC (2007) (regarding CO2 emissions reductions), del Rı´o and Gual (2004)(regarding RES-E promotion schemes) and Harmelink et al. (2008) and Sorrell (2004)(regarding EE promotion). The lists in Table 1 do not pretend to be comprehensive. In addition, instruments belonging to the EE and RES-E promotion categories could be included in the CO2 emissions reduction category.

ARTICLE IN PRESS P. del Rı´o / Energy Policy 38 (2010) 4978–4989

4980

Table 1 Key quantity and price-based instruments to promote CO2 emissions reductions, EE improvements and RES-E. Source: own elaboration. Market-based instruments

Other instruments

Quantity-based

Price-based

CO2 emissions reductions EE promotion

ETS

Carbon taxes

TWCs

Taxes FITs

RES-E promotion

Quotas with TGCs tendering/bidding

FITs

 

RES-E investments undertaken by power companies depend on the wholesale price of electricity (Pw) and the add-on for RES. The add-on represents the unitary public support per kWh. EE investments mostly depend on the retail electricity price (Pr)6. Pr is the result of adding Pw to the add-on for RES-E and the add-on for EE support. Concerning static efficiency, the focus is on consumer costs (as shown by variations in Pr). Regarding dynamic efficiency, we analyse the impact on RES-E investments in general and, in particular, on the currently least mature technologies.

Following Konidari and Mavrakis (2007), interactions between instruments might be positive, when the performance of one or both examined instruments against a criterion increases because of their coexistence, or negative when the combined policies lead to negative impacts that would not have occurred by either alone.

4. A general analysis of the interactions between the instruments: A comparison of scenarios Several papers have analysed the impact of RES-E and EE support on a pre-existing ETS (NERA, 2005; Sorrell et al., 2009). The main results of such analyses are summarised in Table 2. In contrast, the analysis of interactions between RES-E and EE has been scarce. Adding RES-E support to an ETS increases RES-E deployment and reduces conventional electricity generation and Pw. Pr would increase, modestly encouraging the adoption of EE measures and reducing electricity demand (secondary effects). CO2 emissions would be unaffected, if the emissions reduced by RES-E are already covered by the ETS. Adding EE support to an ETS reduces electricity demand and Pw. Since the policy is paid by electricity consumers in their bills, Pr will likely increase. The reduction in electricity demand and Pw will reduce electricity generation in general and RES-E generation and investments in particular.7 The lower Pw discourages investments in renewables. Again, CO2 emissions are not affected. Adding EE support to RES-E support reduces electricity demand, Pw and the incentive for RES-E investments. However, the extent of this later reduction partly depends on the type of RES-E support scheme and its design elements (see Section 5) but it is likely to 6 It is assumed in this paper that RES-E investments are carried out by power companies, not end consumers. 7 Sorrell et al. (2009) argue that the impact on the output of existing renewable energy generators will depend on their position within the plant merit order. Since existing RES have low short-run marginal costs they should take preference in the merit order. Therefore, they are likely to provide base-load supply and hence be unaffected by small changes in demand.

Emission standards, information and education campaigns, voluntary agreements Investment subsidies, fiscal exemptions, tax rebates, soft loans, EE standards, voluntary agreements, information campaigns, labelling, procurement and demand-side management Investment subsidies, fiscal/financial incentives

be small, because the incentives provided by RES-E support dominate (against reductions in Pw). Again, CO2 emissions remain unaffected. Since these policies have to be paid for (double addon), Pr prices will be higher than in their absence. Adding RES-E support to EE support generally leads to the same final results than the previous case. Since EE and RES-E technologies benefit from the existence of a carbon price (as provided by an ETS), an ETS should be introduced first. This price signal is a necessary albeit not sufficient condition for their uptake and its implementation would reduce (but not remove) the need to apply the other instruments.8 The difference between the combination RES-E/ETS with respect to EE/ETS is that demand in the later combination is reduced whereas RES-E generation is increased in the former. Emissions, abatement costs and Pr will be at similar levels in both combinations, since both policies have to be paid for and emissions are covered by the ETS. Therefore, regarding CO2 emissions reductions, both instruments are redundant with respect to a previously established ETS and increase the abatement costs with respect to a single ETS. Thus, the reason for their coexistence must lay elsewhere, i.e., in the dynamic efficiency effects (see Section 3.2) and non- CO2 benefits (security of supply) brought about by RES-E and EE. Whether this is acceptable depends on the level of those benefits and society’s willingness to pay for them. In turn, there is not a strong interaction between EE and RES-E support, since they have different scopes. Their mutual impacts are mediated through the electricity market, with one instrument affecting the supply side and the other the demand side.

5. Do different support schemes lead to different interaction results? In this section, we analyse whether the interactions between RES-E and EE support schemes are affected by different RES-E support schemes (TGCs and FITs), leaving for future research how different EE support instruments influence those interactions. We start by analysing the interactions with a TGC scheme and then identify whether the results of the interactions change with a FIT. The addition of TGC or FIT to a pre-existing EE instrument (i.e., a TWC scheme) would indirectly influence the uptake of EE measures, due to an increase in Pr. This would increase the uptake of EE measures, reduce electricity demand and, thus, make it easier to achieve energy saving targets. 8 A carbon price might not be enough to encourage RES-E investments (see, e.g., Reinaud, 2003). In the case of EE, there might be sector-specific market failures to EE which a carbon price is unlikely to remove (see, e.g., Sorrell et al. (2004) and Schleich (2009).

ARTICLE IN PRESS P. del Rı´o / Energy Policy 38 (2010) 4978–4989

4981

Table 2 Comparison of scenarios. Source: own elaboration. Scenario

Elect. demand

RES generation

Conventional generation

CO2 em.

Pw Pr Addon

CO2 abatement costs

RES investments

Adding RES-E support to an ETS Adding EE support to an ETS Adding EE support to RES-E supportn Adding RES-E support to EE supportn

¼ o o

4 o ¼

o ¼ ¼

¼ ¼ ¼

o o o

¼ ¼ ¼

4 4 4

4 4 4

4 o o

¼

4

o

¼

o

¼

4

4

4

Where: (¼) indicates no change, ( 4) indicates an increase in the considered variable when the new instrument is added and ( o ) indicates a reduction. Pw ¼ Wholesale price of electricity; Pr¼ Retail price of electricity. A pre-existing ETS is in place.

