Green Certificates and Market Power on the Nordic Power Market

Transcription

1 Green Certfcates and Market Power on the Nordc Power Market Erk S. Amundsen* and Lars Bergman** The purpose of ths study s to elucdate under whch crcumstances, how, and to what extent market power on a Tradable Green Certfcates (TGC) market can be used to affect an entre electrcty market. There are bascally two reasons for beng concerned wth ths. One s that a small number of companes may have exclusve access to frst rate stes for wnd power generaton. The other s that wthdrawal of a small number of TGCs mples a multple reducton of electrcty consumpton, wth correspondng ncreases of end user prces. We formulate both an analytcal model to nvestgate economc prncples and a numercal model based on that to nvestgate the Swedsh TGC market operatng n a settng of a common Nordc electrcty market. The analyss shows that Swedsh producers may exercse market power usng the TGC-market but that ths problem wll be elmnated by openng the TGC-market for other Nordc countres. Keywords: Renewable energy, Electrcty, Green certfcates, Market power, Nordc power market 1. INTRODUCTION In Sweden a market for Tradable Green Certfcates (TGCs) was ntroduced n The purpose was to stmulate nvestments n electrcty generaton based on renewable energy sources wthout usng drect governmental subsdes * Correspondng author. Department of Economcs, Unversty of Bergen, Fosswnckelgate 6, N- 5007, Bergen, Norway. E-mal: erk.amundsen@econ.ub.no, Phone: , Fax: ** Stockholm School of Economcs, Box 6501, SE Stockholm, Sweden. E-mal: Lars.Bergman@hhs.se. The authors are grateful to Gjermund Nese, SNF for complng cost data and updatng the numercal model, and to SNF, Elforsk and the Nordc Energy Research Program for fnancal support. Constructve comments by Rchard Green (on an earler verson of the paper) are gratefully acknowledged. The ntal steps n ths project were taken wthn the SESSA project, fnanced by the EU Commsson. The usual dsclamer apples. The Energy Journal, Vol. 33, No

2 102 / The Energy Journal to renewable energy. More precsely the am was to create a market where dfferent types of renewable electrcty can compete on equal terms, thus relevng governments and publc agences from beng drectly nvolved n power ndustry nvestment decsons. Lke any other market a TGC-market conssts of sellers and buyers. The sellers are the generators of green electrcty. 1 The generators obtan an amount of TGCs correspondng to the amount of green electrcty they feed-nto the network.e. one MWh of green electrcty generated gves rse to one TGC. The sellers thus get revenues both from sellng the electrcty on the electrcty wholesale market and from mmedately sellng the TGCs receved from the ssung body on the TGC-market. However, they may also decde to sell the certfcates at a later date, or even not at all. The buyers of TGCs are the consumers and retalng companes that are requred to acqure certfcates correspondng to a certan percentage of the total consumpton of electrcty ( the percentage requrement ). In other words the demand for TGCs s drectly derved from the demand for electrcty. A market clearng prce of TGCs s determned by the nterplay of supply and demand. An mportant aspect of a TGC system s that the percentage requrement creates a drect lnk between the electrcty market and the TGC-market. What happens n the electrcty market has a drect mpact on the TGC-market and vce versa. Moreover, as the demand for electrcty tends to be nelastc, the derved demand for TGCs s even more nelastc. Several studes have focused on varous mplcatons of the drect lnk between the two markets. 2 It seems to us that many of the problems dscussed n these studes were not well understood when TGC-markets were desgned and ntroduced. Another ssue that also seems to have been overlooked s the rsk for and mpact of market power n the TGC-market. The purpose of ths study s to elucdate under whch crcumstances, how, and to what extent market power n the TGC-market can be used to affect the entre electrcty market. There are bascally two reasons for beng concerned wth market power n TGC-markets. The frst s the fact that the ndustry average cost curve for green electrcty tends to be upward slopng. Ths s because the cost of wnd power, one 1. By green electrcty s meant electrcty produced by means of wnd-, sun-, wave- and geothermal energy as well as by bomass and peat and (new) small hydropower plants. 2. Amundsen and Mortensen (2001), as well as Amundsen and Nese (2004) show that an ncrease of the percentage requrement does not necessarly lead to an ncrease of the generaton of green electrcty and that harsher CO 2 constrants (tax or permts) wll defntely lead to less green electrcty wth ths TGC system. (See also Bye (2003) and Fscher (2009) on the effects on electrcty prces wth a TGC system). Amundsen, Baldursson and Mortensen (2004) analysed the volatlty of TGC prce stemmng from the annual varatons of wnd power. Several specfc features of mult-country TGC-markets have been studed by Fnon and Menanteau (2003), Bergman and Radetzk (2003), Nese (2003) and Unger and Ahlgren (2005), Butler and Neuhoff (2004), Søderholm (2008), Amundsen and Nese (2009). The relatonshp between a TGC-market and a CO 2 emsson permt market s nvestgated by Böhrnger and Rosendal (2010).

