Feed-In Tariffs and Quotas for Renewable Energy in Europe

Forum

FEED-IN TARIFFS AND sures. Up to now, these policies have been implemented exclusively on a national level and aim to QUOTAS FOR RENEWABLE fulfil the national targets as set in the RES-E direc- ENERGY IN EUROPE* tive. However, based on the currently implemented policies, these targets will most likely not be met in the majority of countries, which indicates that RES- E support systems are still not designed in a suit- GUSTAV RESCH**, able way. MARIO RAGWITZ***,

ANNE HELD***, Evaluation of policy instruments for promoting THOMAS FABER** AND renewable electricity from a historical perspective

REINHARD HAAS** Classification of policy instruments and development of RES-E policies in the EU enewable electricity has increased significantly Rin recent years on a global scale and especially Within this study, the assessment of direct regulatory within Europe. A major reason for this development promotion strategies is carried out by focusing on a at the European level is the national support strate- comparison between price-driven (e.g. FITs) and gies triggered by Directive 2001/77/EC on renewable quantity-driven (e.g. quotas based on TGCs) strate- energies in the electricity sector (European Par- gies, which can be defined as follows: liament and Council 2001), which set the renewable energy sources (RES-E) target of 21 percent at the Feed-in tariffs (FITs) are generation-based, price- EU-25 level for the year 2010 and specified corre- driven incentives. The price that a utility or supplier sponding targets for all 25 member states. All EU or grid operator is legally obligated to pay for a unit member states have introduced policies to support of electricity from RES-E producers is determined the market introduction of RES-E and most of them by the system. Thus, a federal (or regional) govern- have started to improve the corresponding adminis- ment regulates the tariff rate. It usually takes the trative framework conditions (e.g. planning proce- form of either a fixed amount of money paid for dures, grid connection) as well. The market diffusion RES-E production, or an additional premium on top of new renewable energy technologies has increased of the electricity market price paid to RES-E pro- significantly over the last decade. The existing sup- ducers. Besides the level of the tariff, its guaranteed port instruments encompass feed-in tariffs (FITs), duration represents an important parameter when quota-based tradable green certificates (TGCs), evaluating the actual financial incentive. FITs allow investment grants, tender procedures and tax mea- technology-specific promotion and acknowledge future cost-reductions by applying dynamically decreasing tariffs. * This assessment of the effectiveness and economic efficiency of support schemes for renewable electricity was conducted for the European Commission, DG TREN within the European research project OPTRES (www.optres.fhg.de). For a detailed discussion of Quota obligations based on Tradable Green Certi- the above illustrated topic we refer to Ragwitz et al. 2007. ficates (TGCs) are generation-based, quantity-driven The authors and the whole project consortium gratefully acknowledge the financial and intellectual support of this work provided by instruments. The government defines targets for the Intelligent Energy for Europe – Programme. In particular, spe- cial thanks go to the project officers Beatriz Yordi, DG TREN, and RES-E deployment and obliges a particular party of Ulrike Nuscheler, IEEA. ** Vienna University of Technology, Institute of Power Systems the electricity supply-chain (e. g. generator, whole- and Energy Economics, Energy Economics Group, Vienna, Austria. saler or consumer) with their fulfilment. Once de- *** Fraunhofer Institute Systems and Innovation Research, Karls- fined, a parallel market for renewable energy certifi- ruhe, Germany. E-mail of lead author: [email protected]. cates is established and their price is set following