On the other hand, adding an EE instrument to a TGC scheme can affect the TGC market, although only if the RES-E quota is set in percentage terms.9 In this case, the EE instrument would reduce electricity demand/production and the absolute RES-E requirement (Klink and Langniss, 2006). Jensen and Skytte (2002) observe that Pw and the TGC price (PTGC) move in opposite directions, i.e., a reduction in Pw triggers an increase in PTGC, leaving total support constant. However, support does not remain constant when electricity demand is reduced. Indeed, with a large demand reduction and a highly inelastic RES-E supply curve (MCre) near the quota level,10 there could be both a reduction in Pw and PTGC (Fig. 1), i.e., a double disincentive for RES-E investments. Therefore, there may be negative interactions regarding the effectiveness criteria when EE support is added to RES-E promotion (absolute reduction of RES-E), although this is only the case when a relative quota in a TGC scheme is implemented, not with a FIT, while the interaction is positive in the other direction. Thus, if RES-E and EE are to be promoted, and a relative quota with a TGC scheme is implemented, then the effect of EE support on RES-E should be considered and the relative quota should be set accordingly.11 Since the EE instrument makes it easier to attain the RES-E target (Child et al., 2008), its impact with a relative RES-E target should not necessarily be regarded as negative. Policy-makers would have an additional incentive to apply EE instruments in order to reduce electricity demand, particularly with a stringent RES-E target. In this case, significant RES-E increases may need to be combined with electricity demand reductions. Thus, the more stringent the RES-E target, the lower the costs of applying an EE instrument. This additional incentive to apply EE instruments does not exist under a FIT scheme because, in this case, the EE instrument does not affect the costs of reaching the RES-E target. This calls for some coordination when setting targets in both realms and consideration of the avoided RES-E costs due to the implementation of EE support. The aforementioned negative impact on RES-E investments would be particularly so for the most expensive renewable technologies,

9

Quota targets for RES-E can be set either in relative terms (i.e., as a percentage of total electricity sales) or in absolute terms (i.e., an amount in MWh). All EU countries have set a relative quota target. This is due to the fact that targets in the Renewable electricity Directive (2001/77/EC) and the new Renewable Energy Directive are set in relative terms. In contrast, an absolute quota (in TWh) has been set in Australia. 10 Although the RES cost curves are relatively flat on a large range of cumulative installed capacity, they are much less elastic near the quota target level. 11 This might be difficult, however, because the reduction in electricity demand due to the implementation of EE support should be predicted, which is not an easy task.

/MWh

n

MCre Support (0) Support (1)

PTGC (0) PTGC (1)

PW (0) PW (1) Q1

Q0

MWh

Fig. 1. The impact of electricity demand reduction on the TGC market. Source: own elaboration. Note: For illustrative purposes, it is assumed that the relative quota (%) translates into an absolute amount of RES-E (MWh). Obviously, the reduction in electricity demand does not affect the relative quota, but it reduces the absolute amount of RES-E which is needed to comply with it, as shown by the shift of Q to the left.

which were previously needed to meet the quota.12 This further reduces the already low technological diversity in a TGC scheme, negatively affecting dynamic efficiency.13 In practice, however, since the most expensive technologies are usually promoted in a TGC scheme through secondary instruments,14 they will be less affected by the interaction with EE support. The interaction between EE and RES-E support is more modest with secondary instruments, because these have a lower impact on the electricity market. On the other hand, a changing TGC price is also problematic for RES-E investments. EE measures could increase the volatility of the TGC market as a result of the reduction of the TGC market volume, which reduces its liquidity, making TGC price spikes more likely. Would FITs be affected by the introduction of an EE measure which reduces electricity demand and, thus, Pw? Compared to TGCs, FITs are generally less affected by EE measures, but this depends on the type of FITs implemented. Under a regulated tariff, support does not depend on Pw, but under a feed-in premium generators receive the Pw and the premium and, thus, part of the total support is affected. The reduction in Pw reduces the overall incentive to invest in RES-E. Since the RES cost curves are relatively flat on a large range of cumulative installed capacity

12 Of course, we refer to expensive renewable technologies whose long-term costs would be just below the price of the TGC before the EE instrument was introduced, and which would be no longer needed to meet the quota once the later instrument was applied. Other renewable technologies with even higher costs would not be affected by the introduction of the EE instrument, since they were not needed to meet the quota in the absence of such instrument anyway. In turn, the profitability of inframarginal RES-E investments would be reduced as a result of EE promotion but they would still be undertaken since their costs are lower than the addition of the electricity price and the (lower) TGC price. 13 Only the cheapest technologies will be deployed in a TGC scheme without a technology-specific quota. See Lipp (2007) and Mitchell and Connor (2004), among others. 14 For example, in Italy solar PV is supported with FITs.

ARTICLE IN PRESS 4982

P. del Rı´o / Energy Policy 38 (2010) 4978–4989

(Finon and Pe´rez, 2007), a slight variation in Pw may have major repercussions in terms of quantities produced under a FIT premium. To sum up, the type of support scheme has some impact on the interactions, but it might be significantly mediated by the design elements of those instruments. This is the focus of the next section.

6. Do different design elements in RES-E instruments lead to different interaction results? In order to simplify the analyses, the interactions literature has usually considered ‘‘idealised’’ instruments with simplified designs, disregarding the fact that instruments that are actually implemented, such as RES-E support schemes, may have different design elements15 which may affect the interactions between instruments. This is in line with the importance attached to the impact of the design elements on the success of RES-E support schemes (see, for example, del Rı´o, 2008; IEA, 2008, Ragwitz et al., 2007; Couture and Gagnon, 2010). Therefore, the aim of this section is to identify the impacts of the different design elements of RES-E promotion schemes (Table 3) on these interactions. These are either included in current schemes in Europe or planned. They will be described in the next subsections.16 The analysis is carried out considering, first, that a RES-E support instrument (a FIT or a TGC) is added to EE support (Section 6.2.1) and, second, that EE support is added to RES-E support. A pre-existing ETS is assumed in both cases. We identify the impact on the considered variables. Table 4 summarises the results of this analysis. 6.1. Adding RES-E support to EE support 6.1.1. FITs Fixed premium versus tariff: In countries allowing RES-E generators to opt for either a fixed premium or a tariff (e.g., Spain, the Czech Republic and Estonia), the fixed premium option is more market-oriented, allowing RES-E generators to sell their electricity in the wholesale market, whereas under the tariff option RES-E is generally purchased by distributors under the tariff option. A lower Pw can be expected in the premium case compared to the tariff option, because RES-E would be encouraged to participate in the wholesale market and the low variable 15 For example, NERA (2005, p.104) argues that ‘‘the aim is to understand the basic implications of a TWC scheme without complicating the analysis with market and design features that vary from scheme to scheme’’, although a discussion of the implications on the interactions between instruments of some design elements is provided. In their theoretical paper on the interaction between the EU ETS and TWCs, Sorrell et al. (2009, p.29) state that ‘‘the paper examines the (y) impacts of an idealised TWC scheme’’ and that ‘‘the paper abstracts from the empirical details of individual schemes’’. The design elements of RES-E support schemes (feed-in tariffs and TGCs) widely differ across countries, as shown by Ragwitz et al. (2007) and European Commission (2008). 16 Other aspects and factors which have not been considered in this paper might have a bearing on the results of the interactions, including electricity market structure and design, the electricity liberalisation process, the extra-cost recovery mechanism and the interconnection of domestic electricity markets in Europe, among others. In turn, some design elements within these schemes have not been considered, such as whether the obligation is imposed on RES-E generators rather than on electricity suppliers, as in Italy, whether the revenues from the penalty are returned back to compliant RES-E suppliers (as in the U.K.), recourse to long-term (private) contracts, acceptance limits regarding which production installations can receive certificates, maximum number of TGCs per installation, number of TGCs issued in exchange for which production of sustainable energy (MWh, kg of CO2 emissions avoided) (for TGCs), guaranteed duration of support, annual updating of support levels, revisions of support levels, complement for reactive power (voltage control), payment for capacity guarantee, payment procedure, access to the grid, share of the costs of grid connection and reinforcements and treatment of the costs of deviations (for FITs).