3 Green Certfcates and Market Power on the Nordc Power Market / 103 mportant source of green electrcty, depends on the wnd condtons at the ste of the plant, and that dfferent stes dffer sgnfcantly n ths respect. The stuaton s smlar for envronmentally frendly hydro power, and, to some extent, for other types of green electrcty. In the short and medum term the locaton of sutable stes n relaton to the exstng power grd mples that nvestments n green electrcty generaton, to a varyng extent, should nclude nvestments n transmsson capacty connectng the power plant to the exstng grd. In the case of bomass and peat other knds of transportaton nfrastructure nvestments may be needed. Thus, gven the state of technology and the upward slopng cost curve, there s a lmt on the amount of green electrcty that can be produced wthn a country. Ths means that some generators, by gettng access to the sutable stes, wll become domnatng producers of green electrcty and thus may be able to exercse market power n the TGC-market. Currently the number of green electrcty producers n Sweden s very large, but 96 percent of them produce less than 50 GWh per annum (18 percent of the total producton of green electrcty). On the other hand the three major power companes produce around 25 percent of all green electrcty, and on the bass of current nvestment plans ths share s lkely to ncrease sgnfcantly n the near future. The second reason for beng concerned wth market power n a TGCmarket s that, as a result of the percentage requrement, the wthdrawal of a gven number of TGCs from the market forces a much larger reducton of electrcty consumpton (e.g. f the percentage requrement s 10 the wthdrawal of one certfcate wll nduce a reducton of 10 MWh of consumpton). Thus, relatvely modest exercse of market power n the TGC-market may have a sgnfcant mpact on the prce of electrcty and the allocaton of resources n the power ndustry. Moreover, by havng access to a large share of the best stes for generaton of green electrcty a small power producer may be able to exercse market power n the electrcty market by exercsng market power n the TGCmarket. In order to elucdate the nterplay between the electrcty and TGC-markets a smple analytcal model s presented. Then a numercal verson of the model, depctng the Nordc electrcty market and a Norwegan-Swedsh market for TGCs s used to quantfy the mpact of TGC-market power under varous assumptons of trade and market power exerton. 2. AN ANALYTICAL MODEL The followng model s desgned to capture the nterplay between the electrcty and TGC-markets. The followng varables wll be appled p: End-use prce of electrcty s: TGC prce q: Wholesale prce of electrcty x: Quantty of total electrcty

4 104 / The Energy Journal y: Quantty of black electrcty z: Quantty of green electrcty w: Number of TGCs sold v: Number of TGCs used : Percentage requrement The nverse demand functon for electrcty s assumed gven by p p(x), wth p (x) 0 The cost functon for black electrcty for a gven producer, ( 1,...,n) s assumed gven by and c c (y ), wth c (y ) 0 c (y ) 0 The cost functon for green electrcty s assumed gven by and h (z ), wth h (z ) 0 h (z ) 0 The cost functons are taken to represent long run relatonshps. Otherwse, observe that both black and green electrcty are delvered to a common wholesale market, from where proft maxmzng retalng companes purchase electrcty for end-use delveres. In addton to the wholesale market there s also a market for TGCs where the producers of green electrcty are the sellers and the retalng companes / end-users are the buyers. To gan nsght nto the jont functonng of these two markets we consder two cases. In the frst we assume that the electrcty market s an olgopoly market where agents behave n accordance wth the Cournot model, whle there s perfect competton on the TGC-market. Ths case wll serve as a reference case for the analytcal model to follow. In the second case both markets are olgopolstc, and Cournot behavor s assumed. Both cases concern a sngle country. In order to prepare for our numercal analyss n a Nordc settng we thereafter expand the analytcal model to nclude trade n electrcty and TGCs. In the frst case we assume that the producers take the TGC prce as gven and maxmze profts as a Cournot player on the electrcty market. Clearly,

