Renewable Energy, Subsidies, and the WTO: Where has the ...(Munich), the 2014 EALE conference...

RENEWABLE ENERGY, SUBSIDIES, AND THE WTO: WHERE HAS THE ‘GREEN’ GONE?

Documents de travail GREDEG GREDEG Working Papers Series

Patrice BougetteChristophe Charlier

GREDEG WP No. 2014-20http://www.gredeg.cnrs.fr/working-papers.html

Les opinions exprimées dans la série des Documents de travail GREDEG sont celles des auteurs et ne reflèlent pas nécessairement celles de l’institution. Les documents n’ont pas été soumis à un rapport formel et sont donc inclus dans cette série pour obtenir des commentaires et encourager la discussion. Les droits sur les documents appartiennent aux auteurs.

The views expressed in the GREDEG Working Paper Series are those of the author(s) and do not necessarily reflect those of the institution. The Working Papers have not undergone formal review and approval. Such papers are included in this series to elicit feedback and to encourage debate. Copyright belongs to the author(s).

Groupe de REcherche en Droit, Economie, GestionUMR CNRS 7321

Renewable Energy, Subsidies, and the WTO:Where has the ‘Green’ Gone?∗

Patrice Bougette † Christophe Charlier ‡

GREDEG Working Paper No. 2014–20

This version: August 2015

Abstract

Faced with the energy transition imperative, governments have to decide aboutpublic policies to promote renewable electrical energy production and to protect do-mestic power generation equipment industries. These policies can generate trade fric-tions. For example, in the Canadian renewable energy dispute the EU and Japanclaimed that Feed in Tariff (FIT) programs in Ontario constitute discriminatory sub-sidies because of a local content requirement (LCR) clause that is incompatible withWorld Trade Organization obligations. This paper investigates this issue using an in-ternational quality differentiated duopoly model in which power generation equipmentproducers compete on price. FIT programs including those with a LCR are comparedfor their impacts on trade, profits, amount of renewable electricity produced, and wel-fare. When sales are taken into account, the results confirm discrimination. However,introducing a difference in the quality of the power generation equipment producedon both sides of the border has a moderating effect on the results. Finally, the resultsenable discussion of the question of whether environmental protection can be a reasonfor subsidizing renewable energy producers in light of the SCM Agreement.

Keywords: Feed-in tariffs, Subsidies, Local content requirement, Industrial policy,Canada – Renewable energy dispute, Trade policy.

JEL Codes: F18, L52, Q42, Q48, Q56.

∗The authors would like to thank two anonymous referees for their insightful comments and suggestions.The authors are also grateful to participants at the 2014 FAERE meeting (Le Havre), the 2014 JMA con-ference (Clermont-Ferrand), the FAERE contributed session “International Economics and Environment”of the 2014 AFSE conference (Lyon), the 2014 WCERE conference (Istanbul), the 2014 ETSG conference(Munich), the 2014 EALE conference (Aix-en-Provence), the GREDEG, OFCE, EconomiX, and CRESEseminars. Special thanks to Gwenole Le Velly, Sandra Rousseau, Francesco Vona, Frederic Marty, andYannick Perez for valuable suggestions.†Universite Nice Sophia Antipolis and GREDEG/CNRS. Email: [email protected]‡Corresponding author. Universite Nice Sophia Antipolis and GREDEG/CNRS. Address: GREDEG

CNRS, 250 rue Albert Einstein, 06560 Valbonne, France. Email: [email protected]

1 Introduction

Energy transition has important implications for international trade.1 The equipmentrequired for the generation of renewable energy sources is being produced by newly de-veloped industries (e.g., solar photovoltaic industry) that have emerged to address thechallenges of global warming, air pollution, and energy supply diversification. In the con-text of energy transition, the development of domestic industries to produce equipment forrenewable energy appears strategic. Governments have been tempted to design policies tosupport the development of these sectors. However, specific protection of domestic indus-tries could lead to trade policies that potentially violate the non-discrimination principlethat is the cornerstone of World Trade Organization (WTO) international law.

This situation has already arisen as evidenced by seven trade disputes2 brought to theWTO’s Dispute Settlement Body. These trade disputes account for 14.5% of the total ofdisputes occurring in the period September 2010 (date of the first of the seven disputes)to May 2013. Around two-thirds of those are disputes between newly industrialized andindustrialized countries. The nature of these seven trade disputes is mostly complaintsabout national programs that provide support for the development of renewable energyand preferential use of domestic over imported rival products.

One reason for the significant share of disputes related to energy is the increasinguse of Feed-In Tariff (FIT) programs to develop and sustain renewable energy sources.3

FIT programs provide guaranteed prices for renewable energy supplied to the grid throughlong-term contracts. The guaranteed price is set higher than the wholesale market price forenergy supplied from non-renewable sources. This mechanism helps to offset the highercosts faced by renewable energy producers. The objective is to foster investment andinnovation in renewable energy sectors by removing the cost disadvantage.4 FIT programshave emerged as one of the most popular green policies. For instance, between 1990and 2011, 23 EU members implemented FIT programs (Jenner et al., 2013, pp. 386).

1Energy transition in the EU has the objective of 20% renewable energy in overall European energyconsumption by 2020. Renewable energy sources include wind, solar, geothermal, wave, tidal, hydropower,biomass, landfill gases, sewage treatment plant gases, and biogas.

2In chronological order, these disputes are: Canada – Renewable Energy (DS412), China – Measuresconcerning wind power equipment (DS419), Canada – Feed-In Tariff Program (DS426), European Unionand a Member State – Certain Measures Concerning the Importation of Biodiesels (DS443), EuropeanUnion and Certain Member States – Certain Measures Affecting the Renewable Energy Generation Sector(DS452), India – Certain Measures Relating to Solar Cells and Solar Modules (DS456), and EuropeanUnion – Certain Measures on the Importation and Marketing of Biodiesel and Measures Supporting theBiodiesel Industry (DS459). For further details, see http://www.wto.org/english/tratop_e/dispu_e/dispu_status_e.htm

3For an up-to-date list of FIT programs by countries, see the IEA/IRENA database, http://www.iea.org/policiesandmeasures/renewableenergy/.

4See, e.g., Johnstone et al. (2010). Based on panel data for 25 countries over 26 years, they show thatFIT programs have a positive influence on solar energy patenting. The positive effect of price support oninnovation can be exacerbated when applied to competitive markets since subsidies can make market entrymore attractive and encourage more radical innovations (Nesta et al., 2014).

1

Also, more than 50 developing countries have adopted FIT programs or renewable energystandards (Steer, 2013).

The economic literature on FIT programs is mainly empirical, and investigates the im-pact of this policy in several countries (Chua et al., 2011; Huang and Wu, 2011; Antonelliand Desideri, 2014, e.g.). Despite some design and implementation issues, most studiespoint to the positive effects of FIT programs on the development of renewable energyproduction. Smith and Urpelainen (2014), in a study based on 26 industrialized countriesover 1979 to 2005, show that FIT programs have caused large increases in renewable elec-tricity generation. Jenner et al. (2013) is the first econometric study on the efficacy of FITprograms in Europe. The authors find that FIT policies have driven solar photovoltaiccapacity growth but argue that this effect is overstated by not controlling for country char-acteristics and specific designs. However, Borenstein (2012) among others, criticizes therelevancy and efficiency of FIT programs. Borenstein argues that the underlying marketfailure is due to underpricing of fossil fuel electricity not overpricing of renewable energy.FIT programs artificially decrease the price of energy and may discourage efficient energyconsumption. In this sense, FIT programs could lead to overconsumption of electricityand act as disincentives for energy efficiency. Therefore, the design of renewable energyprograms is of prime importance.

One design that has received great attention includes a local (or domestic) contentrequirement (LCR). A LCR means that the subsidy is “contingent. . . upon the use ofdomestic over imported goods” according to the WTO Subsidies and Countervailing Mea-sures (SCM) Agreement). With regard to renewables, all FIT programs in this settingmust include a minimum amount of locally produced goods and services. This specific de-sign has been used by both developing and developed countries (Wu and Salzman, 2013;Sahoo and Shrimali, 2013, e.g.). Beyond the energy transition objective, the purpose ofthe LCR is to foster job creation within the local area. However, there is no direct evidenceof these positive effects (see Charnovitz and Fischer (2015) for a discussion).

In Canada – Renewable Energy (DS412 and DS426), Japan and the EU complain aboutthe FIT program implemented by the Canadian Province of Ontario, which guarantees thepurchase price of wind and solar electricity production. Under the 2009 Green Energy Act,the FIT program was implemented to diversify Ontario’s supply-mix and to help replacethe generation capacities that would be lost by 2014 due to the closure of coal-fired facilitiesin Ontario. According to the plaintiffs, these programs make the award of a contractallowing guaranteed prices for renewable electricity subject to a LCR in the productionfacilities (for solar energy projects only). More precisely, to qualify for the FIT rate, 50%to 60% of power generation equipment (including services) should be manufactured in theProvince of Ontario. The complaints focus on discriminating subsidies. By benefitingonly electricity producers whose facilities meet the national content criterion, the LCR

2

policy treats imported and domestic power generation equipment asymmetrically, andgoes against the GATT 1994 Agreement principle of national treatment. By guaranteeinga purchase price for electricity produced from renewable sources higher than the wholesaleprice, the FIT program constitutes a subsidy within the meaning of the SCM Agreement.