Table 3 Main design elements of RES-E support schemes. Source: own elaboration based on Ragwitz et al. (2007), European Commission (2008), del Rı´o (2008), Nielsen and Jepessen (2003), Mendonca and Jacobs (2009). FITs

TGCs

Fixed premium versus fixed tariff RES-E support tied to electricity prices Technologies excluded Existing plants included Cap system Floor system Stepped tariffs (technology-specific) Degression.

Target (absolute/relative). Technology-specific quota Minimum prices Maximum prices Technologies included/excluded. Existing plants included. Banking. Borrowing.

costs of RES-E generation would tend to push down Pw (if the share of RES-E in total electricity is high). In contrast, the add-on under the premium can be expected to be higher than under the tariff alternative, offsetting the lower Pw and leading to a greater Pr and incentive to invest in EE.17 The reason is that the premium option is more risky for RES-E investors because the trend of part of the support (Pw) is unknown, and highly dependent on difficult to predict factors. In order to compensate for the greater risk, governments may increase the premium, i.e., the premium +Pw would be greater than the tariff.18 This would lead to the same RES-E increase in both options, but with a higher support level (add-on) in the premium case.19 Support tied to electricity prices (linkage): In some countries, FIT support (either regulated tariffs or premiums) have been linked to electricity prices. For example, this was so in Spain between 2004 and 2007 (del Rı´o, 2008), Germany between 1991 and 2000 (Butler and Neuhoff, 2008) and Denmark between 1993 and 2001 (Lipp, 2007). In this case, RES-E support is more affected by variations in electricity prices than if support was not tied to those prices. In turn, the impact on Pw, the add-on and Pr is endogenous to the evolution of electricity prices. Since either an increase (Spain, 2004–2007) or a reduction (Germany, late 90s) in those prices is possible, there could either be a higher or a lower RES-E level, Pr and EE incentives in the two options. A priori, a greater risk in the linkage option would result in lower levels of RES-E investments. However, the Spanish case shows that the higher risk was offset by the higher support, leading to significant increases in RES-E investments (wind and solar). Stepped FIT (technology-specific): Technology-specific FITs are one modality of stepped FITs, whereby support levels are differentiated across technologies to reflect technology-specific generation costs, with higher support given to the most expensive technologies and lower support to the cheaper ones. In contrast, with a flat FIT, the same level of support would be granted to different technologies.

17 For example, the total (average support) received by RES-E generators in Spain under the premium in the period 2004–2007 was estimated to be 0.3 ch/ kWh higher than the regulated tariff option (CNE, 2004). This is exactly the incentive to participate in the wholesale market. Ragwitz et al. (2007) also report a greater remuneration in the premium option in the Czech Republic, but not in Slovenia, where the overall remuneration is supposed to be the same for, both, the premium and the fixed options. 18 Indeed, Ragwitz et al. (2007) empirically show that the fact that risks are greater in the case of the premium has made some countries chose the level of the premium in a way that the overall remuneration of this option is higher than the fixed tariff option (i.e., the Czech Republic and Estonia) in order to compensate for the higher risk for RES-E producers. In contrast, in Slovenia the overall remuneration is supposed to be the same for both options. 19 With an ETS in place, the risks for RES-E generators of the premium option are higher, but revenues are also likely to be higher, because the ETS pushes the Pw upwards.

Table 4 Impacts of RES-E support design elements on the interactions with EE (RES-E-EE) and impacts of EE with different RES-E design elements (EE-RES-E). Source: own elaboration. Criteria/ variable

FIT Fixed premium (reference: fixed tariff)

RES-E-EE

EE-RES-E

CO2 Pw

RES-E addon

Pr

EE (secondary effect)

RES gen.

RES I

CO2

[ ¼] [ o] (reduction in both cases, but lower Pw with premium) [ ¼] [?]a

[ 4] (increases in both cases, but greater with premium). [?]a

[ 4] (increases in both cases, but greater with premium). [?]a

[ 4] (increases in both cases, but greater with premium). [?]a

[ 4]

[¼ ]

[?]

[o]

[o]

[o]

De

EE RES add-on addon

RES gen.

RES I

[¼ ]b [ ¼]c [ ¼]c [ ¼]a [ o] (lower add-on with premium)

[ o] (increase in both cases but less with premiums, unless adjusted).

[ o] (reduction with premiums, uncertain with tariffs)

[?]a

[¼ ]b [ ¼]c [ ¼]c [ ¼]a [ o] (lower add-on with linkage)

[ o] (lower increase [ o] (lower with with linkage) linkage)

[ o] (reduction with premiums, uncertain with tariffs) [ o] (lower with linkage)

[o]

[o ]

[¼ ]b [ ¼]c [ ¼]c [ ¼]a [ ¼] (lower in both cases)

[ ¼] (increase in both cases)

[ ¼] (reduction in both cases)

[¼ ]b [ ¼]c [ ¼]c [ ¼]a [ 4] (lower reduction with floor) [¼ ]b [ ¼]c [ ¼]c [ ¼]a [ ¼] (lower in both cases).

[ 4] (lower reduction with floor) [ ¼]a

[ 4] (lower reduction with a floor).