5 Green Certfcates and Market Power on the Nordc Power Market / 105 the mplcaton of ths s that the producers wll sell all TGCs generated. Hence, the maxmzaton problem of the producer may be formulated as MaxP (y z ) qy [q s]z c ( y ) h (z ) The frst order condtons read P y P z p(x) x q c ( y ) 0 p(x) x q s h (z ) 0 In dervng these condtons, we use q q q (p s) p, y z Hence, we assume that the producer recognzes that the sale of electrcty wll affect the wholesale prce of electrcty through ts effect on end-user demand. Also we assume for ths case that the TGC-prce, s s perceved as gven. The frst of the two frst order condtons states that the producer n generatng black electrcty wll equate the margnal revenue of sellng black electrcty wth the margnal cost of generatng black electrcty. Lkewse, the second condton states that the producer n generatng green electrcty wll equate the margnal revenue of sellng green electrcty and the correspondng TGC generated wth the margnal cost of generatng green electrcty. Hence, the producer has focus on the electrcty market and recognzes that he can nfluence the market prce of electrcty by choosng the level of output of black and green electrcty. Equlbrum condtons are p* q* s* z* x* y* z*, where x* x *, y* y *, z* z * p(x) x * q* c ( y *), p(x) x * q* s* h (z *), A more compact way of characterzng the soluton n market equlbrum may be obtaned by successve substtuton of the above condtons.e.

6 106 / The Energy Journal p p(x*) x * (1 )c (y *) h (z *), Ths condton states that the producer s margnal revenue n equlbrum s equal to a lnear combnaton of the ndvdual producer s margnal cost of provdng black and green electrcty, wth the percentage requrement as the combnaton weght. Thus, under market equlbrum the optmalty condtons for the producer smply bols down to the standard olgopoly condton except that the relevant margnal cost now s composed of two margnal cost elements. As the margnal cost of generatng green electrcty s larger than the margnal cost of generatng black electrcty n optmum t s also clear that ths s not a least cost soluton of generatng electrcty as ths would necesstate equal margnal cost of generaton for the two knds of electrcty. Next we turn to the second case of ths secton and consder an electrcty producer that behaves as a Cournot player n both markets. 3 In so dong we assume that the producer recognzes that hs sale of TGCs wll nfluence the end-user prce of electrcty and hence behaves accordngly. In part the producer has to decde on how many TGCs to sell, w. The producer s, however, constraned by the amount of TGCs generated,.e. we have w z The electrcty producer, thus, faces the followng optmzaton problem MaxP ( y,z,w ) q[ y z ] sw c( y ) h(z ), s.t. w z To solve ths problem, formulate the Lagrangan functon L ( y,z,w ) q[ y z ] sw c( y ) h(z ) k [w z ] The frst order condtons are L y p(x) x q c ( y ) 0 3. We use standard Cournot assumptons for multple markets,.e. we assume that the producer takes the prces n the other markets as gven when operatng n one of the markets. Hence, we do not assume what has been called smultaneous gamng. For a dscusson of problems related to smultaneous gamng wth a TGC-market of the Nordc type, see Amundsen and Nese (2004). Otherwse, market power of nterconnected markets (e.g. emsson permt markets and energy markets) s dealt wth n Montero (2009). Furthermore, Traber and Kemfert (2009) analyze the mpacts of the German Feed-n tarffs on both electrcty prces, emssons and frms whle takng account of market power and olgopolstc behavour.

7 Green Certfcates and Market Power on the Nordc Power Market / 107 L z p(x) x q k h (z ) 0 L 1 p(x) w s k 0 w In dervng these condtons we use p q s 1 p(x), w w Hence, we assume that the producer recognzes that hs sale of TGCs splls over on the electrcty market and nfluences the end user prce of electrcty. Thus, f the producer sells one extra TGC he perceves that the effect on total electrcty demand s equal to 1/ that results n a prce drop of (1/ )(dp/dx) n the electrcty market. Observe that f the producer does not fnd t optmal to sell all TGCs receved (.e. w z), we must have k 0. Elmnatng k * the equlbrum condtons may be expressed as p* q* s* w* x* y* z* p(x*) x * q* c (y *), p(x*) w * * x * q* s* h (z *), Comparng the last two of these condtons wth the correspondng condtons of the prevous case, we see that the optmalty condton for black electrcty generaton s the same, whle the optmalty condton for green electrcty generaton has been changed. In partcular, the negatve element of the margnal revenue expresson has become an extra element, mplyng that the producer now recognzes that hs sale of TGC trggers an ncrease of demand equal to ( w */ ) that wll affect the electrcty prce negatvely. Hence, the producer recognzes that hs sale of TGCs also affects the market prce negatvely n addton to the negatve prce effect of hs electrcty sale. Consequently, he cuts back on hs electrcty generaton as compared wth what he would have had wthout market power on the TGC-market. Upon successve substtuton of the above condtons, the followng compact relatonshp appears