This dispute has attracted the attention of legal scholars (Cosbey and Mavroidis,2014; Rubini, 2012; 2014; Shadikhodjaev, 2015). The main point discussed is how thedispute ruling highlights the possibility for WTO members to subsidize renewable energyproduction under the SCM Agreement. Most of the literature criticizes the reasoning of thePanel and the Appellate Body (AB) leading to the conclusion that claimants have failed toshow that FIT programs constitute a subsidy. This finding, in a case where a subsidy wouldbe obvious from an economic perspective, has been interpreted as introducing some legaluncertainty with regard to WTO subsidy laws. Economic analysis related to Ontario’sFIT program are presented in Yatchew and Baziliauskas (2011); Pirnia et al. (2011);Mabee et al. (2012); Dewees (2013); Qudrat-Ullah (2014); Winfield and Dolter (2014).The resulting evaluations emphasize its relative cost and suggests some improvement oralternative scenarios. With the notable exception of Pirnia et al. (2011) empirical studieson Ontario do not consider welfare issues. Edenhofer et al. (2013)’s survey of the economicsof renewable energy sources calls for an analysis of policy instruments designed to tacklethe welfare issue.

Our study is the first to present a theoretical model – with numerical simulations– that allows comparison of FIT programs with and without LCR. We contribute to atheoretical literature focusing on how environmental regulation in a downstream sectorimpacts on the upstream market for abatement equipment. For instance, under quantity-setting, Greaker and Rosendahl (2008) study whether a government should strategicallyimplement stringent national environmental regulation in order to give advantage to itsexporting pollution abatement industry. They show that this choice has ambiguous ef-fects on export, compared to a direct subsidy for R&D in the pollution abatement sector.However, their results show also that these two policies should be seen as complementary.In this context, and in relation to supporting investment in abatement technologies, Fis-cher et al. (2014) demonstrate that subsidizing upstream R&D would be preferred overdownstream subsidies such as price support. Upstream subsidies both reduce the price ofabatement technology and emissions leakages, and provide upstream firms with strategicadvantage.

In this article, which focuses on the quality dimension of upstream competition, weaddress the question of whether FIT programs (with and without LCR) confer benefitsto domestic power generation equipment suppliers at the expense of foreign suppliers.We show that upstream suppliers capture the advantages of a FIT program largely as asubsidy, and even foreign suppliers can benefit from it. Our results confirm that LCR is

3

discriminatory but reveal that under certain conditions, foreign suppliers may also gain.We inquire whether FIT programs (with and without LCR) necessarily increase the totalamount of renewable energy produced. We show also that the total amount of renewableenergy produced may decrease under a FIT program with LCR, which contradicts theenergy transition objective. Finally we investigate the consequences of FIT programsfor domestic welfare. We show that the impact on domestic welfare is ambiguous anddepends mostly on the social cost of carbon, as well as the social cost of public funds. Ourtheoretical results and simulations shed light on the Canada – Renewable Energy disputeruling and broader open research questions at the intersection of trade, energy, industrypolicy, and the environment. The same questions will arise in the context of an ongoingWTO dispute between the U.S. and India regarding India’s LCR-based support policiesfor solar energy.5

The rest of the paper is organized as follows. Section 2 presents the model. Section 3investigates the effects of FIT programs within this framework. Section 4 calibrates themodel to the Ontario’s data and provides numerical simulations. Section 5 discusses theresults and offers some concluding remarks and policy implications.

2 The Model

We consider a standard vertical product differentiation duopoly model in which two manu-facturers – domestic and foreign – produce renewable energy power generation equipment.Take the example of solar panels that convert sunlight to electricity. Pillai and McLaugh-lin (2013) provide an analysis of the solar module industry and show that it is a far fromperfectly competitive industry. Also, due to the importance of R&D and new technology,the industry is largely protected by patents. Section 2.1 presents the two markets in themodel (an upstream manufacturing duopoly, and the renewable electricity market) andthe regulator’s objective. Section 2.2 resolves the equilibrium and analyzes the effects ofFIT programs.

2.1 The power generation equipment and electricity markets

Power generation equipment differs only in quality. There are two rival technologies inthe solar industry – crystalline silicon modules, and thin film modules.6 Equipment pur-chasers are renewable energy producers. The domestic renewable energy supply consistsof numerous individual small producers.7 These producers take the electricity price and

5In this dispute (DS456, see Footnote #2), the panel was composed on 24 September 2014.6For further details on solar technologies, see, e.g., Sahoo and Shrimali (2013).7This assumption allows us to take account of micro FIT contracts used in the Province of Ontario.

These contracts refer to micro renewable energy generation programs producing less than 10kW. SeeYatchew and Baziliauskas (2011) for further details.

4

the price of power generation equipment as given. Also, an impartial regulator aims atincreasing the domestic supply of electricity produced from renewable energy sources.

Power generation equipment producers. We assume that power generation equip-ment is produced by two– domestic and foreign– firms competing in the domestic market.8

The two manufacturers differ in output quality and production costs. The costs of pro-ducing power generation equipment are assumed to be a quadratic function of the powergeneration equipment quality (the power generation equipment’s marginal product Γ) andactual demand Q: ci (Γi, Qi) for i = d, f where c′i(Γ) > 0, c′i(Q) > 0, c′′i (Γ) > 0, andc′′i (Q) > 0 (see, e.g., Moorthy (1988)). Whenever necessary, we will use the following form

for the cost function

ci (Γi, Qi) = aΓ2i Q

2i , for i = d, f with a > 0 . (1)

Power generation equipment producers d and f compete on price a la Bertrand in themarket. Prices are denoted kd and kf (ki > 0 with i = d, f).

Renewable electricity producers. In the domestic economy, a continuum of potentialrenewable electricity producers is considered. The size of each generating facility is as-sumed to be fixed (normalized to 1 unit of power generation equipment). The productionfrom each generating facility depends on two elements: the marginal product of powergeneration equipment Γ (which is positive Γ ∈]0, ∞[) and an exogenous parameter θ de-noting the difficulty involved in producing electricity due to natural conditions (e.g., lowwind speeds, sunshine hours, etc.).

Electricity producers are heterogeneous in θ, which is a cost index that is uniformly dis-tributed on [0, 1]. These costs are efficiency losses for a particular installation; for example,θ represent the share of time without full sun or lost solar potential more generally. Theelectricity production of a given plant is assumed to decrease with θ. The marginal revenueof one producer running a generating facility using equipment with marginal product Γ ispg (1− θ) Γ, where pg is a guaranteed price set by the regulator above the market price p(see infra). One unit of power generation equipment costs k. Other costs of production(variable cost) are assumed to be negligible. Under these assumptions, the (marginal)profit of a generating facility is given by the expression:

pg (1− θ) Γ− k . (2)

A producer’s willingness-to-pay for one unit of power generation equipment with a8We only consider the domestic market. The impact of domestic FIT programs on foreign markets

for electricity and equipment is left to future research. In this more general framework, the effect woulddepend a priori on the type of policy chosen in the foreign market.

5

marginal product Γ is pg (1− θ) Γ. This willingness-to-pay increases in Γ and decreases inθ. Given pg, θ, (Γd, kd), and (Γf , kf ), a renewable electricity producer can choose amongproducing one unit of electricity using power generation equipment d, producing one unitof electricity using power generation equipment f , or not producing.

Two critical levels θd and θf , representing the maximum rates of efficiency loss, can bedefined equating the profit of producing electricity to zero:9

θi = 1− kipgΓi

for i = d, f . (3)

Using power generation equipment i = d, f is profitable as soon as θ ≤ θi.10

Let θ be the level at which the marginal electricity producer is indifferent about powergeneration equipment quality Γd at price kd and power generation equipment quality Γfat price kf , given a guaranteed electricity price pg:

θ = 1− kd − kfpg (Γd − Γf ) . (4)

The critical values of θ defined in (3) and (4) allow definition of demand for domesticand foreign equipment presented below. In the rest of the paper we assume that Γd > Γf .This assumption avoids the situation where a domestic subsidy is given to the less effi-cient technology under a FIT program with LCR.11 When the domestic power generationequipment is more efficient than the foreign equipment (Γd > Γf ), producers for which θ

is smaller than θ prefer the domestic power generation equipment, and those for which θ isbigger than θ prefer the foreign power generation equipment. Note that θ can be outsideof [0, 1], in which case one technology dominates.