[ ¼] (reduction in both cases) [ 4] (lower reduction with a floor). [ ¼]c

[ ¼] [ o] (small effect)

[ 4]

[ 4]

[ 4]

[ 4]

[4 ]

Stepped (technology-specific). (ref.: flat rate) Degression. (ref.: flat rate across time)

[ ¼] [ o]

[ 4] (likely)

[ 4] (likely)

[ 4] (likely)

[ 4]

[4 ]

[ ¼] [ 4]

[o]

[o]

[o]

[o]

[o ]

[¼ ]b [ ¼]c [ ¼]c [ ¼]a [ o] Degression effect dominates adjustment effect (see text)

[ o] Degression effect dominates adjustment effect (see text)

[ o] Degression effect dominates adjustment effect (see text)

[ o] Degression effect dominates adjustment effect (see text)

[ 4]

[o]

[o]

[o]

[o]

[¼ ] if immature supported with FITs

[¼ ]b [ ¼]c [ ¼]c [ ¼]a [ 4] (lower reduction with exclusion of immature)

[ 4] (lower reduction with exclusion of immature)

[ 4] (lower reduction with exclusion of immature)

[ 4] (lower reduction with exclusion of immature)

[ o]

[ 4]

[ 4]

[ 4]

[ 4]

[4 ]

[¼ ]b [ ¼]c [ ¼]c [ ¼]a [ o] (greater reduction with low penalty)

[ o] (greater reduction with low penalty)

[ 4] (greater reduction with low penalty)

[ o] (small effect)

[ 4] (small effect)

[ 4] (small effect)

[ 4] (small effect)

[ 4]

[4 ]

[¼ ]b [ ¼]c [ ¼]c [ ¼]a [ 4] (greater reduction without minimum prices)

[ 4] (greater reduction without minimum prices)

[ 4] (greater reduction without minimum prices)

[ ¼]

[ 4]

[ 4]

[ 4]

[¼]

[4 ]

[¼ ]b [ ¼]c [ ¼]c [ ¼]a [ ¼] (reduction in both cases)

[ ¼] (reduction in both cases)

[ ¼] (reduction in both cases)

[ o] (greater reduction with low penalty) [ 4] (greater reduction without minimum prices) [ 4] (reduction in both cases,

QUOTA WITH TGCS Immature [ ¼] technologies excluded (ref.: all technologies included). Low penalty [ ¼] (ref.: appropriate penalty). Minimum [ ¼] prices. (ref.: no minimum prices). Existing [ ¼] plants noneligible.

[ ¼]c

4983

Floor (ref: no floor)

ARTICLE IN PRESS

Pr

P. del Rı´o / Energy Policy 38 (2010) 4978–4989

Support tied to electricity prices Cap. (ref.: no [ ¼] [ 4] (small) cap)

Pw

ARTICLE IN PRESS

c

b

Increase in both cases. Unchanged in both cases. Reduction in both cases.

Obviously, immature technologies would have a better chance to penetrate the market with a stepped than with a flat FIT. Therefore, there would be more RES-E deployment and a slightly lower Pw under the stepped alternative, because investments in both mature and immature technologies would be promoted. It is difficult to tell whether the add-on would be higher in the stepped or the flat option. Fig. 2 shows why. While the penetration of the immature technologies entails greater costs with a stepped FIT (areas ‘‘B+ C’’ in Fig. 2b), the difference between technology costs and the support level would be greater in the flat alternative for the mature technologies (area ‘‘A’’ in Fig. 2a). Whether one factor offsets the other is an empirical matter, and depends on the level of the flat FITs, the stepped FITs and the costs of the different renewable energy technologies. Since there is a lower Pw in the stepped FIT option, a lower consumer cost seems likely (and, thus, a lower Pw and incentive for EE investments). Cap price: This design element is only currently implemented in Spain.20 By capping the total (unitary) support to be obtained by RES-E generators (Pw+ premium), it limits the influence of interactions between the EU ETS and FITs and the impact on consumer costs. If the cap is activated, then a lower level of support (add-on) and, thus, a lower RES-E level (generation and investment) could be expected. There would also be a slightly higher Pw (given lower RES-E levels), but also a lower Pr and incentive to invest in EE measures than without a cap. Floor: By putting a lower limit on the support level (premium+Pw), the floor limits the influence of interactions ‘‘downstream’’ between an ETS and FITs. If the floor mechanism is activated, then a higher level of support than without the floor and, thus, a higher level of RES-E could be expected. A slightly lower Pw (given higher RES-E levels) but a higher add-on, Pr and incentive to invest in EE measures would result. Degression refers to reductions over time in support levels for new plants with respect to the flat alternative. Pw would be greater with degression, given the lower RES-E. The add-on, Pr, EE investments and RES-E would be lower.21

a

[?] [ o ,¼bank. (borrow.] [o] [o]

[o] [o] [o] (ref.: eligible).

Technology[ ¼] [ ¼] specific quota (ref.: non-specific quota) Banking and [ ¼] [ ¼] borrowing (ref.: no bank. and borrow.).

Pr CO2 Pw

RES-E addon

[o]

[¼ ] [¼]

[¼ ]b [ ¼]c [ ¼]c [ ¼]a

[¼ ]b [ ¼]c [ ¼]c [ ¼]a

but lower under noneligible case). [ 4] (lower [ 4] (lower [ 4] (lower reduction in [ 4] (lower reduction in the reduction in the the non-specific quota reduction in non-specific quota non-specific quota case) the noncase) case) specific quota case) [ o] (reduction in [ o] (reduction in [ ¼?] (reduction in both [ 4] both cases, but both cases, but cases, but ambiguous (reduction price spikes less price spikes less whether greater lower with likely with banking/ likely with banking/ reduction with banking/ banking/ borrowing) borrowing) borrowing) borrowing)

RES I RES gen. Pr EE RES add-on addon De Pw CO2 RES I RES gen. EE (secondary effect)

EE-RES-E RES-E-EE Criteria/ variable

Table 4 (continued )

Note: Between parentheses, the impact of measures is shown, whereas brackets refer to the differential impact of one versus another design option. CO2 ¼ CO2 emissions; Pw ¼ Price of electricity (wholesale); Pr¼ Price of electricity (retail); EE ¼Energy efficiency; RES gen. ¼ RES-E generation; RES I ¼Investments in RES-E; De ¼electricity demand.

P. del Rı´o / Energy Policy 38 (2010) 4978–4989

4984

6.1.2. TGCs Quota target: A relative quota may lead to a greater or a lower absolute amount of RES-E than an absolute target because the level of electricity demand and electricity sales cannot be predicted. Thus, the other variables (add-on, Pr, EE, etc.) can also be either higher or lower. An absolute quota is preferable for investors, given the higher certainty on the needed amount of RES-E generation, reducing the costs of financing (lower risk premium). Technologies excluded (immature): A frequent criticism of TGC schemes is that they do not encourage the least mature technologies, whose costs are higher than the marginal technology needed to comply with the quota (i.e., TGC price). This has led to either a promotion of these expensive options with other instruments, to granting more TGCs per MWh to the most expensive technologies or to setting technology-specific quota. The first option is discussed below. 20 An antecedent was the cap applied for wind on-shore in Denmark in 2003– 2004, and abolished in 2005. 21 Nevertheless, investment risks are slightly lower with a degressive tariff, because it provides some clues on the evolution of RES-E support in the future. Whether this offsets the lower support is uncertain. In addition, if the degression design element was anticipated by RES-E investors, then some earlier investments would occur. The impact of this anticipation on the variables considered is likely to be small, however.