CESifo DICE Report 4/2007 26 Forum demand and supply conditions (forced by the obliga- electrical power. The definition of effectiveness used tion). Hence, for RES-E producers, financial support in this analysis is given in the following equation: may arise from selling certificates in addition to the revenues from selling electricity on the power mar- i i i G G ket. In principle, technology-specific promotion is E = n n1 n ADD POT i also possible in TGC systems. But it should be noted n that separate markets for different technologies will i En Effectiveness indicator for RES lead to much smaller and less liquid markets. technology i for the year n

i Gn Existing normalised electricity generation by RES Figure 1 shows the evolution of the main support technology i in year n ADD POT i instrument for each country. Only 8 of the 15 coun- n Additional generation potential of RES tries regarded did not experience a major policy shift technology i in year n until 2020 during the period 1997–2006. The current discussion within EU member states focuses on the comparison This definition of effectiveness has the advantage of of two opposed systems, the FIT system and the being unbiased with regard to the available potential quota regulation in combination with a TGC-mar- for individual technologies in a specific country. ket. The latter have replaced existing policy instru- Member states need to deploy RES-E capacities ments in some European countries, such as Belgium, proportional to the given potential in order to Italy, Sweden, the UK and Poland. Other policy demonstrate the comparable effectiveness of their instruments, such as tender schemes, are no longer instruments. This appears to be a meaningful used in any European country as the dominating approach since the member state targets, as deter- policy scheme. However, there are instruments like mined in Directive 2001/77/EC, are also mainly production tax incentives and investment incentives based on the realisable generation potential of each which are frequently used as supplementary instru- country. ments. Only Finland and Malta apply them as their main support scheme. Figure 2 shows the average annual effectiveness indicator for wind onshore electricity generation for Effectiveness of policy instruments 1998–2005 for EU-15 countries. Several findings can be derived from these figures. Firstly, the three mem- The effectiveness of a policy for renewable electrici- ber states showing the highest effectiveness during ty is based on its ability to increase the generation of the considered period – Demark, Germany, and Figure 1 Spain – applied fixed feed-in tariffs during the entire period EVOLUTION OF THE MAIN POLICY SUPPORT SCHEMES IN EU-15 1998–2005 (with a relevant sys- MEMBER STATES tem change in Denmark in 2001). The resulting high investment 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 All RES-E security as well as low adminis- AT technologies All RES-E trative barriers stimulated a BE technologies All RES-E strong and continuous growth in DK technologies All RES-E Feed-in tariff wind energy during the last FI technologies Wind decade. It is often claimed that FR Bioenergy Quota / TGC PV the high level of the feed-in tar- All RES-E DE technologies iffs is the main driver for in- Tender All RES-E GR technologies vestments in wind energy, espe- All RES-E Tax incentives / IE technologies Investment cially in Spain and Germany. Wind grants However, as will be shown in the IT Bioenergy PV section below, the tariff level is All RES-E Change of the LU technologies system not particularly high in these two All RES-E NL technologies countries compared with the All RES-E Adaptation of PT technologies the system other countries analysed here. All RES-E ES technologies This indicates that a long-term All RES-E SE technologies and stable policy environment is All RES-E UK technologies actually the key criterion for the

27 CESifo DICE Report 4/2007 Forum

Figure 2 higher than the costs of generation. The reasons for the higher EFFECTIVENESS INDICATOR FOR WIND ONSHORE average effectiveness indicator 1998–2005 support level expressed by the % 10 current green certificate prices include still immature TGC 8 markets, the non technology- specific design of the currently 6 applied TGC-systems as well as