8 108 / The Energy Journal p(x*) p(x*) [x * w *] (1 )c ( y *) h (z *), Ths condton states that the producer s margnal revenue from the electrcty and the TGC-market n market equlbrum s equal to a lnear combnaton of the ndvdual producer s margnal cost of provdng black and green electrcty, wth the percentage requrement as the combnaton weght. By comparng ths condton wth the prevous condton where market power on the TGC-market was not exercsed, we see that the margnal cost concept s the same, but as referred to above, that the margnal revenue expresson has been altered. Thus, the negatve element of the margnal revenue expresson s enlarged when the producer possesses market power on the TGC-market n addton to market power on the electrcty market. 4 The effect of ths n the market s to ncrease the end user prce of electrcty and to reduce the consumpton of electrcty. Hence, exerton of market power on the TGC-market splls over to the electrcty market n terms of lower electrcty generaton and a hgher end user prce. 3. A TWO-COUNTRY MODEL The objectve of ths part of analyss s to nvestgate the jont functonng of an electrcty and a TGC-market n a Norwegan-Swedsh settng. Hence, we next expand the analytcal model to consder two countres, A and B, (such as Norway and Sweden) that trade electrcty and/or green certfcates wth each other. In ths settng polcy measures taken n one country may effect the decsons made n the other country. Under autarky each country wll have to satsfy the same general set of equlbrum condtons as dscussed above. Prces and quanttes wll, however, be specfc to each country as determned by the demand functon, the cost functons and the polcy measure appled (percentage requrement). Openng for trade n electrcty whle stll keepng separate TGC-markets, the electrcty wholesale prce wll become the same n both countres. Thus t holds that q A* q B* q*. In ths settng black and green electrcty flow freely between the two countres, whereas TGCs can only be sold n the country where they orgnate. Snce both the TGC prce and the percentage requrement may dffer between the countres, end user prces may also dffer. As always, wth ntegraton of markets, market power gets dluted due to the smaller relatve sze of each producer n the common market as compared wth the relatve sze under autarky for each of the jonng countres. For the case where market power s not exercsed on the TGC-markets the effect of openng for trade n electrcty whle stll keepng separate TGC-markets s smply to reduce market power on the 4. The consequence of ths s that the relevant margnal revenue curve becomes steeper. Wth a lnear demand curve, the margnal revenue curve pvots nwards around the pont where the demand curve hts the prce axs.

9 Green Certfcates and Market Power on the Nordc Power Market / 109 electrcty market. The consequence of ths s an ncrease of the jont generaton of electrcty n the two countres as compared wth what t was pror to market ntegraton. In the above case where market power on the TGC-markets s not exercsed and where the TGC-markets are separate wthout any trade a producer can not exercse market power on the electrcty market specfcally related to hs own country. Ths s, however, stll an opton f market power s exercsed on the separate TGC-markets. Consderng the case of market power exerton n both markets but wth natonal TGC markets the followng condtons apply: p * q* s *, j A,B j j j x * * y * z * y * z * A B A A B B x * y * z *, j A,B j j j p j(x j*) j x * q* c (y *),, j, j j j j A,B p j(x j*) w j* x j* q* s j* hj (z j*),, j j A,B j j j j j j In the above relatonshps we have x * *.e. x * s the aggregate electrcty generaton n country j A,B. Lkewse, we have y * y * for black electrcty and j z * z * for green electrcty. j As s seen from the relatonshps above, the decsons of the sngle producer are equal to what they were pror to market ntegraton. Hence, a producer of green electrcty wll sell hs TGCs n order to nfluence the TGC-prce and the end user prce of electrcty n hs own country. Consequently, szable market power may stll persst n a country even after ntegraton of the electrcty market has taken place. Otherwse, t may be nterestng to note that even for ths case a compact verson of the producer s producton decsons n market equlbrum may be derved from the above condtons.e. p j(x j*) p j(x j*) [x j* w j*] (1 j)cj ( y j*) j hj (z j*),, j j The nterpretaton of ths condton s the same as the one gven for the correspondng expresson n the prevous secton. It should be observed that the composton of margnal cost n terms of weght (the percentage requrement) put on the two cost components s the same for every producer n a gven country, but j