A result for the position of θ, θd and θf can be stated and will be useful to analyze thestructure of demand for the two types of power generation equipment.

Lemma 1 If the marginal product of power generation equipment d is higher, Γd > Γf ,then θf > θd when θd > θ, whereas θf < θd when θd < θ.

The proof is given in the Appendix �

The structure of the demand for power generation equipment in the domestic marketcan now be specified if Γd > Γf . The following settings could emerge: i) no powergeneration equipment is purchased because electricity production is not profitable, ii)

9Note that these critical values are always smaller than 1 and can be negative.10Insufficient rate of return is therefore the only reason for non-adoption.11This assumption is not necessarily realistic. However, it allows for a scenario where the FIT program

with LCR should not be automatically rejected. For further information on the quality differences of solartechnologies, see http://sroeco.com/solar/table/.

6

either domestic or foreign power generation equipment is produced (monopoly cases), iii)both types of power generation equipment are purchased, with the critical values for θordered as follows: 0 < θ < θd < θf . In the rest of the paper we consider only thisduopoly case as illustrated in Fig. 1.

10 θdθ θf

Figure 1: The duopoly market structure

Therefore, demand for power generation equipment is Qd = θ and Qf = θf − θ. Notethat in this market setting θf denotes renewable electricity market coverage. Becauseθi < 1 (i = d, f), the market is never fully covered.

The regulator. An impartial regulator aims at maximizing total domestic welfare bysetting guaranteed prices for renewable energy.12 In the domestic economy, we considerthat the production of renewable energy compared to fossil fuels or nuclear generations issmall.13 Renewable generation is assumed (realistically) to cost more than non-renewablegeneration. Furthermore, the structure of demand for renewable energy power generationequipment above shows that there are always potential electricity producers that choosenot to produce at all. In this context, and because of the objective to initiate energytransition, the regulator aims to foster renewable electricity supply in the market.

The domestic market electricity price is assumed exogenous and given. This pricemay be too low to incite sufficient numbers of renewable electricity producers to enter themarket. The regulator aiming at developing the supply of renewable energy in the market,offers a higher price, pg > p, implementing a FIT program. The difference between theguaranteed price and the market price, pg − p, can therefore be considered a subsidy. Thesituation where pg = p corresponds to the no-policy setting (hereafter the unregulatedcase).

We assume that an increase in the amount of renewable energy decreases the productionof fossil electricity. Unlike fossil fuel generation, renewable energy is assumed to implyno external costs. As a consequence, developing renewable electricity (i.e., reducing non-renewable generation) with the subsidy pg − p always decreases the external cost of non-renewable energy. In the model, this relation is considered assuming that the amount

12We do not address the global welfare issue theoretically since the energy policy, here FIT programs, isset at national levels. Therefore, domestic welfare is more appropriate. However, in the simulations belowwe provide global welfare to allow comparison.

13This assumption is quite realistic since for instance, conventional fossil-fuel and nuclear generatorsprovided 87% of U.S. electricity in 2011 (Schmalensee, 2013).

7

of renewable energy produced is positively linked to the environmental benefit (cf infra).Subsidizing renewable energy development requires public funds, and therefore its socialmarginal cost must be considered. Note that if one assumes no externality and no socialcost of public funds, setting a FIT program would systematically decrease domestic welfareunder price competition. Only the presence of the externality of the fossil electricitytheoretically justifies the use of subsidies in such an analytical framework.

2.2 Feed-in tariff equilibrium

The domestic and foreign firms compete on price a la Bertrand. They determine powergeneration equipment prices and produce the output that meets renewable electricity pro-ducers’ demands.

Demands. When 0 < θ < θd < θf , both firms producing power generation equipmenthave positive demands and form a duopoly:

Dd (kd, kf ) = θ = 1− kd − kfpg (Γd − Γf ) , (5)

Df (kd, kf ) = θf − θ = kdΓf − kfΓdpgΓf (Γd − Γf ) . (6)

Profits. Using expressions (1), (5), and (6), the two profit functions are given by:

Πd (kd, kf ) = kdDd (kd, kf )− aΓ2dDd (kd, kf )2 , (7)

Πf (kd, kf ) = kfDf (kd, kf )− aΓ2fDf (kd, kf )2 . (8)

The two firms compete on price choosing kd and kf non-cooperatively, anticipating thecompetitor’s price. Since each profit function Πi (k) = Πi (ki, kj), for i = d, f , j = d, f ,i 6= j, is continuous in k and is concave in ki for each value of kj , an equilibrium pointexists.

Maximization of the two profit functions gives the following best-response price func-tions:

kd (kf ) =(2aΓ2

d + pg (Γd − Γf ))

(kf + pg (Γd − Γf ))2aΓ2

d + 2pg (Γd − Γf ),

kf (kd) = kdΓfΓd

2aΓdΓf + pg (Γd − Γf )2aΓdΓf + 2pg (Γd − Γf ) .

where power generation equipment prices are strategic complements (best-response func-tion slopes are positive). Solving kd (kf ) and kd (kf ) gives the (unique) unregulated Nashequilibrium, (k∗d, k∗f ). The resulting equilibrium quantities D∗d = θ∗ and D∗f = θ∗f − θ∗ arederived using demand functions (5) and (6). Note that, as expected, both firms are active

8

on the domestic market.Finally, equilibrium profits are easily computed using (7) and (8). Equilibrium results

are therefore functions of all the exogenous variables: pg, Γd, Γf , and the cost parametera. For the sake of clarity the equations for the exposition equilibrium are not includedhere.14

Surplus of the renewable electricity producers. Two types of renewable electricityproducers can be distinguished, those that use domestic power generation equipment atprice kd and those that use foreign power generation equipment at price kf . One cantherefore compute the total surplus of renewable electricity producers as follows

PS =θ∗∫0

(pg (1− θ)Γd − kd)dθ +

θ∗f∫θ∗

(pg (1− θ)Γf − kf )dθ , (9)

with power generation equipment prices at their equilibrium value (k∗d, k∗f ). Note that asexpected, the total surplus PS is an increasing function of the electricity price pg.

Total amount of renewable energy. One can compute the total amount of renewableenergy, denoted G, produced at equilibrium (i.e., with θ and θf calculated with k∗d andk∗f ):

G = Gd +Gf

=θ∫

0

(1− θ)Γddθ +θf∫θ

(1− θ)Γfdθ . (10)

Unsurprisingly, like the producer surplus PS, G grows with the guaranteed electricityprice pg.

Domestic welfare. Total domestic welfare Wd is computed as the sum of the domesticpower generation equipment producer’s profits, the renewable electricity producers surplus,and the environmental benefits of renewable generation (i.e., the economy of the externalcosts of fossil electricity), minus the total cost of the FIT program:

Wd = Πd + PS + G (γ − λ (pg − p)) , (11)14Expressions are available upon authors’ request.

9

where γ > 0 is the social cost of fossil electricity, and where λ > 1 stands for the MarginalCost of Public Funds (MCPF).15 Two tacit assumptions are made in the expression (11).First, electricity consumer surplus is fixed (i.e., does not depend on the type of electricityconsumed). Second, fossil electricity producers have no surplus. In the numerical analysisbelow, we refer to WT for the global welfare that includes the foreign power generationequipment producer’s profit (WT = Wd + Πf ).

In the rest of the paper we investigate the impact of a FIT program explored for twovariants, i.e. applied to both types (domestic and foreign), and one type (i.e., LCR) ofpower generation equipment sold in the domestic market. We compare both situations tothe unregulated equilibrium (pg = p) and make different propositions relative to marketcharacteristics such as power generation equipment prices, renewable electricity producers’surplus, and welfare.16

3 The Effects of Feed-in Tariff Programs

In the analysis below, we focus on two types of FIT programs. In the first, the priceis ensured irrespective of the power generation equipment used for electricity generation(most frequent scheme, used, e.g., in France and Germany). In the second, the guaranteedprice is conditional on the ‘local content’ of the electricity generation, i.e. on the use of thedomestic power generation equipment. This is a LCR exemplified by the WTO disputebetween Canada versus Japan and EU mentioned earlier. For example, under Ontario’sFIT program, government pays high fixed prices for electricity produced with renewableenergy sources but only if producers use a certain level of locally-produced components.

3.1 The Unconditional FIT

We first consider the unconditional FIT program. In order to assess the characteristicsof this policy, we derive results on power generation equipment prices and demands, theamount of renewable electricity produced, profits, and welfare.

Proposition 1 Following unconditional feed-in tariffs, the equilibrium prices of both na-tional and foreign power generation equipment producers rise.


15The MCPF measures “the loss incurred by society in raising additional revenues to finance governmentspending” (Dahlby, 2008, pp.1). Dahlby and Ferede (2012) provide estimations of λ around 1.31 in theProvince of Ontario.

16The benchmark could be considered as the situation where no power generation equipment (and norenewable energy) is produced because of lack of profitability. However, this setting might excessively favorthe FIT programs in the comparison.