4 the higher risk premium re- quested by investors. In the case 2 of Spain and Germany, the support level indicated in Figure 3 0 AT BE DK FI FR DE GR IE IT LU NL PT ES SE UK EU15 appears to be above the average Feed-in-tariffs Quota/TGC Tender Tax incentives/Investment grants level of generation costs. How- Note: This figure depicts the effectiveness indicator for wind onshore electricity in the period 1998 to 2005 ever, the low cost potentials in the EU-15 showing the relevant policy schemes during this period. have already been exploited in these countries due to recent success of developing RES-E markets. As can be success in market growth. Therefore a level of sup- observed in a country like France, high administra- port that is moderately higher than average costs tive barriers can significantly hamper the develop- seems to be reasonable. ment of wind energy even under a stable policy environment combined with reasonably high feed-in tar- Expected revenues and profits for investors iffs. In order to correlate the effectiveness of an instrument Economic efficiency from society’s point-of-view with the efficiency of support as defined in the previ- ous section, the levelised profit of potential wind ener- In order to analyse the economic efficiency of sup- gy investments was calculated for Austria, Belgium, the port from a historical perspective we compare the Czech Republic, France, Germany, Ireland, Italy, level of support in the case of wind energy onshore Lithuania, Spain, Sweden and the UK for the year and the corresponding costs of electricity genera- 2004. Thus, calculations are based on the effective sup- tion. Based on this definition the analysis shows (see port conditions in each country during 2004. Figure 3) that for many countries the support level and the generation costs are very close. Countries By plotting the effectiveness versus the levelised with costly potentials frequently show a higher sup- profit as shown in Figure 4, the correlation between port level. A clear deviation from this rule can be the levelised profit for investments and the level of found in the three quota systems in Belgium, Italy effectiveness attained by the support instrument in and the UK, where support is presently significantly the respective year is analysed.

Figure 3 In Figure 4, the expected lev- COMPARISON OF SUPPORT LEVELS AND GENERATION COSTS FOR elised profits as well as the effec- IND NSHORE EUR/MWh W O tiveness show a broad spectrum 160 for the countries under consider- Minimum to average generation costs 140 Average to maximum support level ation. It should be pointed out 120 that the different instruments 100 have different levels of maturity 80 and that policy schemes in some 60 countries – in particular quota 40 obligation systems – are still in a 20 transitional phase. It is striking 0 that Italy, the UK and Belgium, AT BE DK FI FR DE GR IE IT LU NL PT ES SE UK which transformed their markets Note: This figure depicts support level ranges (average to maximum support) for direct support of wind by introducing quota systems as onshore in EU-15 Member States (average tariffs are indicative) compared to the long-term marginal generation costs (minimum to average costs). the main support instrument

CESifo DICE Report 4/2007 28 Forum

Figure 4 Prospective analysis based on the model Green-X EFFECTIVENESS INDICATOR VERSUS LEVELISED PROFIT FOR WIND ONSHORE effectiveness indicator in % In this section we aim to signpost 20 ES- ES- the way forward by presenting a Fixed Price Market Option prospective analysis of possible 15 future RES-E support options at DE the European level. The effec- 10 IE tiveness and efficiency of sup- AT port schemes is based on the 5 BE BE Flanders IT UK results obtained from simulation Wallonia SE FI CZ FR runs using the Green-X model 0LT 01234567 (www.greenx.at). This tool en- expected annuity, EUR Cent/KWh ables us to make a comparative Feed-in-tariffs Quota/TGC Tender Tax incentives/Investment grants and quantitative analysis of the Note: This graph shows a possible levelised profit per unit of electricity generated by an investment in wind future deployment of RES up to onshore in 2004. 2020 in all energy sectors (i.e. electricity, heat and transport) between 1999 and 2002, are characterised by expect- based on applied energy policy strategies in a ed high levelised profits but low effectiveness. The dynamic context. Geographically the assessment high levelised profit results in particular from the refers to the European Union as of 2006, comprising extrapolation of the presently observed certificate 25 member states (EU-25). prices. The results show that certificate systems lead to higher producer revenues than FITs, which com- Figure 5 indicates the investigated scenario paths pensate for high investment risks. Furthermore, the and the resulting RES-E deployment – comprising a recent development of certificate prices does not business-as-usual (BAU) case based on a continua- show any decreasing tendency. On the other hand, tion of current national support schemes (BAU), a countries with FITs seem to be typically more effec- national improvement and a harmonisation of RES- tive at generally moderate levelised profits per unit E support at the European level based on either of electricity generated. The fact that expected prof- technology-specific support, i.e. a feed-in tariff sys- itability from the investor’s perspective is signifi- tem with technology-specific differentiated tariffs, or cantly lower for FITs is directly linked with a higher uniform support, i.e. a quota obligation based on efficiency of this strategy because additional costs TGCs commonly applied for all RES-E options. for consumers are lower. Results with regard to non-harmonised conditions – BAU & Figure 5 improved national policies- OVERVIEW OF INVESTIGATED CASES scenario