10 110 / The Energy Journal dfferent between countres f the percentage requrements dffer. The reason for ths s that TGCs wll only be sold n the country where they orgnate. The next step s to open for trade also for TGCs n addton to the electrcty trade. The effect of ths s to equate TGC prces so that they become the same for both countres.e. so that s A* s B* s* where s* s some ntermedate level as com- pared wth the ntal TGC prces of the two countres. In order to gve an analytc llustraton of ths case there s a need for an addtonal varable and an addtonal equaton (statng that total certfcates appled must be equal to total certfcates sold) to descrbe the equlbrum soluton. Hence, we descrbe the number of certfcates used n country j by v j. Net mport of certfcates for country j s then equal to v w j j. The set of equlbrum condtons now reads p j* q* js*, j A,B x * * y * z * y * z * A B A A B B x * y * v *, j A,B j j j v * v * w * w * A B A B p(x*) x j * q* c j ( y j *),, j, j A,B p(x*) w j * x j * q* s* h j (z j *),, j, j A,B j The nterpretaton of these condtons s smlar to the nterpretaton of the condtons of the prevous case consdered. There s, however, one mportant dstncton: as the TGCs may be used n both countres the sale or wthdrawal of an addtonal TGC does not only affect the end user prce n the country where the sale s takng place but ndeed n both countres. Therefore, market power gets dluted due to the smaller relatve sze of each producer of green electrcty n the common TGC-market. Hence, market ntegraton of the TGC-markets wll result n smaller market power exerton showng up as an ncrease of the jont generaton of electrcty n the two countres as compared wth what t was for the case where only electrcty was traded. Otherwse, t should be observed that end user prces may stll dffer between the two countres f the percentage requrements are dfferent. It should also be observed that the relatve weght on the margnal cost of black and green electrcty generaton now s the same n market equlbrum for all producers rrespectve of where producton s takng place. The combnaton weght wll be ntermedate to the level of each country s percentage requrement and depends on the sze of the markets.

11 Green Certfcates and Market Power on the Nordc Power Market / 111 A specal case of the analytcal model apples f only one of the two countres has a TGC system and generators n the other country are allowed to partcpate n that system. For nstance, Norwegan generators of green electrcty could receve Swedsh TGCs and be allowed to partcpate n the Swedsh market for TGCs. Ths s the case we set out to nvestgate n the followng. 4. A NUMERICAL MODEL In order to shed some lght on the real world stuaton a numercal model wth the same basc structure as the analytcal model has been used. The model s an updated verson (see Amundsen and Nese, 2009) of the statc model of the Nordc electrcty market,.e. the ntegrated electrcty market of Denmark, Fnland, Norway and Sweden, ntally used n Bergman and Radetzk (2003). The basc features and assumptons of the numercal model are as follow: In each country there are 2 5 major electrcty generaton frms, actng as Cournot players, and a number of small generaton frms formng a compettve frnge. There s no entry of new frms, but the ncumbent frms may nvest n new capacty. Margnal cost curves are step-wse ncreasng and lnear, reflectng unt costs and capacty lmts for varous technologes. 5 Green electrcty encompasses electrcty generated by wnd, water and bomass. However, only electrcty generaton n new small hydropower plants are consdered green, whereas electrcty generaton n exstng hydropower plants are consdered black 6 just as electrcty generated n nuclear-, gas-, coal- and ol power plants. Free-trade n electrcty between the four Nordc countres, but nterconnector capacty lmts may lead to dfferent wholesale prces n the varous countres Two alternatve trade regmes for the TGC-market: Autarky or free trade between the four Nordc countres Two alternatve assumptons about the behavor of the major frms n the TGC-market: Prce takng (perfect competton) or Cournot behavor. The frnge frms always behave as prce takers both n the electrcty and the TGC-market. Constant elastc demand curves n each country (prce elastcty: 0,3) 5. The cost of producton n an electrcty generaton unt s equal to the varable cost n exstng unts and the sum of the varable cost and the annualzed captal cost n a new unt. 6. For Norway black electrcty generaton s almost exclusvely takng place n water power plants whereas black electrcty generaton n Sweden also ncludes electrcty from nuclear-, gas- and coal power plants.