10

The intuition behind Proposition 1 is straightforward: since the guaranteed price ishigher than the market price for electricity, renewable electricity producers have a higherwillingness to pay for both types of power generation equipment. This situation allowspower generation equipment producers to increase their prices.

Proposition 2 Guaranteeing a price for renewable electricity higher than the market pricefor electricity has the following three effects: i) an increase in market coverage θf , ii) anincrease in demand for the domestic power generation equipment, and iii) an increase indemand for the foreign power generation equipment but only if the market electricity priceand the gap between the prices of the domestic and the foreign power generation equipmentare not too high:

pg < 2aψ (Γd,Γf ) and k∗d − k∗f < 2aϕ (Γd,Γf ) ,

where ϕ (Γd,Γf ) and ψ (Γd,Γf ) are two positive functions of Γd and Γf .


The intuition behind Proposition 2 is as follows. Because the guaranteed price pg ishigher than p, some electricity producers using foreign power generation equipment underp will find it profitable to use domestic rather than foreign power generation equipment(increase of θ with p), while others will find market entry attractive using the foreign powergeneration equipment (the rise of θf with p). Therefore, the demand for foreign generationpower equipment increases in p only if the existing producers switching to domestic powergeneration equipment are less numerous than the producers entering the electricity marketand running their installations using the foreign power generation equipment. Proposi-tion 2 states that this result holds only if two conditions are met simultaneously. First,the difference in power generation equipment prices cannot be larger than a threshold thatdepends on the power generation equipment qualities. This condition reveals that the gapbetween domestic and foreign power generation equipment qualities cannot be too large.17

Second, the FIT program pg cannot be too high, in order to avoid a situation where theproducers that would use foreign power generation equipment under p but domestic powergeneration equipment under pg become too numerous.

We denote G as the total amount of renewable energy the FIT program. G is computedas in the expression (10) with θ and θf calculated using the new equilibrium prices.

17Remember that the conditions to maintain a duopoly (i.e., 0 < θ < θd < θf ) are such that a positivedifference in power generation equipment quality (Γd > Γf ) must be translated into a positive differencein power generation equipment prices at equilibrium (k∗d > k∗f ). An important gap between k∗d and k∗ftherefore, reveals an important difference in the power generation equipment qualities.

11

Corollary 1 The production of renewable energy in the domestic country increases undera feed-in tariff program.


The rise in renewable energy production under a guaranteed price results from theincrease in market coverage (increase of θf ) and from the fact that under a guaranteedprice more producers choose the more efficient power generation equipment (increase ofθ).

The results of a guaranteed price for renewable energy can also be assessed in relationto profits and welfare. We can show that whatever the parameter values under duopolyconstraints, the power generation equipment producers’ profits and the electricity produc-ers’ surpluses rise following implementation of an unconditional FIT program.

Proposition 3 Under a FIT program, domestic and foreign power generation equipmentproducers’ profits and renewable electricity producers’ profits rise.


The intuition behind Proposition 3 is straightforward. First, guaranteeing a price pghigher than the market price for electricity p increases the profits of both power genera-tion equipment producers since the profit functions are strictly increasing in p. Second,the electricity producers fall into two categories: those using domestic power generationequipment located in the segment

[0; θ], and those using foreign power generation equip-

ment found in the segment[θ; θf

]. Implementing a FIT program pushes up θ and θf .

The rise in θ means that at least one producer θ finds it more profitable to use domesticthan foreign power generation equipment. As a consequence, the profit of each renewableelectricity producer under this θ increases under a FIT program. The rise of θf meansthat at least one producer θ finds it profitable to enter the market and produce using theforeign power generation equipment. As a consequence, the profit of each producer usingforeign power generation equipment under this θ increases under a FIT program. Thesetwo effects imply that electricity producers’ total profits increase when a guaranteed pricehigher than the market price is proposed.

This first series of results sheds some light on the WTO disputes we are interestedin. First, there is a “pass-through effect”. Subsidizing renewable electricity producersclearly also benefits power generation equipment producers. Second, regarding non-discrimination, the model shows that both foreign and domestic equipment producers’profits increase under a FIT program (Proposition 3). However, the quantity of lower-

12

quality foreign equipment sold does not systematically rise, contrary to the case for do-mestic equipment (Proposition 2). Third, vis-a-vis the environmental dimension, a FITprogram pushes up renewable electricity production (Corollary 1). This last result isessential for justifying a FIT program. An impartial regulator is aware of the welfareconsequences of its policy choices. We present the final two results before introducing aLCR in the FIT program address this issue. They show that the impact of the guaranteedprice on total domestic welfare is ambiguous.

Proposition 4 Domestic welfare may be strictly decreasing under the FIT program.

Corollary 2 A guaranteed price pg maximizing domestic welfare does not exist systemat-ically.

The proofs are given in the Appendix �

Proposition 4 and Corollary 2 question the economic rationale of the FIT programespecially when domestic welfare is decreasing. In such a setting, a FIT program couldnot be justified from a welfare perspective.18 These results highlight one of the maininconveniences attached to subsidizing renewable electricity generation in order to reducefossil fuel generation.19 In effect, some electricity producers would produce in the absenceof a subsidy and would receive a subsidy under a FIT program. Therefore, in this modelpublic funds are expended without any increase in the supply of renewable electricity,since producers are assumed to use only one unit of power generation equipment. Theseresults show that welfare maximization does not necessarily resolve the question of whereto set pg. Note that the target of maximizing renewable energy production to fix pg doesnot help either since the production of renewable energy is increasing with the electricityprice (Corollary 1). Thus, political motivations (exogenous in our model) are central indeciding about implementation of this type of regulatory scheme.

3.2 The FIT with LCR

In a conditional FIT program, benefiting from a guaranteed price depends on a LCR.Specifically, the electricity price pg is ensured only for renewable electricity producers usingdomestic power generation equipment, while the producers using foreign power generationequipment obtain the market electricity price p. Therefore, the critical values for θ changecompared to their values in the unconditional FIT scenario.

18For instance, the numerical simulations in Table D.1 provide an example where unconditional FITincreases renewable energy but implies a decrease in (domestic and global) welfare.

19See Jaffe et al. (2005) and Fischer and Newell (2008).

13

Under the assumption of a duopoly setting,20 both demands for power generationequipment result in the following expressions:

Dd,c (kd, kf ) = θc = 1− kd − kfpgΓd − pΓf

,

Df,c (kd, kf ) = θf,c − θc = kd − kfpgΓd − pΓf

− kfpΓf

.

We can then deduce equilibrium power generation equipment prices k∗d,c and k∗f,c. Thedemands for power generation equipment at equilibrium are denoted D∗d,c and D∗f,c.

Proposition 5 Under the conditional FIT program, the price of the domestic power gen-eration equipment always increases compared to the unregulated equilibrium, while thechange in the price of the foreign good is ambiguous.


From Proposition 5, the producer of domestic power generation equipment alwayscharges higher prices under a conditional FIT program. In contrast, the LCR disadvan-tages the foreign producer in the domestic market. Maintaining a duopoly setting, insome circumstances, may require a reduction in k∗f,c compared to the unregulated case.However, this effect is not systematic as shown by Fig. C.2 in the Appendix. Comparedto the unconditional FIT (Proposition 1), this result shows that the “pass-through effect”of the subsidy to k∗f,c may not appear.

The consequences of a LCR in a FIT program related to generation equipment quan-tities sold, on the quantity of renewable energy produced, and on the foreign power gen-eration equipment producer’s profit can now be evaluated.

Proposition 6 Compared to the unconditional FIT program, the LCR has a negativeimpact on the demand for foreign generation equipment and may decrease the demand forthe domestic one. The reduction of the latter appears for example when Γf < Γf (Γd),where Γf (Γd) is a positive fonction of Γd.


The decrease in the demand for domestic equipment may appear surprising. However,a FIT program with LCR may give more market power to the domestic equipment pro-ducer. The more the domestic and foreign equipment would be differentiated (i.e., the gap

20Note that the introduction of the LCR could lead to a domestic monopoly. We do not consider thiscase since discrimination appears obvious.

14

between Γd and Γf is large), the more the domestic equipment producer would be in theposition to impose a high price to maximize its profit at the expense of the quantity sold.

Corollary 3 Compared to the unconditional FIT program, the production of renewableenergy in the domestic country may decrease under a FIT program with a LCR.


Proposition 7 Compared to the unregulated equilibrium, the LCR does not necessarilylead to a reduction in the foreign power generation equipment producer’s profit.