Historical development 1156 TWh (improved national In 2004 the total amount of RES- 35% Indicative RES-E Target (2010) & harmonised policies) Introduction of harmonised policies (2015) E generation within the EU-25 30% BAU-forecast Strengthened national policies was around 460 TWh, corre- 25% Technology-specific sponding to a share of about deployment harmonised FIT scheme 951 TWh 20% Non technology-specific (BAU) 15 percent of gross electricity harmonised TGC system RES-E 15% demand. Without any changes to the current support schemes of 0% the various member states, RES- 1990 1995 2000 2005 2010 2015 2020 E would achieve a demand share No Harmonisation Harmonisation in 2015 of 18.2 percent in 2010 at EU-25 Business-as-usual Improved national Technology-specific Non technology- (Bau) policies support specific support level. If RES-E support is accom- Continuation of current Efficient & effective Feed-in tariffs Quota obligation based panied by energy efficiency mea- national policies national policies -harmonised on TGCs – harmonised up to 2020 sures as assumed for a sensitivity variant to the BAU case, a higher

29 CESifo DICE Report 4/2007 Forum

Figure 6 ing RES-E plants installed prior to 2005, which account for COMPARISON OF RES-E GENERATION AT EU25 LEVEL 431 TWh (equal to a share of 2005 TO 2020 RES-E - electricity generation 45 percent in total RES-E gen- BAU case "improved national policies"-variant eration) by 2020 in the BAU

TWh/year TWh/year case. “Improved national poli-

1 200 Wind offshore 1 200 cies” will induce a much higher Wind onshore deployment of new RES-E in 1 000 Tide & wave 1 000 Solar thermal electricity the investigated period: by 2020 800 Photovoltaics 800 Hydro small-scale this will amount to 725 TWh 600 Hydro large-scale 600 from new RES-E plants in- Geothermal electricity 400 Biowaste 400 stalled between 2005 and 2020, Solid biomass 200 corresponding to 63 percent of Biogas 200 Total stock (end of 2004) the total RES-E generation of 0 0 1156 TWh. 2005 2007 2009 2011 2013 2015 2017 2019 2005 2007 2009 2011 2013 2015 2017 2019

Figure 7 illustrates the required demand share of 18.8 percent is feasible in 2010. By consumer expenditure for both cases investigated at 2020, these differences will become more apparent: a the EU-25 level due to the underlying national RES- share of 23.6 percent is projected for the default E policies and the corresponding induced RES-E BAU case, whilst deployment in relative terms is deployment. In this context, the consumer / societal 27 percent for BAU with accompanying DSM. expenditure due to the support for RES-E represents a net value based on the direct costs of apply- In contrast, it would be feasible to meet the Euro- ing a certain support scheme. This figure also illus- pean target as set by the RES-E Directive by im- trates both the technology-specific shares of new proving the support conditions for RES-E rigorous- RES-E plants and the expenditures associated with ly and immediately in all EU countries, including a the stock of existing RES-E plants (indicated by the removal of non-financial deficiencies and the imple- blue area). mentation of energy efficiency measures. In the “improved national policies” case, a RES-E share of The required consumer expenditures will increase 20.9 percent is reached in 2010, rising to 34.1 percent steadily over the next ten years with BAU. In relative in 2020. terms, expressing the expenditures as a premium per MWh total demand, these are projected to rise from The dynamic development of RES-E generation in a level of 2.1 EUR/MWhDEMAND in 2005 up to about both cases is depicted in absolute terms at the EU-25 5.0 EUR/MWhDEMAND in the final years 2019 and level in Figure 6. This graph illustrates the tech- 2020. Obviously, within the “improved national poli- nology-specific deployment for new RES-E plants and shows Figure 7 the total RES-E stock (indicated by the blue area) comprising all COMPARISON OF NECESSARY CONSUMER EXPENDITURE plants installed up to the end of AT EU25 LEVEL, 2005 TO 2020 yearly transfer cost for consumers due to RES policy 2004. If currently implemented RES-E policies are maintained, BAU case "improved national policies"-variant as assumed in the BAU case, billion EUR/year billion EUR/year 28 Wind offshore 28 the total amount of RES-E gen- Wind onshore 24 24 eration will increase from 460 Tide & wave 20 Solar thermal electricity 20 TWh in 2004 to about 951 TWh Photovoltaics 16 Hydro small-scale 16 in 2020. This 2020 figure com- Hydro large-scale 12 12 prises almost equal contribu- Geothermal electricity Biowaste 8 8 tions of new RES-E installations Solid biomass (from 2005 to 2020) in the order 4 Biogas 4 Total stock (end of 2004) of 520 TWh (55 percent of total 0 0 2005 2007 2009 2011 2013 2015 2017 2019 RES-E) and the stock of exist- 2005 2007 2009 2011 2013 2015 2017 2019