12 112 / The Energy Journal Usng 2001 as ts base year the model projects annual equlbrum prces and quanttes n, and cross border tarffs between, the electrcty markets n Denmark, Fnland, Norway and Sweden. It also projects equlbrum prces and trade n the markets for TGCs n Norway and Sweden. In the Swedsh system the producers of certfcate-enttled electrcty can be essentally classfed nto two groups: one consstng of a few large companes, whch produce most of the electrcty, and the other consstng of a large number of producers who ndvdually have relatvely small producton quanttes (Swedsh Energy Agency, 2009). In accordance wth ths we splt the producers n two groups and assume that the rghts to explot the most favorable stes for wnd power producton have been acqured by the major power producers. Thus a frnge power producer who wants to enter the market for green electrcty only has access to more costly wnd power stes,.e. stes where wnd condtons and/ or the cost of connectng to the grd s hgher than for the stes avalable for the major generators. Unfortunately lack of avalable data on ownershp prevents us from usng exact nformaton and the above assumptons may turn out to be too strong and thus exaggerate the extent of market power of the natonal Swedsh TGC market. Nevertheless, there are clear ndcatons that the major producers do have a sgnfcant share of the new wnd power stes elgble for TGCs n Sweden. For nstance, Vattenfall owns the large Lllgrund Wnd Farm, whle one of the largest land-based wnd farms n Europe currently under constructon, Blaken Vnd AB, s owned by Fortum and Skellefteå Kraft, the second and fourth largest power producers n Sweden. Ths project alone accounts for some 15 percent of the total nstalled wnd capacty n Furthermore, Vattenfall s about to assess the project Kregers Flak 2 wth a capacty of 2.5 tmes that of Blaken. In the followng we wll use the numercal model to analyze to what extent a requrement to have a certan percentage of green electrcty n Sweden affects the Nordc electrcty market. We consder both autarky and free trade of TGCs. Moreover we analyze to what extent that market power n the TGC-market affects the electrcty market. The case of no TGC-market s consdered as a benchmark. The focus s on the year As the effects of the Swedsh TGC system are smlar n the three other Nordc countres we only present the results for Norway (beng a sgnfcantly larger natonal electrcty market than Denmark and Fnland together). 5. SIMULATION RESULTS Usng the numercal model we have consdered three alternatve cases, or projectons and compared the outcome for prces and quanttes wth the correspondng data for the benchmark case (denoted Base 2010). Note that the Base case s also a projecton, and that electrcty consumpton ncludes transmsson and dstrbuton losses. The dfferent cases are defned n the followng way:

13 Green Certfcates and Market Power on the Nordc Power Market / 113 Table 1: Results of the Numercal Model Base 2010 Case 1 Case 2 Case 3 Electrcty consumpton n Norway, TWh Electrcty consumpton n Sweden, TWh Producer prce of electrcty n Sweden, /MWh Producer prce of electrcty n Norway, /MWh Prce of TGCs, /MWh Consumer prce n Sweden, /MWh Base 2010 In ths case there s free trade n electrcty between the Nordc countres,.e. Denmark, Fnland, Norway and Sweden. There s no TGC system n Sweden or n any other Nordc country. Case 1 Agan free trade n electrcty between the Nordc countres s assumed, but now a TGC system wth the percentage requrement equal to s ntroduced n Sweden. Perfect competton n the Swedsh TGC-market s assumed, but there s no cross-border trade n TGCs. Case 2 Ths case s lke Case 1, except that the major generators act as Cournot players n the market for TGCs Case 3 Ths case s lke Case 2, except that generators n Denmark, Fnland and Norway are allowed to partcpate n the Swedsh TGC system, and to trade TGCs across the Nordc borders. We present the smulaton results by par-wse comparsons of the four alternatves. The key results are summarzed n Table 1 above. Base 2010 vs Case 1 In order to analyze the mpact of a TGC system on the electrcty market we compare Base 2010 and Case 1. Note that n Case 1 the Swedsh TGC-market 7. The offcal percentage requrement for 2010 s 17.9 percent. However, as electrcty ntensve ndustres are exempt from the requrement to use a certan fracton of green electrcty the average percentage requrement on all electrcty consumpton s 12.7.