Propositions 6 and 7 throw light on the WTO disputes generated by FIT programswith LCR from two different perspectives. Proposition 6 shows that the LCR clausecan be considered a strategic trade policy. The discrimination it implies results in areduction in demand for foreign power generation equipment compared to the unregulatedequilibrium. If the duopoly assumption were relaxed at the extreme, the LCR could leadthe foreign power generation equipment producer to exit the market. However, Proposition7 shows that the foreign power generation equipment producer’s profit could increaseas a consequence of the LCR compared to the unregulated case. This situation occurswhen the price effect of the FIT program with LCR is positive and counterbalances thenegative demand effect.21 This result shows that the FIT program with LCR should not beconsidered as automatically favoring the domestic power generation equipment producer atthe expense of the foreign one. When the foreign power generation equipment producer’sprofit increases following a LCR, discrimination could be questioned.

Proposition 6 contains another disruptive result: The demand for domestic powergeneration equipment can decrease compared to the unconditional FIT program. Whenboth demands for power generation equipment decrease, a reduction in the renewableenergy generated occurs, as stated in Corollary 3. Therefore, unlike in the case of anunconditional FIT program, inclusion of an LCR can lead to an outcome that is oppositeto the energy transition objective. This result clearly questions the economic rationale forLCR (see, e.g., simulations in Table D.1 in the Appendix where conditional FIT cannot bejustified from the renewable energy production, neither from a global welfare perspective).

21In the Appendix, the simulation in Table D.1 presents the situation where the foreign firm’s profits doincrease following the introduction of a FIT program with LCR.

15

4 Simulations for the Ontario Case

The objective of this Section is to calibrate the model using the Ontario’s data, to comparethe two FIT programs, and to provide several critical values of particular interest in termsof policy implications.

4.1 Results with a calibrated model

A first numerical simulation of the effects of a FIT program is given in Table 1 (first threecolumns). The use of the unconditional FIT analyzed above is compared with the un-regulated equilibrium. The simulation is based on real 2012 world data from the Canada– Renewable Energy dispute. The value of λ in the Province of Ontario is estimated byDahlby and Ferede (2012), solar efficiency data are collected from SRoeCo Solar websitethat provides comparisons of solar panels.22 The chosen Canadian product is manufac-tured by Canadian Solar, and the Japanese one is made by Sharp. Both are 250-wattmonocrystalline solar panels. The market prices for electricity and the FIT rates areavailable in Winfield and Dolter (2014). Finally, an estimation of the social cost of fossilelectricity in $/MWh (γ) is taken from Dewees (2013).

Characteristics Unregulated Unconditional FIT Conditional FIT

kd 51.8348 219.5856 524.371kf 48.9512 193.1990 33.3157Dd 0.0988838 0.389198 0.393909Df 0.1033441 0.377570 0.063137Πd 2.57846 46.0031 166.134Πf 2.54565 39.3997 1.16542PSd 0.942676 187.7375 69.1434PSf 0.327662 59.0452 0.1223Gd 1.51707 5.05925 5.10551Gf 1.34662 2.44428 0.556439

γG 85.9110 225.1064 169.859Wd 89.7598 26.4100 34.4005WT 92.3054 65.8097 35.5659

Table 1: Simulation of the effects of a FIT Program(a = 1, Γd = 16.14, Γf = 15.34, γ = 30, λ = 1.31, p = 4, and pg = 54)

This parametrization illustrates the results demonstrated in Propositions 1-2: increasesin power generation equipment prices, profits, and renewable electricity produced followingimplementation of an unconditional FIT program. Plotting the respective best-responseprice functions from the two settings shows the “pass-through effect” of the subsidy toinput prices (k∗d, k∗f ) (Fig. C.1 in the Appendix). The simulated unconditional FIT pro-gram appears to be welfare-decreasing (Fig. C.3), which is one of the cases detailed in

22See http://sroeco.com/solar/table/.

16

Proposition 4. As already underlined from a theoretical point of view, there is no inter-national competition issue involved in this policy since both power generation equipmentproducers’ profits increase under the FIT program. A different perspective is offered ifFIT programs are conditional on a LCR.

The fourth column of Table 1 presents such a situation. The results show a decrease inthe foreign power generation equipment producer variables (price, demand, and profit) andan increase in these variables for the domestic power generation equipment producer. Inthis parameter setting, introducing a LCR clearly decreases the total amount of renewableelectricity produced in the domestic industry (7.50 vs. 5.66). The FIT program with LCRis less environmentally friendly as the program without LCR but performs better than ifno support is provided. In terms of domestic welfare, the FIT program with LCR leads tohigher welfare compared to the case with no LCR but remains lower than the unregulatedequilibrium. If global welfare is taken into account, the unconditional FIT is preferred tothe conditional one.

4.2 Computation of critical parameters values

The model calibrated with Ontario’s data is also used to compute critical values for theexternal cost of fossil electricity (γ), the MCPF (λ), and the FIT rate (pg). We pro-vide a sensitivity analysis to these critical parameters, which gives a clear cutoff of whatis required to justify the policy and how these values are sensitive to the parameters,particularly the external cost and the MCPF.23

A floor value for the external cost of fossil electricity, above which welfare increasesunder pg = 54 compared to the unregulated equilibrium, is first computed. Results arepresented in Table 2, and show that the Ontario’s conditional FIT would require anexternal cost higher than C$ 49 in order to improve domestic (Wd) or global welfare(WT ).24 This requirement is weakened as soon as the MCPF decreases, or when the FITbecomes unconditional.

FIT without LCR FIT with LCR

λ = 1 λ = 1.31 λ = 1 λ = 1.31

γ min. for a Wd increase 18.58 43.65 18.42 49.78γ min. for a WT increase 10.64 35.71 18.00 49.36

Table 2: Critical values of γ (unregulated benchmark)(a = 1, Γd = 16.14, Γf = 15.34, p = 4, and pg = 54)

In the same vein, a ceiling value for the MCPF below which welfare increases witha FIT pg = 54 is calculated. Results in Table 3 show that Ontario’s MCPF of C$ 1.31

23We would like to thank an anonymous referee for pointing out the interest of the sensitive analysis.24Remember that the external cost of fossil electricity estimated by Dewees (2013) amounts to C$ 30.

17

estimated by Dahlby and Ferede (2012) is actually too high to ensure an enhanced welfare.The constraint is more flexible when global welfare is considered rather than the domesticone, or when the LCR is canceled.


λ max. for a Wd increase 1.14 1.114λ max. for a WT increase 1.24 1.118

Table 3: Critical values of λ (unregulated benchmark)(a = 1, Γd = 16.14, Γf = 15.34, γ = 30, p = 4, and pg = 54)

The FIT that shall not be exceeded is computed when the external cost of fossilelectricity is estimated at C$ 30. The results in Table 4 show that this maximum FITamounts to C$ 27.41, which is well below the Ontario’s FIT of C$ 54 reported by Winfieldand Dolter (2014). Justifying such a guaranteed price for renewable electricity (underLCR) would require an external cost of fossil electricity γ estimated at C$ 50.


γ = 30 γ = 50 γ = 30 γ = 50

pg max. for a Wd increase 33.57 64.29 27.41 54.30pg max. for a WT increase 45.10 80.98 28.21 54.88

Table 4: Critical values of pg (unregulated benchmark)(a = 1, Γd = 16.14, Γf = 15.34, γ = 30, λ = 1.31, and p = 4)

Finally, the requirement in terms of minimum external cost of fossil electricity is weakerunder a FIT without LCR than with a LCR (35.71 vs. 43.65 in Table 2). The sametype of results applies with the MCPF and the guaranteed price pg, too. The maximumMCPF and highest pg are larger under the FIT without LCR. Besides, the global welfareperspective gives more flexibility on the critical values computed. For instance, societycan bear a MCPF of C$ 1.24 under a FIT without LCR compared to C$ 1.14 with LCR(Table 3).

5 Discussion

This paper investigated the consequences of a FIT program in the context of energytransition, using an international quality differentiated duopoly model in which powergeneration equipment producers compete on price. The results are applied to a discussionof the Canada – Renewable Energy case. The ruling in this dispute was based on twomain points: National treatment and subsidies. In this case, the LCR was found tobe incompatible with the GATT (General Agreement on Tariffs and Trade) principle of

18

National Treatment since it discriminated between domestic and foreign power generationequipment in favor of the former. The results of our model confirm this idea as long as salesare taken into account in a duopoly setting. Proposition 6 establishes that, as a result ofsuch a regulatory scheme, demand for domestic power generation equipment may increasewhile demand for the foreign power generation equipment decreases. However, introducinga difference in the quality of the power generation equipment produced on both sides of theborder raises questions about this discrimination. First, our results show that a rise in theprofits of the foreign power generation equipment producer could be due to the inclusionof LCR in the FIT program (this situation is depicted in the model simulation presentedin Table D.1 in the Appendix). In this case, treating the two producers differently shouldnot be considered ‘damaging’ discrimination. Second, a decrease in the demand for foreigngeneration equipment could also emerge under a FIT program with no LCR, as shown inProposition 2. In this case, LCR would exacerbate discrimination.