ARTICLE IN PRESS P. del Rı´o / Energy Policy 38 (2010) 4978–4989

Q (flat)

STEPPED

MCre

/MWh

/MWh

FLAT

4985

Q (non-mature) MCre

Q (mature)

D A

C B MWh

MWh

Fig. 2. Consumer costs with different FIT designs. Source: own elaboration.

At least two cases are possible: one in which an expensive technology which would be part of the quota was excluded from meeting such quota and another in which the technology would be too expensive to even be part of the quota. Only the first case is worth analysing. Excluding those technologies would involve reducing the quota accordingly. However, since they are expensive, their impact on the marginal costs of reaching the quota is greater than on the level of the quota itself. Therefore, excluding those technologies implies that the quota is lower and is achieved at much lower marginal and total costs.22 Therefore, RES-E generation, the add-on, Pr and EE investments would all be lower.23 Low penalty: An appropriate penalty (i.e., above the marginal costs of the marginal technology which sets the TGC price) discourages non-compliance. Absent this, a lower level of RES-E generation than the quota could result. This would trigger a higher Pw. Consumer would be better off because the higher Pw would be more than offset by the significantly lower add-on and Pr. EE investments would be lower. Minimum TGC prices (as in Sweden before 2008) would be attractive for investors, since they ensure a minimum level of revenue. If minimum prices were activated (which may never occur), the add-on, Pr, EE and RES-E generation would probably be higher than without a minimum price. This design element works very much like a floor price for FITs24 but, since TGC prices are more volatile than FIT premiums, its benefits for dynamic efficiency may be greater. Technology-specific quota: Targets for different technologies may exist, leading to a fragmentation of the TGC market, with one quota for the mature and another for the non-mature technologies. The static efficiency gains of such a reduced market would be

22 Of course, target setting is not independent on the technologies available. In reality, policy makers take into account the available technologies when setting the target. 23 If the immature technologies are promoted with a different instrument (FIT), then the same amount of RES-E generation would result, but still at much lower costs than if they were part of the quota. The reason is that these would be the marginal technologies and would set the level of the TGC price. If the immature are not eligible for the target, this would result in a lower TGC price, which would reduce the overall costs of support. Therefore, there could be the same level of support for the immature (with FITs and TGCs) but a much lower level of support for the mature when the immature are not eligible for the quota, since the remuneration (TGC price) for the mature is lower. 24 For example, in Belgium, minimum prices are ensured if the sum of the electricity market price and the TGC price is insufficient to cover the costs of electricity generation (Ragwitz et al., 2007).

lower. The TGC price would be more volatile. In turn, the TGC price in the mature technology market would be lower compared to a single TGC market, because the costs of the marginal technology needed to reach the target are higher in a single TGC market. With a technology-specific quota, this marginal technology would set the price of TGCs in the immature market, but not in the mature one, leading to a lower level of RES-E support in the case of technology-specific quotas. However, this is (partially) offset by the lower trade benefits in fragmented markets and, thus, higher compliance costs. The first effect is likely to dominate and, thus, the add-on would be higher in the case of a single TGC market. Pr would also be higher, encouraging subsequent increases in EE. Existing plants eligible: Existing plants should not be eligible under a TGC scheme because, otherwise, only a very small additional capacity would be promoted with significant windfall profits for existing plants.25 Obviously, the overall costs would be lower if existing capacity was included, because cheaper options would be enough to comply with the quota.26 The inclusion of existing plants would negatively affect both new RES-E and EE investments (given the lower Pr). Banking: In general, banking would lower costs (intertemporal efficiency). It would probably boost RES-E generation in the shortterm and less so in the longer run.27 On average, the add-on would probably be lower with banking, leading to a lower Pr, EE and RES-E investments. However, banking tends to reduce the volatility of the TGC price, which favours RES-E generation. Therefore, the impact on RES-E investments is inconclusive. Borrowing: Same effects as banking result, although, in contrast to banking, short-term RES-E investments are discouraged. To sum up, several RES-E design elements have a significant effect on the interaction with a previously implemented EE support instrument, including a premium and floor (FITs) and the

25 Sweden is one example where certificates are issued for old and new capacity in the same way (Wang, 2006). 26 The overall costs of the scheme including existing capacity are lower than when only new capacity is eligible for TGCs, but the costs of promoting new capacity are very high (dividing the overall costs of the scheme by the small amount of new capacity being promoted). 27 Amundsen et al. (2006) show that the introduction of TGC banking may reduce price volatility considerably and lead to increased social surplus. Banking lowers average prices but may lead to higher TGC prices in the short-term. Indeed, Finon and Perez (2007) note that banking can increase the lack of liquidity in a period of tight supply of TGCs. Therefore, the add-on and the Pr are higher in the short-run and lower in the long-run. On average (i.e., short-run and long-run) the add-on is lower with banking than without it.

ARTICLE IN PRESS 4986

P. del Rı´o / Energy Policy 38 (2010) 4978–4989

eligibility of existing plants, the existence of a technology-specific quota and banking and borrowing (TGCs). 6.2. Adding EE policies to RES-E support The impact of EE measures on RES-E can take place if they either affect Pw or the add-on, i.e., the premium or the tariff (in case of FITs) or PTGC. 6.2.1. FITs Fixed premium: Compared to tariffs, RES-E generators and investors under a fixed premium are much more affected by reductions in Pw due to the implementation of EE measures. With a fixed premium, RES-E generators receive two streams of revenue (Pw and the premium). Variations in Pw directly affect the profitability of RES-E. With tariffs, a total remuneration is granted, which is relatively independent of Pw. Therefore, this design element is relevant to explain the different impact of EE measures on RES-E. The add-on would be lower in the premium case since the lower Pw triggers a lower RES generation. However, the government may increase the premium when Pw is reduced in order to maintain revenues constant for RES-E generators (see Section 6.1.1), which would lead to constant and similar incentives for RES-E deployment and consumer costs (add-on+ Pw) in both FIT options. The implementation of a premium is assumed in the rest of this section.28 Support tied to electricity prices (linkage). Obviously, the impact of EE on RES-E support is greater when the later is tied to electricity prices, reducing both revenue streams for RES-E generators/investors: Pw and add-on (compared to just a reduction in Pw in the case of non-linkage). Therefore, lower levels of RES-E investments could be expected. Cap: This mechanism is irrelevant in this context because it is activated only when the aggregation of Pw and the premium exceeds a certain limit and the implementation of EE measures would only reduce Pw, not increase it. Floor prices would limit the impact on RES-E of a reduction in electricity demand and Pw triggered by EE measures because the total remuneration for RES-E plants would not fall below the floor price. If the floor price was activated, then the remuneration would be higher than without it and the risks for investors would be lower. This would encourage RES-E investments. Stepped (technology-specific): There is not a significant impact of EE measures on the key variables in this case. The government is more likely to adjust the premium after a reduction in Pw when there is a technology-specific premium than under the flat alternative. It is more likely to increase it for the cheapest technologies, which would be relatively more affected by a reduction in Pw, since it represents a larger share of total support than in the most expensive technologies. Degression: The reduction of Pw due to EE support can be problematic for RES-E investments with degressive premiums, because it adds to the reduction of the premium over time (i.e., both streams of revenue). This impact could be mitigated with a floor price. 6.2.2. TGCs The following relevant design elements make a difference regarding the impact of EE. Absolute versus relative quota: The implementation of EE support reduces electricity demand and the absolute amount of 28 The reason is that the impact of the other design elements can be more clearly illustrated under this design element.