14 114 / The Energy Journal s assumed to be perfectly compettve,.e. no generator can, or decdes to, exercse market power on the TGC-market. Also note that producer prces n Sweden are equal to producer prces n Norway. As can be seen n the table the ntroducton of a TGC system leads to a reducton of the producer prces of electrcty,.e. the system and area prces determned at the common Nordc power exchange Nord Pool. Ths s well n lne wth theory as one would expect producer prces to fall as a consequence of ntroducng what amounts to a tax on black electrcty generaton. The consumer prce of electrcty n Sweden, however, ncreases as a result of the percentage requrement and the postve prce of TGCs. Thus electrcty consumpton ncreases n Norway (where producer and consumer prces are the same n the model) and decreases n Sweden. 8 The prce of TGCs n Case 1 s somewhat lower than the actual Swedsh TGC prces observed , whch vared wthn the range /MWh. For the perod , when the percentage requrement was lower, the prces projected by the model were very close to the observed prces. The somewhat hgher prces n 2008 and 2009 could be a result of market power beng exercsed, but t could also reflect relevant costs not fully taken nto account by the model. 9 Case 1 vs Case 2 Next we analyze the potental mpact of market power on the Swedsh market for TGCs,.e. we compare Case 1 and Case 2. As can be seen n Table 1 there s consderable potental market power on the TGC-market. If ths market power s fully exercsed the equlbrum prces of TGCs would be several tmes hgher than the prces htherto observed. Moreover, the mpact on the electrcty market would be sgnfcant, reducng electrcty consumpton n Sweden by around 15 percent. The numbers may seem unrealstc, but they clearly ndcate that the percentage requrement, whch s the core feature of a TGC system, opens up a new possblty to exercse market power n the electrcty market. One reason for the sgnfcant potental market power created by the TGC system s that the TGC-market s a qute small natonal market wth a small 8. The effect on producer prces of ntroducng a TGC system s extensvely dealt wth n the lterature. Hence, Amundsen and Mortensen (2001) as well as Bye (2003) show that an ncrease of the percentage requrement always results n a reduced producer prce n a jont compettve electrcty and TGC-market. However, the effect on end-user prce and total electrcty consumpton s ndetermnate. 9. Only a few emprcal analyses exst for the Swedsh TGC system e.g. Kåberger et al. (2004), Knutsson et al. (2006) and Bergek and Jacobsson (2010). Two of these are rather crtcal to the functonng of the system. Hence, Kåberger et al. fnd that the Swedsh TGC system s not really effcent whle Bergek and Jacobsson fnd that the Swedsh TGC market functons as a rent-generatng machne that only to a small extent stmulates techncal change. None of the studes address the queston of market power. Otherwse nformaton on the functonng of the Swedsh market s reported annually by the Swedsh Energy Agency.

15 Green Certfcates and Market Power on the Nordc Power Market / 115 number of relatvely large players. Consequently t s an olgopolstc market n whch each player has sgnfcant market power. As dscussed n the analytcal model one way of dealng wth that problem s to enlarge the market and open up for nternatonal trade n TGCs. Case 3, where generators n the other Nordc countres are allowed to partcpate n the Swedsh TGC-market, represents one possblty along these lnes. Case 3 vs Case 2 and Case 1 Case 3 mples that, from a Swedsh polcy perspectve, green electrcty generated n Denmark, Fnland and Norway s seen as a perfect substtute for green electrcty generated n Sweden. Consequently TGCs are allocated to generators n the other Nordc countres on the bass of the same prncples as TGCs are allocated to Swedsh generators. Moreover, all generators n the Nordc countres can sell ther TGCs n the Swedsh TGC-market. Needless to say a full verson of the Case 3 model has not been serously consdered by polcy makers n the Nordc countres. However, t llustrates a serous drawback of the current Swedsh TGC-market. As can be seen n the table the addton of new partcpants on the Swedsh TGC-market ncreases competton and sgnfcantly reduces the possbltes to exercse market power. In fact, openng up for cross-border trade wth TGCs (Case 3) essentally has the same effect on the electrcty market as perfect competton n a natonal Swedsh market for TGCs (Case 1). 6. CONCLUDING REMARKS In connecton wth the restructurng of the electrcty markets n the Nordc countres market power was seen as a major potental problem. The reason was that each one of the natonal markets (n Denmark two dsconnected regonal markets) was domnated by one major power company. However, by ntegratng the natonal markets ths problem was sgnfcantly mtgated. Our results ndcate that market power s a potental problem n the Swedsh market for TGCs, but agan the soluton seems to be market ntegraton. That s, to open up the Swedsh market for green power producers n the other Nordc countres. The argument for dong ths becomes even stronger f a TGC system s ntroduced n the other Nordc countres. In fact, ths has recently happened as a jont Swedsh-Norwegan TGC market was establshed on January 1, An alternatve to applyng a TGC-market for stmulatng the generaton of renewable electrcty s to use a system of so called feed-n tarffs whch smply amounts to a subsdy per unt of green electrcty produced. Ths s a system that s n use n many countres (e.g. Germany and Denmark). There are smlartes as to how a TGC market and a system of feed-n tarffs functon but there are also dfferences. One smlarty s that both systems provde the producer wth a subsdy for each unt of green energy produced. However, as the feed-n tarff