CESifo DICE Report 4/2007 30 Forum cies” variant, characterised by a Figure 8 40 percent higher RES-E de- COMPARISON OF NECESSARY CUMULATED CONSUMER EXPENDITURE ployment in the investigated pe- FOR NEW RES-E riod 2005 to 2020, even greater Harmonised non technology-specific support 34.9 financial support is required to (from 2015 on) - neglecting risk achieve the ambitious RES-E tar- Harmonised non technology-specific support 36.6 get set for 2010. Accordingly, a (from 2015 on) steeper increase in expenditure in Harmonised technology-specific support 20.4 with less novel technologies (from 2015 on) the period up to 2017 occurs, cul- Harmonised technology-specific support 21.8 minating in a peak at 7.7 EUR/ (from 2015 on)

MWhDEMAND in 2017. Improved national policies 25.8

BAU - continuation 32.6 Harmonisation: Technology- of current national RES-E policies specific versus uniform support 0 5 10 15 20 25 30 35 40

EUR/MWhRES

Note: This depiction shows the necessary cumulated consumer expenditure (i.e. the cumulated present value Besides the above discussed na- (2005) of yearly transfer cost) due to the support of new RES-E (installed 2005 to 2020), expressed per MWh tional support options (i.e. BAU induced RES-E generation for the investigated cases. and “Improved national policies”) the following policy options at the European fer costs due to the promotion of new installations in level are investigated below: the observed period 2005 to 2020 as well as the resid- ual costs after 2020 – is shown in Figure 8 for the – Harmonisation of support based on a uniform investigated cases. This figure illustrates both the (non technology-specific) support scheme, i.e. a cost-efficiency and the effectiveness of RES-E sup- quota obligation based on TGCs commonly port options, expressing the cumulated consumer applied for all RES-E options. expenditures in specific terms, i.e. per MWh induced – Harmonisation of support based on a technology RES-E generation. The following conclusions are specific support scheme, i.e. a feed-in tariff system drawn from this diagram: with technology-specific differentiated tariffs. – The cumulated transfer costs for consumers are In addition, a further variant of each harmonised lowest when applying technology-specific support RES support case is also taken into consideration. harmonised throughout Europe achieved by Thereby, in case of technology-specific support it is applying feed-in tariffs. There are marginal differ- assumed that the support is limited to less novel ences between the two variants, i.e. by considering RES-E technologies, whilst in the case of non-tech- or neglecting novel RES-E options. nology-specific support the variant refers to the neg- – Improved national policies with a similar deploy- ligence of the investor’s risk (as commonly associat- ment of new RES-E result in slightly higher spe- ed with uncertain earnings in the TGC market). cific costs corresponding to an increase of +18 percent compared to the technology-specific sup- One target is assumed for future RES-E deployment port provided within a harmonised scheme in 2020 in all cases based on harmonised support in (including novel RES-E options). order to be able to compare the economic efficiency – Higher specific costs can be expected from con- of the different policy options – i.e. it is assumed that tinuing current RES-E support. With BAU, the about 1156 TWh have to be generated by RES-E at specific costs are 49 percent higher compared to the EU-25 level by 2020, similar to the “improved harmonised technology-specific support. It is national policies” case. Note that regarding har- worth mentioning that the overall deployment of monised support options a transition period is taken new RES-E is 29 percent lower with BAU than into account. Accordingly, new and improved har- with all other policy options. monised policies offering equal financial incentives – The most inefficient policy option in terms of throughout Europe are then applied to new RES-E costs is harmonised, but non technology-specific installations from 2015 onwards. support as provided by a uniform EU-wide TGC system, which results in much higher consumer A comparison of the cumulated consumer expendi- expenditures ranging from + 60 to + 68 per- ture for new RES-E installations – i.e. the total trans- cent compared to its technology-specific coun-