The question of whether the guaranteed price for electricity under Ontario’s FIT pro-gram constituted a subsidy was the subject of heated debate in the Canada – RenewableEnergy dispute. Under Article 1 of the SCM Agreement, a “subsidy shall be deemed toexist if there is a financial contribution by a government” conferring a ‘benefit’ to its re-cipient. On the benefit determination question, the Panel and the AB expressed differentviews. The question of what price (or equivalently what market) can be used as bench-mark became central to the dispute. The Panel considered that the guaranteed price forelectricity should not give “more than adequate remuneration” to the renewable electricityproducers, and that this ‘adequacy’ had to be assessed compared to regular market condi-tions. The Panel refused to consider requests from Japan and the EU that the wholesaleprice for electricity should be considered as the benchmark because it was quite clear thatthis price was distorted by the Government of Ontario’s policy (being simultaneously abuyer and a producer of electricity). Therefore, the Panel concluded that the claimantshad failed to establish the existence of subsidization.25 The AB reached the same con-clusion but chose the competitive market for renewable energy rather than the wholesalemarket as the benchmark.26

If the FIT program had been considered as constituting a subsidy, the question ofwhether such subsidies were prohibited would inevitably have been addressed. In lightof the Panel’s conclusion about national treatment, and Article 3 of the SCM Agreementprohibiting subsidies contingent “upon the use of domestic over imported goods”, it is

25Against this background, a dissenting opinion expressed by one Panel member was recorded in thePanel’s Report. According to this member, the fact that every party recognized that the guaranteed pricefixed in the FIT program is a condition for the existence of renewable energy producers, shows clearlythat the FIT program conferred a benefit on these producers. See (Cosbey and Mavroidis, 2014) for adiscussion.

26This position has attracted critique from legal scholars. However, this is beyond the scope of thepresent paper (see Cosbey and Mavroidis (2014) and Charnovitz and Fischer (2015)).

19

very likely that the FIT program would have been found incompatible with the SCMAgreement, without consideration of its environmental dimension.

In the model, the FIT program (in whatever form) is assumed to constitute subsidiesin light of the preceding debate: the guaranteed price is set higher than the wholesaleprice. Nevertheless, we believe that this theoretical framework and simulations provideinteresting insights into the debate.

The first point that can be discussed is how subsidies are identified. In the verticalrelation between power generation equipment producers and electricity producers, pre-sented in the model, Proposition 3 shows that subsidizing the latter clearly benefits theformer (with a ‘pass-through effect’). The problem to be addressed of protectionism, reststherefore on the power generation equipment not the electricity market. Japanese andEuropean electricity generators are not in competition with Ontario’s power generators.The competition is among power generation equipment producers. As a consequence, theeffects of the subsidies implied by the FIT program on prices, quantities, and profits,should be assessed on the power generation equipment market where the LCR is poten-tially damaging.

The second point worth discussing is the way that LCR is excluded from debate oversubsidies. As Cosbey and Mavroidis (2014) pointed out, “the Panel and the AB dividedthe challenged measure in two (...) as if the two were unrelated”. However, Japan and theEU would not have claimed that Ontario’s FIT program constituted subsidies if the LCRhad not been part of the FIT program. Therefore, while this requirement is at the heartof the dispute, the Panel did not consider it when discussing subsidies to avoid the risk ofruling on FIT programs more generally (and dangerously as we argue below). Our modelshows that the FIT program without LCR can distort trade (on the power generationequipment market), but not systematically however.

Finally, the third point is the way the environment was excluded from the Panel’sruling. If one agrees with the view of the dissenting Panel member referred to above, thequestion of justification for the subsidies is unavoidable: Can environmental protection andenergy transition be put forward as reasons for subsidizing renewable energy producersin light of the SCM Agreement? Our model shows that the FIT program without LCR,indisputably raises the production of renewable energy (Corollary 1). As Rodrik (2014)argues, “industrial policies have an indispensable role in putting the global economy ona green growth path.”27 What is the real weight of such an argument in this kind ofdispute especially in the face of discrimination? Should renouncing the LCR be seen asthe ultimate solution? Discussing the case, Cosbey and Mavroidis (2014) point out thateven with no LCR, FIT programs could be found incompatible with the SCM Agreementsince the AB finding leaves this door open. Therefore, there is an urgent need for WTO

27See also Rubini (2012).

20

members to decide whether the environmental protection motive should be introduced inthis Agreement.

This debate over the possibility of subsidies for environmental purpose takes place in aparticular context. During the first years of the WTO’s existence, such “green subsidies”were possible under Art. 8 of the SCM Agreement. However, this article had a limitedlife span of five years and was not extended. At the end of the day, the Panel and the ABmay have been wise in their approach to this issue.28 What would have happened werethe FIT program to have been classified as an illegal subsidy without discussion of theappropriateness of such a policy? The AB would have been deciding a sensitive issue onbehalf of WTO Members. This would have questioned all the renewable energy fundingprograms across Europe, Canada, and other countries. Were such a debate on ‘greening’the SCM Agreement to take place, the economic reasoning developed in our model clearlyestablishes the positive characteristic of a FIT program as well as its main limitations, i.e.possible welfare reduction and eventual decreased production of renewable energy if theLCR were adopted. Our simulations point out the critical values for the external cost offossil electricity and the MCPF to justify a given FIT rate.

28For Cosbey and Mavroidis (2014) the “acrobatics” in both the Panel and AB reasoning was motivatedby the wish to avoid finding on the FIT as a subsidy.

21

References

Antonelli, Marco and Desideri, Umberto. (2014). ‘The Doping Effect of ItalianFeed-in Tariffs on the PV Market’, Energy Policy 67, 583–594.

Borenstein, Severin. (2012). ‘The Private and Public Economics of Renewable Elec-tricity Generation’, Journal of Economic Perspectives 26(1), 67–92.

Charnovitz, Steve and Fischer, Carolyne. (2015). ‘Canada-Renewable Energy: Im-plications for WTO Law on Green and Not-So-Green Subsidies’, World Trade Review14, 177–210.

Chua, Shing Chyi, Oh, Tick Hui and Goh, Wei Wei. (2011). ‘Feed-in Tariff Outlookin Malaysia’, Renewable and Sustainable Energy Reviews 15(1), 705–712.

Cosbey, Aaron and Mavroidis, Petros C. (2014). ‘A Turquoise Mess: Green Subsi-dies, Blue Industrial Policy and Renewable Energy: The Case for Redrafting the Sub-sidies Agreement of the WTO’, Journal of International Economic Law 17(1), 11–47.

Dahlby, Bev. (2008), The Marginal Cost of Public Funds: Theory and Applications, TheMIT Press, Cambridge, Massachusetts.

Dahlby, Bev and Ferede, Ergete. (2012). ‘The Effects of Tax Rate Changes on TaxBases and the Marginal cost of Public Funds for Canadian Provincial Governments’,International Tax and Public Finance 19, 844–883.

Dewees, Donald N. (2013). ‘The Economics of Renewable Electricity Policy in Ontario’,Working Papers from University of Toronto, Department of Economics No. 478 .

Edenhofer, Ottmar, Hirth, Lion, Knopf, Brigitte, Pahle, Michael, Schlomer,Steffen, Schmid, Eva and Ueckerdt, Falko. (2013). ‘On the Economics of Renew-able Energy Sources’, Energy Economics 40, S12–S23.

Fischer, Carolyn, Greaker, Mads and Rosendahl, Knut Einar. (2014). ‘RobustPolicies against Emission Leakage: The Case for Upstream Subsidies’, Cesifo WorkingPaper No. 4742 .

Fischer, Carolyn and Newell, Richard G. (2008). ‘Environmental and TechnologyPolicies for Climate Mitigation’, Journal of Environmental Economics and Management55(2), 142–162.

Greaker, Mads and Rosendahl, Knut Einar. (2008). ‘Environmental Policy with Up-stream Pollution Abatement Technology Firms’, Journal of Environmental Economicsand Management 56(3), 246–259.

Huang, Yun-Hsun and Wu, Jung-Hua. (2011). ‘Assessment of the Feed-in TariffMechanism for Renewable Energies in Taiwan’, Energy Policy 39(12), 8106–8115.

Jaffe, Adam B., Newell, Richard G. and Stavins, Robert N. (2005). ‘A taleof Two Market Failures: Technology and Environmental Policy’, Ecological Economics54(2-3), 164–174.

22

Jenner, Steffen, Groba, Felix and Indvik, Joe. (2013). ‘Assessing the Strength andEffectiveness of Renewable Electricity Feed-in Tariffs in European Union Countries’,Energy Policy 52, 385–401.

Johnstone, Nick, Hascic, Ivan and Popp, David. (2010). ‘Renewable Energy Policiesand Technological Innovation: Evidence based on Patent Counts’, Environmental andResource Economics 45, 133–155.

Mabee, Warren E., Mannion, Justine and Carpenter, Tom. (2012). ‘Comparingthe Feed-in Tariff Incentives for Renewable Electricity in Ontario and Germany’, EnergyPolicy 40, 480–489.