16 116 / The Energy Journal system nvolves governmental bodes n determnng the sze of the subsdy ths s left to the market (as determned by the percentage requrement) for the TGCsystem. Another mportant dfference s that the feed-n tarff system avods the problem of market power exerton as the sze of the feed-n tarffs can not be drectly nfluenced by the producers. On the other hand, a feed-n tarff system s not well suted to stmulate cooperaton among countres n nstallng new capacty of renewable energy. Hence, f two countres have a common target of achevng a certan share of renewable energy for both countres vewed as one, a common TGC-system may realze mutual gans of cost reductons by lettng the country wth the lowest cost take a larger share than the other whle mplctly beng pad by the other country n terms of TGCs receved from that country. Ths s not easly done wth a feed-n tarff system as ths would mply a system of drect cross border subsdes. In concludng ths paper t should further be noted that the success of ntroducng TGC-markets not only depends on the ablty to mtgate market power but also on other factors. One potental problem s the volatlty, both short term and on an annual bass, of TGC prces resultng from natural varatons n wnd condtons (see Amundsen, Baldursson and Mortensen (2006)). In Denmark, for nstance, the supply of wnd power may vary by 25 percent (compared wth the annual average) between wndy and calm years. As the margnal cost of wnd power generaton s close to zero for exstng capactes compettve wnd power generators wll at all tmes produce what s feasble and thus generate erratc and prce nelastc supply. Hence, the number of TGCs ssued and avalable for sale wll also be hghly volatle and ths wll lead to a consderable uncertanty wth respect to the remuneraton (.e. the sum of the TGC prce and the wholesale prce) of nvestment n green technologes. Ths n ts turn may also nfluence the requred rate of return for nvestors n renewable electrcty.e. the requred rate of return would be hgher as compared wth what would be necessary f subsdes were stable and certan. However, an ntegraton of the Swedsh TGC-market wth other Nordc TGC-markets would be a soluton, or at least a remedy, to ths problem. Ths s because wnd varatons n varous parts of the Nordc area probably are not closely correlated. Hence, an extenson of the TGC-market would stablze TGC prces. REFERENCES Amundsen, E.S. and J.B. Mortensen (2001). The Dansh green certfcate system: Some smple analytcal results, Energy Economcs, 23, Amundsen, E.S. and G. Nese (2004). Market Power n Interactcve Envronmental and Energy Markets: The Case of Green Certfcates, Workng paper, Department of Economcs, The Unversty of Bergen.

17 Green Certfcates and Market Power on the Nordc Power Market / 117 Amundsen, E.S., M. Baldursson and J.B. Mortensen (2006). Prce volatlty and Bankng n Green Certfcate Markets. Envronmental and Resource Economcs, 35: /s Amundsen, E.S. and G. Nese (2009). Integraton of tradable green certfcate markets: What can be expected?, Journal of Polcy Modelng, 31, Bergek, A. and S. Jacobsson (2010). Are tradable green certfcates a cost-effcent polcy drvng techncal change or a rent-generatng machne? Lessons from Sweden , Energy Polcy, 38, Bergman, L. and M. Radetzk (2003). Global klmatpoltk, SNS Förlag, Stockholm. Böhrnger C. and K.E. Rosendahl (2010). Green Serves the Drtest. On the Interacton between Black and Green Quotas, Journal of Regulatory Economcs, 37, /s Butler, L. and K. Neuhoff (2004). Comparson of Feed-n Tarff, Quota and Aucton Mechansms to support Wnd Power Development, Workng Paper Department of Appled Economcs, Unversty of Cambrdge. Fnon, D. and P. Menanteau (2003). The Statc and Dynamc Effcency of Instruments of Promoton of Renewables, Energy Studes Revew, 12(1): Fscher, C (2009). Renewable Portfolo Standards: When Do They Lower Energy Prces? The Energy Journal, 31(1): Montero, P. (2009). Market Power n Polluton Permt Market, The Energy Journal, 30 (Specal Issue 2), , Kåberger, T, T. Sterner, M. Zamanan and A. Jurgensen (2004). Economc Effcency of Compulsory Green Electrcty Quotas n Sweden, Energy and Envronment, 15(4): Knutsson D., S. Werner and E. O. Ahlgren (2006). Combned heat and power n the Swedsh dstrct heatng sector mpact of green certfcates and CO 2 tradng on new nvestments, Energy Polcy 34, Nese, G. (2003). Essays n Lberalzed Energy Markets Doctoral dssertaton, Department of Economcs, The Unversty of Bergen. Søderholm, P. (2008). The poltcal economy of nternatonal green certfcates markets, Energy Polcy 36, Swedsh Energy Agency (2009). The electrcty certfcate system, 2009, Stockholm. Traber, T. and C. Kemfert (2009). Impacts of the German Support for Renewable Energy on Electrcty Prces, Emssons, and Frms The Energy Journal, 30(3): /ISSN EJ-Vol30-No3-8. Unger, T. and E.O. Ahlgren (2005). Impacts of a common green certfcate market on electrcty and CO 2 emsson markets n the Nordc countres Energy Polcy, 33(16), /j.enpol

18