31 CESifo DICE Report 4/2007 Forum

terpart incl. novel RES-E options – depending specific support is implemented. In contrast, if har- whether the investor’s risk is neglected or taken monisation meant putting all the RES-E options in into account. one basket and giving equal support to all the RES- E technologies considered, then the accompanying consumer expenditures would increase significantly Conclusions if the RES-E target is ambitious. Consequently, a harmonised non technology-specific support would The empirical findings presented in this paper show decrease efficiency of support. that instruments which have proven to be effective also tend to be economically efficient. Feed-in systems, which are implemented in the majority of EU References member states, have initiated significant growth of European Parliament and Council (2001), Directive 2001/77/EC of renewable energy generation at moderate costs for September 27th 2001 on the Promotion of Electricity Produced from Renewable Energy Sources in the Internal Electricity Market. society. The main reason for this observation is the long-term price security of the system combined Ragwitz, M., A. Held, G. Resch, T. Faber, R. Haas, C. Huber, P.E. Morthorst, S.G. Jensen, R. Coenraads, M. Voogt, G. Reece, I. with technology diversification of support. Com- Konstantinaviciute and B. Heyder (2007), OPTRES: Assessment and Optimisation of Renewable Energy Support Measures in the pared to short-term trading in renewable certificate European Electricity Market, final report of the research project OPTRES of the European Commission, DGTREN, Intelligent markets, the intrinsic stability of feed-in systems Energy for Europe - Programme appears to be a key element for success. (Contract No. EIE/04/073/S07.38567), Karlsruhe, Germany.

The key criterion for achieving an enhanced future deployment of RES-E in an effective and efficient manner, besides the continuity and long-term stability of any implemented policy, is the technology specification of the necessary support. Concentrat- ing on only the currently most cost-competitive technologies would exclude the more innovative technologies needed in the long run. Furthermore, it would not be possible to achieve any moderate to ambitious RES-E target without considering these novel RES-E options. In other words technology neutrality may be cost-efficient in the short term but is more expensive in the long term.

Even in the short term, the producer profits involved in the promotion of RES-E as well as observable cost differences among cheap to moderate RES-E options suggest a diversification of support. Most of the European success stories of promoting RES-E over the past decades in an effective and economically efficient way were driven by feed-in tariffs, which are implemented in a technology-specific manner.

The results of the modelling exercise clearly indicate that the major part of possible efficiency gains can already be exploited by optimising RES-E support measures at the national level – about two thirds of the overall cost reduction potential can be attributed to optimising national support schemes. Further efficiency improvements at a considerably lower level (about one third of the overall cost reduction potential) are possible through an EU wide harmonisation of the support schemes provided that technology-

CESifo DICE Report 4/2007 32