Moorthy, K. Sridhar. (1988). ‘Product and Price Competition in a Duopoly’, MarketingScience 7(2), 141–168.

Nesta, Lionel, Vona, Francesco and Nicolli, Francesco. (2014). ‘EnvironmentalPolicies, Competition and Innovation in Renewable Energy’, Journal of EnvironmentalEconomics and Management 67(3), 396–411.

Pillai, Unni and McLaughlin, Jamison. (2013). ‘A Model of Competition in the SolarPanel Industry’, Energy Economics 40, 32–39.

Pirnia, Mehrdad, Nathwani, Jatin and Fuller, David. (2011). ‘Ontario Feed-in-Tariffs: System Planning Implications and Impacts on Social Welfare’, The ElectricityJournal 24(8), 18–28.

Qudrat-Ullah, Hassan. (2014). ‘Green Power in Ontario: A Dynamic Model-basedAnalysis’, Energy 77, 859–870.

Rodrik, Dani. (2014). ‘Green Industrial Policy’, Oxford Review of Economic Policy30(3), 469–491.

Rubini, Luca. (2012). ‘Ain’t Wastin’ Time No More: Subsidies for Renewable Energy,The SCM Agreement, Policy Space, and Law Reform’, Journal of International Eco-nomic Law 15(2), 525–579.

Rubini, Luca. (2014). ‘The Good, the Bad, and the Ugly. Lessons on Methodologyin Legal Analysis from the Recent WTO Litigation on Renewable Energy Subsidies’,Journal of World Trade 48(5), 895–938.

Sahoo, Anshuman and Shrimali, Gireesh. (2013). ‘The Effectiveness of DomesticContent Criteria in India’s Solar Mission’, Energy Policy 62, 1470–1480.

Schmalensee, Richard. (2013). ‘The Performance of U.S. Wind and Solar GeneratingUnits’, NBER Working Paper No. 19509 .

Shadikhodjaev, Sherzod. (2015). ‘Renewable Energy and Government Support: Timeto Green the SCM Agreement?’, World Trade Review FirstView, 1–28.

Smith, Michael G. and Urpelainen, Johannes. (2014). ‘The Effect of Feed-in Tariffson Renewable Electricity Generation: An Instrumental Variables Approach’, Environ-

23

mental and Resource Economics 57(3), 367–392.Steer, Andrew. (2013). ‘Resource Depletion, Climate Change, and Economic Growth’,

Working Paper of the Global Citizen Foundation, June 2013 .Winfield, Mark and Dolter, Brett. (2014). ‘Energy, Economic and Environmental

Discourses and their Policy Impact: The Case of Ontario’s Green Energy and GreenEconomy Act’, Energy Policy 68, 423–435.

Wu, Mark and Salzman, James. (2013). ‘The Next Generation of Trade and En-vironment Conflicts: The Rise of Green Industrial Policy’, Mimeo, draft April 2013.

Yatchew, Adonis and Baziliauskas, Andy. (2011). ‘Ontario Feed-in Tariff Programs’,Energy Policy 39(7), 3885–3893.

24

A Prices, demands, and profits at equilibrium

A.1 FIT program with no LCR

k?d = 2pgΓd (Γd (2aΓd + g)− pgΓf ) (Γd (aΓf + pg)− pgΓf )ΓdΓf

(2aΓd (2aΓd + pg)− 5p2

g

)+ pgΓ2

f (pg − 2aΓd) + 4pgΓ2d (aΓd + pg)

k?f = pgΓf (pgΓf − Γd (2aΓd + pg)) (pgΓf − Γd (2aΓf + pg))ΓdΓf


g

)+ pgΓ2


D?d = 2pgΓd (Γd (aΓf + pg)− pgΓf )

ΓdΓf(2aΓd (2aΓd + pg)− 5p2

g

)+ pgΓ2


D?f = pgΓd (Γd (2aΓd + pg)− pgΓf )


g

)+ pgΓ2


π?d =4p2gΓ2

d (Γd (aΓd + pg)− pgΓf ) (pgΓf − Γd (aΓf + pg)) 2(ΓdΓf


g

)+ pgΓ2


)2

π?f =p2gΓdΓf (Γd (2aΓd + pg)− pgΓf ) 2 (Γd (aΓf + pg)− gΓf )(


g

)+ pgΓ2


)2

A.2 FIT program with LCR

k?d,c = 2pgΓd (Γd (2aΓd + pg)− pΓf )(pgΓd (aΓf + p)− p2Γf

)ΓdΓf

(2aΓd

(2apgΓd + 2p2

g − p2)− 5pgp2

)+ pΓ2

f (p2 − 2apgΓd) + 4pgpΓ2d (aΓd + pg)

k?f,c = pΓf (pΓf − Γd (2aΓd + pg))(p2Γf − pgΓd (2aΓf + p)

)ΓdΓf

(2aΓd

(2apgΓd + 2p2

g − p2)− 5pgp2

)+ pΓ2


D?d,c = 2pgΓd

(pgΓd (aΓf + p)− p2Γf

)ΓdΓf

(2aΓd

(2apgΓd + 2p2

g − p2)− 5pgp2

)+ pΓ2


D?f,c = pgpΓd (Γd (2aΓd + pg)− pΓf )

ΓdΓf(2aΓd

(2apgΓd + 2p2

g − p2)− 5pgp2

)+ pΓ2


π?d,c =4p2gΓ2

d (Γd (aΓd + pg)− pΓf )(p2Γf − pgΓd (aΓf + p)

) 2(ΓdΓf

(2aΓd

(2apgΓd + 2p2

g − p2)− 5pgp2

)+ pΓ2


)2

π?f,c = pgp2ΓdΓf (Γd (2aΓd + pg)− pΓf ) 2 (pgΓd (aΓf + p)− p2Γf

)(ΓdΓf

(2aΓd

(2apgΓd + 2p2

g − p2)− 5pgp2

)+ pΓ2


)2

B Proofs of Lemmas, Propositions, and Corollaries

Most proofs use Mathematica 10 software. Computing codes and details are availableupon request.

25

B.1 Proof of Lemma 1

Consider the situation where power generation equipment d is the most efficient: Γd > Γf .Using (3) and (4) permits to write θd > θ as kdΓf−kf Γd

Γd(Γd−Γf ) > 0. As Γd > Γf , we must havekdkf

> ΓdΓf

. When θd is greater than θf , we have 1 − kdpgΓd

> 1 − kf

pgΓf, or Γd

Γf> kd

kf. We

therefore have θf > θd when θd > θ.Under the same reasoning we can show that θd < θ implies that kd

kf< Γd

Γfwhen Γd > Γf ,

so that we can deduce that we must have θd > θf .

B.2 Proof of Proposition 1

Both derivatives of k∗d and k∗f with respect to price p are always positive under the duopolyconstraint. As a consequence, the effect of a price increase will always lead to a rise in powergeneration equipment prices at equilibrium. Remember that power generation equipmentprices are strategic complements. If the domestic firm increases its price, the foreign rival’sbest strategy is to do the same.

In addition, the derivative of k∗d with respect to price p is higher than the derivativeof k∗f . In other words, ceteris paribus, an unconditional FIT program has a larger impacton the domestic power generation equipment price than on the foreign power generationequipment price.


Plugging the equilibrium prices k∗d and k∗f into the demand functions gives the two followingequilibrium demands for power generation equipment goods:

D∗d = 2pgΓd (Γd (pg + aΓf )− pgΓf )4pgΓ2

d (pg + aΓd) + ΓdΓf(2aΓd (pg + 2aΓd)− 5p2

g

)+ pgΓ2

f (pg − 2aΓd)(B.1)

D∗f = pgΓd (Γd (pg + 2aΓd)− pgΓf )4pgΓ2

d (pg + aΓd) + ΓdΓf(2aΓd (pg + 2aΓd)− 5p2

g

)+ pgΓ2

f (pg − 2aΓd)(B.2)

The derivative of D∗d with respect to pg is always positive. Thus, guaranteeing a pricepg > p increases domestic demand for power generation equipment. The derivative of D∗fwith respect to pg is positive only if two conditions are met simultaneously:

k∗d − k∗f < 2aϕ (Γd,Γf ) and pg < 2aψ (Γd,Γf ) ,

26

with ϕ (Γd,Γf ) and ψ (Γd,Γf ) having the following strictly positive values:

ϕ (Γd,Γf ) = 1√2Γd−Γf

Γ3dΓf (Γd−Γf ) −

1Γ2

d

ψ (Γd,Γf ) =√√√√ Γ5

dΓf (2Γd − Γf )

(Γd − Γf )(

2Γ2d − 2ΓdΓf + Γ2

f

)2 + Γ2dΓf

2Γ2d − 2ΓdΓf + Γ2

f

.

B.4 Proof of Corollary 1

The rise in renewable energy production under a guaranteed price results from the increasesin θf and θ.