RES-E generation under a relative RES-E quota whereas this does not occur with an absolute quota. The add-on and the Pr are reduced under a relative target compared to the absolute target case.29 In turn, a lower Pr negatively affects EE (rebound effect), although this secondary effect is unlikely to offset the ‘‘main’’ effect of EE support. See UKERC (2007) for an analysis of rebound effects in energy efficiency.30 Technologies excluded (immature): Under a relative quota, EE reduces the absolute quantity of RES-E and, thus, the RES-E addon. This happens whether expensive/immature technologies are eligible for the quota or not. However, when they are included, such reduction in the add-on (and, thus, in consumer costs) is greater. This is shown in Fig. 3. Part (a) in Fig. 3 represents the overall TGC market (i.e., both mature and immature technologies), whereas (b.1) shows the TGC market for the mature technologies only. For illustrative purposes it is assumed that the non-mature technologies are also supported through a specific quota and TGC scheme (b.2), although this is highly unlikely in reality.31 Q0 and PTGC(0) represent the initial levels of the quota and TGC prices, respectively.32 EE support reduces electricity demand, and a lower absolute quantity of RES-E is needed to comply with the quota (shift in the quota line to the left).33 This triggers a reduction in the TGC price, since this is set by the marginal costs of the marginal technology. The reduction is greater in the overall market (a) than in the TGC market with only mature technologies (b.1), leading to a greater cost saving for consumers (given by the reduction in the PTGC times RES generation) in the overall market: graphically, area A is greater than areas B1 + B2.34 Appropriate penalty: Too low a penalty would result in compliance problems, possibly causing a lower RES-E generation, add-on and Pr and a lower Pr than if the penalty was set at an appropriate level (high above PTGC). EE support mitigates the problem caused by a low penalty because, as shown before, it would push down the TGC price. If this reduction was significant, then the PTGC could come close to the low penalty (Fig. 4). Nonetheless, demand would have to be greatly reduced to significantly push TGC prices down (to PTGC(1)). Whether this occurs is an empirical issue. In addition, expecting EE to reduce the PTGC is not an appropriate manner to deal with the problem of a low penalty, but rather to set the penalty according to the TGC price (i.e., a percentage above it).

29 Note that, in this case, the variations in Pr result from the variations in Pw, the add-on for RES promotion and the add-on for EE promotion. A reduction in RES-E generation tends to (modestly) increase Pw and (significantly) reduce the RES-E promotion add-on whereas the EE measure tends to (significantly) put downward pressure on Pw and (modestly) increase the EE promotion add-on. Thus, a reduction in Pr can be expected. In the case of an absolute target, RES-E is not reduced as a result of the implementation of EE measures. Therefore, the final effects on Pr are uncertain, although a small increase is possible. 30 In order to better illustrate the impact of the other design elements, it is assumed in the rest of this subsection that a relative quota is implemented. This is the most common design element in existing quotas with TGC schemes (see IEA, 2008). 31 However this assumption is not crucial for our analysis. Indeed, the results would not differ substantially whether a quota with a TGC scheme or a FIT system was applied to the most expensive (immature) technologies. 32 The quota is represented in this figure as an absolute quantity of RES, corresponding to a percentage of total sales. This is done for illustrative purposes only and has no bearing on the results. 33 Obviously, the quota in percentage terms does not change, but there is a reduction in the absolute amount of RES-E needed to meet the quota. 34 It has been assumed that the reduction in RES is the same in both cases (i.e., QI0–QI1 ¼(QE0–QE1) + (Qe0 –Qe1). In other words, the total reduction in RES-E experienced in the overall TGC market (a) is the same that the combined reductions of RES in the TGC market with mature technologies only (b.1) and with immature technologies only (b.2).

ARTICLE IN PRESS P. del Rı´o / Energy Policy 38 (2010) 4978–4989

4987

/MWh MCre

PTGC(0) A

PTGC(1)

(a)

QI0

QI1

MCre

PTGC(0) PTGC(1)

QRES

B1

(b.1)

QE1

PTGC(0)

QE0

QRES

MCre

PTGC(1)

B2

Qe1 Qe0

(b.2)

QRES

Fig. 3. Analysing the impact of EE measures when immature technologies are excluded. Source: own elaboration.

Minimum TGC prices: Although it is very unlikely that EE support by itself pushes PTGC below the minimum price, it could add to other factors leading to this situation. If this mechanism was activated, RES-E and consumer costs would be greater than if it was not (i.e., a greater add-on would result). Technology-specific quota: The implementation of EE support is more likely to reduce the add-on and consumer costs of RES-E promotion when there is not a technological differentiation than when there is one. This is so because, in the non-differentiation case, there is only a TGC price set by a relatively expensive renewable energy technology. Thus, this price is likely to be higher than the TGC price in the mature market with a technology-specific quota. The cost-saving effect triggered by EE measures with non-differentiation is greater because these would significantly reduce the PTGC compared to the existence of separate markets. Existing plants not eligible: If existing plants are eligible for the quota, the share of new RES capacity needed to meet the quota would be small even without EE support. But a reduction of demand due to EE support would further reduce the possibility that new RES capacity meets the quota, since it is more costly than existing (almost paid-off) capacity.35 The add-on and Pr would be the same in both cases, however. Banking/borrowing. EE support reduces the amount of TGCs needed to comply with the quota. A smaller market increases the volatility of the TGC price and further discourages RES-E investors. By allowing intertemporal transfer of TGCs, banking and borrowing mitigate this problem. Theoretically, this mechanism would increase intertemporal cost-effectiveness in RES-E deployment and, thus, reduce the add-on. However, price spikes, which are less likely with banking/borrowing, may encourage RES-E generation by existing plants. Therefore, the impact on generation is