Guaranteeing a price pg higher than the market price for electricity p raises the profits ofboth power generation equipment producers since the profit functions ki(pg)Di(kd, kf ) −Ci(Di(kd, kf )) (i = d, f) are strictly increasing in pg at equilibrium.

The electricity producers fall into two categories: those using domestic power gener-ation equipment located in the segment

[0; θ], and those using foreign power generation

equipment in the segment[θ; θf

]. Implementing a FIT program pushes up θ and θf . The

rise in θ means that at least one producer θ finds it more profitable to use domestic thanforeign power generation equipment. As a consequence, the profit of each producer underθ increases under a FIT program. The rise in θf means that at least one producer θ findsit profitable to enter the market and produce using the foreign power generation equip-ment. Thus, the profit of each producer using foreign power generation equipment underθ increases under a FIT program. These two effects imply that electricity producers’ totalprofits increase when a guaranteed price higher than the market price is proposed.

B.6 Proof of the Proposition 4 and Corollary 2

The welfare expression defined in (11) is the sum of two parts. The first part, Πd + PS,corresponding to the sum of the profits of the domestic equipment supplier and of therenewable electricity producers is strictly increasing in pg at an increasing rate. Thesecond part, G (γ − λ (pg − p)), where G is strictly increasing at a decreasing rate in pg,compares the external benefit of renewable energy to the cost of public funds. Two broadcases can be distinguished. In the first one, λ(pg − p) > γ so that G (γ − λ (pg − p)) isstrictly decreasing in pg with a undetermined second derivative’s sign. In the second caseλ(pg − p) < γ. In this case, the sign of the second derivative of G (γ − λ (pg − p)) withrespect to pg is always negative, but the sign of the first derivative is undetermined. As aconsequence, pg maximizing the domestic welfare does not systematically exist.

27


The implementation of a FIT program with LCR has three effects on the reaction functionskf (kd) and kd(kf ) compared to unregulated equilibrium: kf (kd) rotates to the right andkd(kf ) shifts out to the right and becomes less steep. The domestic power generationequipment price kd always increases as a result (k∗d,c > k∗d), whereas the net effect of thesemoves on the foreign power generation equipment price kf is ambiguous.


The demands for domestic and foreign equipment under unconditional and conditional FITprograms are given in Appendix A. First, the impact on the demand for foreign equipmentof FIT program with LCR (computed as D∗f,c−D∗f ) is always negative. Second, the impacton the demand for domestic equipment (D∗d,c−D∗d) is ambiguous. For example, the demanddecreases under the following conditions:

Γf < Γf (Γd) (B.3)

a < a (B.4)

For the sake of readibility and space constraints, the analytical expressions of Γf (Γd) anda are available upon request.

B.9 Proof of Corollary 3

Proposition 6 shows that the FIT program with a LCR systematically decreases the de-mand for foreign power generation equipment. Also, Proposition 6 shows that the programincreases demand for domestic power generation equipment only under certain circum-stances. Therefore, the production of renewable energy may decrease under this program.


Conditioning the FIT program to a LCR has an ambiguous effect on the foreign firm’sprofit. Contrary to what the intuition suggests, it is not always negative. The FIT programcan have a positive impact (Π∗f,c > Π∗f ) under the types of constraints below.

a < a1, with pg > p, or,

a2 < a < a3, with p < pg < pg ,

with a1, a2, a3, and pg positive functions of Γd, Γf , and p. These threshold values werecomputed with Mathematica.

28

The first derivative of the foreign firm’s profit with respect to g under a FIT programwith LCR confirms this finding. We have the following expression:

∂Πf

∂pg= ∂kf∂pg

∂Df

∂kf

∂kf∂pg− ∂cf∂Df

∂Df

∂kf

∂kf∂pg

. (B.5)

The sign of (B.5) can be positive or negative. First, this is negative when ∂kf

∂pg< 0 or

when ∂kf

∂pg> 0 with − ∂cf

∂Df

∂Df

∂kf

∂kf

∂pg>

∂kf

∂pg

∂Df

∂kf

∂kf

∂pg. Second, this is positive when ∂kf

∂pg> 0 and

− ∂cf

∂Df

∂Df

∂kf

∂kf

∂pg<

∂kf

∂pg

∂Df

∂kf

∂kf

∂pg.

C Graphs

FIT Program

Unregulated

0 50 100 150 200 250kd

50

100

150

200

250k f

Figure C.1: Simulations of the Ontario FIT program on the power generation equipmentequilibrium prices: unregulated vs. FIT

29

FIT without LCR

FIT with LCR

0 100 200 300 400 500 600 700kd

100

200

300

400

500k f

Figure C.2: Simulations of the Ontario FIT program on the power generation equipmentequilibrium prices: the LCR effect

FIT with LCR

FIT without LCR

20 30 40 50 60pg

-1200

-1000

-800

-600

-400

-200

200

Wd

Figure C.3: Simulations of the Ontario FIT program on domestic welfare

30

D Simulations

Characteristics Unregulated Unconditional FIT Conditional FITkd 83.3333 115.124 220.503kf 38.8889 51.4673 63.4429Dd 0.555556 0.575621 0.551257Df 0.347222 0.3386 0.290135Πd 38.5802 57.9844 113.957Πf 11.5741 15.5924 17.0602PSd 154.321 241.195 177.932PSf 24.1127 34.3951 16.8357Gd 20.0617 20.4976 19.9657Gf 3.76157 3.45479 3.52428

γG 23.8233 23.9524 23.49Wd 240.837 213.813 191.275WT 252.411 229.406 208.335

Table D.1: The effects of a FIT program on market characteristics(a = 0.01, Γd = 50, Γf = 40, γ = 1, λ = 1.2, p = 10, and pg = 15)

31

DOCUMENTS DE TRAVAIL GREDEG PARUS EN 2014GREDEG Working Papers Released in 2014

2014-01 Cristiano Antonelli & Francesco Quatraro The Effects of Biased Technological Changes on Total Factor Productivity: A Rejoinder and New Empirical Evidence2014-02 Michaël Assous & Roberto Lampa Lange’s 1938 Model: Dynamics and the “Optimum propensity to consume”2014-03 Agnès Festré & Pierre Garrouste Do People Stand by their Commitments? Evidence from Classroom Experiments2014-04 Samira Demaria & Sandra Rigot IFRS Standards and Insurance Companies: What Stakes for Long-Term Investment? A French Case Explanatory Analysis2014-05 Gérard Mondello Splitting Nuclear Parks or Not? The Third Party Liability Role2014-06 Frédéric Marty Towards an Economics of Convention-based Approach of the European Competition Policy2014-07 Giovanni Dosi, Giorgio Fagiolo, Mauro Napoletano, Andrea Roventini & Tania Treibich Fiscal and Monetary Policies in Complex Evolving Economies2014-08 Séverine Borderon La nature devenue projet de compensation écologique2014-09 Benjamin Montmartin & Marcos Herrera Internal and External Effects of R&D Subsidies and Fiscal Incentives: Empirical Evidence Using Spatial Dynamic Panel Models2014-10 Olivier Brette, Thomas Buhler, Nathalie Lazaric & Kevin Marechal Reconsidering the Nature and Effects of Habits in Urban Transportation Behaviour2014-11 Raphaël Chiappini Institutional Determinants of Japanese Outward FDI in the Manufacturing Industry2014-12 Olivier Bruno, André Cartapanis & Eric Nasica Bank Leverage, Financial Fragility and Prudential Regulation2014-13 Anaïs Carlin Consumer Choice Theory and Social Learning2014-14 Djamila Elidrissi, Valérie Hauch & Sabrina Loufrani-Fedida Compétences relationnelles et entreprises à internationalisation rapide et précoce : une approche multi-niveaux2014-15 Franck Léon La construction des Business Models des fournisseurs d’Infrastructure as-a-Service : l’étude de deux “purs players” français2014-16 Maëlle Della Peruta & Dominique Torre Virtual Social Currencies for Unemployed People: Social Networks and Job Market Access2014-17 Lauren Larrouy Challenging Standard Non-Cooperative Game Theory? From Bacharach’s “Variable Frame Theory” to “Team Reasoning”

2014-18 Cléo Chassonnery-Zaigouche & Lauren Larrouy Reshaping Standard Microeconomics for Political Action: Kenneth J. Arrow and Thomas C. Schelling’s Rand Corporation Projects on Racial Issues2014-19 Nabila Arfaoui Eco-innovation and Regulatory Push/Pull Effect in the Case of REACH Regulation: Empirical Evidence from Survey Data2014-20 Patrice Bougette & Christophe Charlier Renewable Energy, Subsidies, and the WTO: Where has the ‘Green’ Gone?

Renewable Energy, Subsidies, and the WTO: Where has the ...(Munich), the 2014 EALE conference...

Documents

Transcript of Renewable Energy, Subsidies, and the WTO: Where has the ...(Munich), the 2014 EALE conference...