35 In contrast, when existing plants are not eligible, only new capacity can comply with the target. Of course, part of this new capacity would also be affected by the EE measures reducing De, but only a small share (the highest cost), not all.

ambiguous, whereas the effect on RES-E investments is clearer. Again, EE support is unlikely to create this narrower market problem although it can contribute to it. To sum up, the impact of EE measures on RES-E deployment is greater when a TGC scheme (with a relative quota) has previously been implemented. In the case of FITs, the impact is negligible if a tariff is adopted and greater with a premium due to the effect of EE support on Pw (i.e., not on the premium itself). The negative impact of EE can be mitigated with ex-post adjustments in the premium or with a floor price. In the case of TGCs, the adaptation could involve changes in the target to capture the influence of EE support or minimum TGC prices. In theory, the interaction of a EE support added to a TGC scheme might be particularly problematic for the most expensive renewables, reducing the incentives to their deployment. However, what could be regarded as problematic in the first place might not turn out to be so if those interactions are taken into account and the respective EE and RES-E targets are adjusted accordingly. In reality, however, countries with a TGC scheme use a non-TGC instrument to promote the most expensive (immature) technologies and, thus, this mitigates the impact of EE support. In contrast, these negative dynamic efficiency effects are less likely to occur under a FIT because only part of the support would be affected (Pw, and only in the premium case), which represents a very small part of the overall support received by the most expensive technologies. Finally, it should be acknowledged that there might be design elements which affect different types of RES-E support schemes alike (see footnote 16). A very relevant one is whether the RES-E support is finally paid by electricity consumers (as it is generally the case) or by taxpayers (as in the Netherlands, see Couture and Gagnon, 2010). Those interactions would be even more limited if RES-E support was charged to taxpayers because, in contrast to RES-E support being paid by electricity consumers, an increase in RES-E support would not translate into a greater electricity price and, thus, the incentive to adopt EE measures would be

ARTICLE IN PRESS P. del Rı´o / Energy Policy 38 (2010) 4978–4989

/MWh

4988

MCre

PTGC(0)

Appropriate penalty Low penalty

PTGC(1)

Q1

Q0

MWh

Fig. 4. Analysing the impact of EE measures with a low penalty. Source: own elaboration.

lower. Therefore, this secondary effect of RES-E support (charged to electricity consumers) on EE should be taken into account when deciding who should bear the financial burden of RES-E support.

7. Conclusions Traditionally, policy analysis of climate policy instruments has been based on an individual approach, i.e., separated from other measures with which they are likely to interact. Considering the effects that instruments have when combined is a relevant research area. The functioning of those combinations depends on how instruments interact which, in turn, is affected by their design elements. If the ‘‘devil is in the detail’’, then careful ex-ante assessment of the impact of different instruments and design choices on those interactions is recommendable. This paper has focused on the interactions between RES-E and EE support schemes. The results show that: 1. The interactions between RES-E and EE support are generally modest because these two instruments have different scopes and no point of direct interaction. Notwithstanding, this should be confirmed by modelling work. 2. The most relevant interactions occur when EE support is added to RES-E promotion, especially when TGCs with a relative quota are implemented. In contrast, the interactions in the opposite direction (i.e., from RES-E support to EE support) are more limited. 3. Different instruments for RES-E support and design elements affect the results of the interactions differently. Particularly, such interaction might be regarded as more conflictive when a TGC scheme and a relative quota are adopted (rather than a FIT). In that case, EE measures put a downward pressure on electricity demand, reducing the absolute requirement for RESE. In the case of FITs, there is no impact of EE on RES-E investment when a fixed tariff is adopted and a lower impact than with TGCs when a fixed premium is implemented. The impact of other relevant design elements on the interactions have also been analysed in this paper. 4. Some conflicts in the interactions between EE support and RES-E generation can be mitigated through changes of instruments or design elements or through coordination between instruments. For example, the aforementioned problem with the relative quota could be circumvented by transforming the absolute RES-E target into a relative one, setting a minimum TGC price, replacing the TGC with a FIT or adapting the quota accordingly before the EE measures are implemented.

5. Indeed, what might be regarded as conflictive in the interactions might not be so problematic when the whole picture is taken into account, i.e., when different goals and different instruments targeting those goals and, in particular, costeffective increases in RES-E and EE are considered. For example, EE facilitates RES-E target achievement, which is particularly welcome with stringent RES-E targets. In turn, the positive impact of RES on the uptake of EE measures suggests that increasing the stringency of RES-E targets can contribute to the achievement of the EE target.36 6. The above suggests that the focus should not be on the functioning of specific instruments with respect to one specific criteria, but on the functioning of the whole policy mix and the interactions leading to conflicts and synergies with respect to several objectives and criteria in this mix. 7. Therefore, the regulator should seek some coordination between the respective targets/instruments, taking into account their interactions. The focus should probably shift from the design of specific instruments to the appropriate design of instrument mixes. Of course, there might be a problem when instruments belong to different territorial/administrative levels but, at the national level, the interaction between RESE and EE instruments should be considered when setting targets in both realms. Although this paper has identified the most direct effects of the interactions on electricity-sector variables, this should be followed by quantitative analysis. In addition, future research could address the interaction effects of some design elements which have not been considered in this paper. Furthermore, while this paper has focused on RES-E instruments, the impact of different EE instruments and design elements should also be analysed. Finally, an interesting topic for future research is whether and how the results of the interactions would change when the funding of RES-E and EE support is charged to the public budget (instead of being paid by electricity consumers in their bills).

Acknowledgements Financial support from CSIC (proyecto intramural) is gratefully acknowledged. 36 Recent simulations with the GREEN-X model confirms that ambitious RES-E support (to achieve the European 20% RES target) needs to be accompanied by a strong EE policy (see Resch et al., 2009). Indeed, in the new RES Directive, EE is a key criteria for RES target achievement, since the RES share is set as a percentage of gross final energy demand.

ARTICLE IN PRESS P. del Rı´o / Energy Policy 38 (2010) 4978–4989

References Amundsen, E., Baldursson, F., Mortensen, J., 2006. Price volatility and banking in green certificate markets. Environmental & Resource Economics 35, 259–287. Bertoldi, P., Huld, T., 2006. Tradable certificate for renewable electricity and energy saving. Energy Policy 34 (2), 212–222. Bertoldi, P., Rezessy, S., Oikonomu, V., Boza-Kiss, B., 2009. Feed-in tariff for energy saving: thinking of the design. ECEEE Summer Study Proceedings. Boonekamp, P., 2006. Actual interaction effects between policy measures for energy efficiency—a qualitative matrix method and quantitative simulation results for households. Energy 31 (14), 2848–2873. Butler, L., Neuhoff, K., 2008. Comparison of feed-in tariff, quota and auction mechanisms to support wind power development. Renewable Energy 35, 1854–1867. Child, R.L.angniss, O. Klink, J., Gaudioso, D., 2008. Interactions of white certificates with other policy instruments in Europe. Energy Efficiency 1, 283–295. CNE, 2004. Report 4/2004. Madrid, /http//www.cne.esS. Couture, T., Gagnon, Y., 2010. An analysis of feed-in tariff remuneration models: implications for renewable energy investment. Energy Policy 38, 955–965. De Jonghe, C., Delarue, E., Belmans, R., D’haeseleer, W., 2009. Interactions between measures for the support of electricity from renewable energy sources and CO2 mitigation. Energy Policy 37 (11), 4743–4752. Del Rı´o, P., 2007. The interaction between emissions trading and renewable electricity support schemes. An overview of the literature. Mitigation and Adaptation Strategies for Global Change 12 (8), 1363–1390. Del Rı´o, P., 2008. Ten years of renewable electricity policies in Spain: an analysis of successive feed-in tariff reforms. Energy Policy 36 (8), 2917–2929. Del Rı´o, P., Gual, M.A., 2004. The promotion of green electricity in Europe: present and future. European Environment Journal 14, 219–234. Del Rı´o, P., Herna´ndez, F., Gual, M.A., 2005. The Implications of the Kyoto project mechanisms for the deployment of renewable electricity in Europe. Energy Policy 33 (5), 2010–2022. European Commission, 2008. The support of electricity from renewable energy sources Commission staff working document. Accompanying document to the Proposal for a Directive of the European Parliament and of the Council on the promotion of the use of energy from renewable sources {COM(2008) 19}. SEC(2008) 57. Brussels. Farinelli, U., Johansson, T., McCormick, K., Mundaca, L., Oikonomou, V., Ortenvik, M., Patel, M., Santi, F., 2005. White and Green: comparison of market-based instruments to promote energy efficiency. Journal of Cleaner Production 13 (10–11), 1015–1026. Finon, D., Perez, D., 2007. The social efficiency of instruments of promotion of renewable energies: a transaction-cost perspective. Ecological Economics 62, 77–92. Harmelink, M., Nilsson, L., Harmsen, R., 2008. Theory-based policy evaluation of 20 energy efficiency instruments. Energy Efficiency 1, 131–148. IEA, 2008. Deploying renewables. Paris. Intergovernment Panel on Climate Change (IPCC), 2007. Fourth Assessment Report. Working Group III. Geneve. Jensen, S.G., Skytte, K., 2002. Interactions between the power and green certificate markets. Energy Policy 30 (5), 425–435. Jensen, S.G., Skytte, K., 2003. Simultaneous attainment of energy goals by means of green certificates and emission permits. Energy Policy 31 (1), 63–71. Klink, J. Langniss, O., 2006. White certificate schemes and (national) green certificate schemes. EuroWhiteCert Project, Work Package 3.2 Task Report. Konidari, P., Mavrakis, D., 2007. A multi-criteria evaluation method for climate change mitigation policy instruments. Energy Policy 35 (12), 6235–6257.

4989

Linares, P., Santos, F.J., Ventosa, M., submitted for publication. Interactions of carbon reduction and renewable support policies in electricity markets: a review of existing results and some recommendations for a coordinated regulation. Climate Policy. Lipp, J., 2007. Lessons for effective renewable electricity policy from Denmark, Germany and the United Kingdom. Energy Policy 35, 5481–5495. Mendonca, M., Jacobs, D., 2009. Feed-in tariffs go global: policy in practice. Renewable Energy World 12 (4), 1–6. Mitchell, K., Connor, P., 2004. Renewable energy policy in the UK 1990–2003. Energy Policy 32, 1935–1947. Morthorst, P.E., 2003. A green certificate market combined with a liberalised electricity market. Energy Policy 31, 1393–1402. NERA, 2005. Interactions of the EU ETS with Green and White Certificate Schemes. NERA Economic Consulting, London. Nielsen, L., Jepessen, T., 2003. Tradable Green Certificates in selected European countries—overview and assessment. Energy Policy 31 (1), 3–14. Oikonomou, V., Jepma, C.J., 2008. A framework on interactions of climate and energy policy instruments. Mitigation and Adaptation Strategy for Global Change 13, 131–156. Oikonomou, V., Patel, M., 2004. An inventory of innovative policies and measures for energy efficiency, Phase I of the EU SAVE ‘‘White and Green’’ project, Copernicus Institute, NWS-E-2004-113, Utrecht University. Ragwitz, M., Held, A., Resch, G., Haas, R., Faber, T., Huber, C., Morthorst, P.E., Jensen, S., Coenraads, R., Voogt, M., Reece, G., Konstantinaviciute, I., Heyder, B., 2007. Assessment and optimisation of renewable energy support schemes in the European electricity market. Final Report of the project OPTRES. Supported by the European Commission (D.G. Energy and Transport), Brussels. Reinaud, J., 2003. Emissions Trading and its Possible Impacts on Investment Decisions in the Power Sector. IEA Information Paper, Paris. Resch, G., Rawgitz, M., Panzer, C., Haas, R., 2009. 20% RES by 2020. European Congress of the International Association for Energy Economics. Viena. Sijm, J., 2003. Interaction of the EU emissions trading directive with climate policy instruments in the Netherlands. Policy Brief INTERACT project. ECN, Amsterdam. Sorrell, S., 2003. Back to the drawing board? Implications of the EU Emissions Trading Directive for U.K. Climate Policy. Executive summary. EU-funded project Interaction in EU Climate Policy (INTERACT). University of Sussex at Brighton, UK. Sorrell, S., 2004. Barrier busting: overcoming barriers to energy efficiency. In: Sorrell, S., Schleich, J., O’Malley, E., Scott, S. (Eds.), The Economics of Energy Efficiency. Edward Elgar, Cheltenham, pp. 287–316. Sorrell, S., Schleich, J., O’Malley, E., Scott, S. (Eds.), 2004. The Economics of Energy Efficiency. Edward Elgar, Cheltenham. Sorrell, S., Harrison, D., Radov, D., Klevnas, P., Foss, A., 2009. White certificate schemes: economic analysis and interactions with the EU ETA. Energy Policy 37, 29–42. Schleich, J., 2009. Barriers to energy efficiency: a comparison across the German commercial and services sector. Ecological Economics 66 (7), 2150–2159. UKERC, 2007. The rebound effect: an assessment of the evidence for economywide energy savings from improved energy efficiency. Unger, T., Ahlgren, E., 2005. Impacts of a common green certificate market on electricity and CO2 emission markets in the Nordic countries. Energy Policy 33, 2152–2163. Wang, Y., 2006. Renewable electricity in Sweden: an analysis of policy and regulations. Energy Policy 34, 1209–1220.