grØn energi innovation i dan- mark 2011... · rammevilkår for innovation, dels forbedring af...

GRØN ENERGI INNOVATION I DAN-MARK ERHVERVSPOTENTIALE FRA GLOBALE UDFORDRINGER | NOVEMBER 2010

Grøn energi innovation i Danmark

2

KOLOFON

Team: Managing Economist Helge Sigurd Næss-Schmidt(project manager), Economist Ulrik Møller, Analyst Holger Jensen, Research Assistant Mette Kildegaard

Kunde: Dansk Energi

Dato: 27. september

Kontakt: SANKT ANNÆ PLADS 13, 2. SAL | 1250 KØBENHAVN TELEFON: 2333 1810 | FAX: 7027 0741 WWW.COPENHAGENECONOMICS.COM


3

Kapitel 1 Sammenfatning ....................................................................................... 4

1.1. De nationale rammevilkår ..................................................................................... 4

1.2. De internationale rammevilkår .............................................................................. 6

Kapitel 2 De globale rammer ............................................................................... 10

Kapitel 3 Innovationspolitikken i Danmark: tre fokusområder ............................ 15

3.1. Styrkelse af midler til forskning, udvikling og innovation..................................... 15

3.2. Øget fokus på innovation og forøgelse af offentlige FU-midler ............................ 17

3.3. fokus på danske styrkepositioner i internationalt perspektiv ................................. 20

Kapitel 4 Internationale indsatsområder: fire fokusområder ................................ 23

4.1. EU's kvotesystem og global klimapolitik .............................................................. 23

4.2. Begyndende harmonisering af EU's VE-politik .................................................... 27

4.3. Styrkelse af EU's forskningsbudget for energirelateret forskning ........................... 29

4.4. Liberalisering af elmarkeder og investeringer i gridkapacitet ................................. 29

Appendiks - Kilder til tabeller og figurer .................................................................... 31

Litteraturliste … ......................................................................................................... 34

INDHOLDSFORTEGNELSE


4

Den globale produktion og anvendelse af energi vil undergå en radikal omlægning i de kommende årtier. En stigende efterspørgsel efter kul, gas og olie fra lande som Kina og Indien vil fastholde vedvarende høje priser på fossile brændstoffer ligesom disse lande også er blevet opmærksomme på gevinsten ved øget ener-gieffektivitet. Samtidigt vil nationale, regionale og globale klima- og energipolitikker kræve reducerede ud-ledninger af CO2 fra energiforbrug og mindre afhængighed af importeret gas og olie fra potentielt ustabile lande. Svaret på disse udfordringer er grøn innovation dvs. mere effektiv udnyttelse af energien og udvikling af low carbon energikilder. Det kræver omstilling i den danske energisektor men giver også nye muligheder og markeder, hvis den er-hvervspolitiske strategi gribes rigtigt an med vægt på to grundprincipper. For det første, at Danmark er et lille land med et begrænset hjemmemarked. Forceret satsning på produkter, der reelt kun kan sælges i Danmark er ikke vejen frem: eksportandelen for de danske frontløbere på energiområdet er ofte over 75 procent. Det er derfor afgørende, at Danmarks er opmærksomme på potentialet ved at udnytte EU´s indre marked som ”hjemmemarked” for vores teknologier, herunder som springbræt til de hastigt voksende ”grønne” markeder uden for EU. For det andet er det afgørende, at Danmark ligger højt i værdikæden, dvs. at det er højproduktive vidensin-tensive job der skabes snarere end ufaglærte industriarbejdspladser: Sidstnævnte ender før eller siden ender med at flytte ud til lande med langt lavere lønniveauer. Erfaringer fra Tyskland med støtte til solceller viser risikoen med en sådan entydig satsning på hurtigt vækst i volumen i stedet for langsigtet kvalitet. Vejen frem for grøn innovation bør derfor funderes på en to-strenget strategi dels styrkelse af nationale rammevilkår for innovation, dels forbedring af internationale rammevilkår for dansk afsætning af grønnere energiproduktion og forbrug. De nationale rammevilkår skal tænkes tæt sammen med de internationale dit-to. Der tænkes her især på samspillet mellem VE-støttesystemer og energiforskningsprogrammer i Danmark og EU.

1.1. DE NATIONALE RAMMEVILKÅR

1.1.1 Sammenhæng mellem kortsigtet driftsstøtte og langsigtet innovation

En styrkelse af den langsigtede innovation kræver en indsats over en bred front, hvor man støtter udvikling hele vejen fra ide til faktisk masseproduktion, der kan erstatte fossile brændstoffer. Støtten skal skrues sam-men så den passer til hvor ”moden” ideen her, men det afgørende er at innovation handler om at tiltag, der forbedrer teknologier og processer. På den lange bane er der tale om midler til grund- og anvendt forsk-ning. I næste fase handler det om at få gjort produkter mere konkurrencedygtige: det kræver at man stiller krav til højere ydelser ved offentlige udbud af VE-projekter f.eks. havvindmøller eller biogasanlæg og giver danske producenter muligheder for at teste produkterne i stærke forskningsmiljøer. Det er i den sammenhæng vigtigt, at der findes den rette balance mellem driftsstøtte til udrulning af VE og satsning på langsigtet udvikling af morgendagens VE-teknologier. Danmark har allerede forpligtiget sig til

Kapitel 1 SAMMENFATNING


5

at opnå en VE andel på 30 procent i 2020: Det kan ikke leveres uden en ganske betydelig driftsstøtte i de kommende 10 år medmindre EU skruer betydeligt op for sine klimaambitioner med højere priser på EU kvoter og fossile brændsler til følge (driftsudgifter i EU i 2020 på op til 500 til 600 mia.kr. alene for el og kraftvarme). I dag gives ca. 4/5 af VE-støtten i Danmark som driftsstøtte. Det er næppe hensigtsmæssigt at øge denne andel yderligere. Også VE-teknologier har en produktionscyklus kendt fra andre industriprodukter, hvor selve den industrielle produktion med tiden bedre kan produceres i lande med meget lave lønninger. Erfa-ringer fra Tyskland med støtte til solcelle produktion er sigende: efter at have brugt knap 400 milliarder kroner i driftsstøtte over de sidste 10 år oplever tyske producenter af solceller faldende produktion og eks-port i kroner og ører medens Kina og Taiwan er endt med at have 50 procent af verdensmarkedet stort set uden hjemlig afsætning indtil nu. Så reelt har høj støtte til solceller i Tyskland, Spanien mv. ikke mindst bidraget til at opbygge en særdeles konkurrencedygtig produktion i Kina og Taiwan betalt af tyske elbruge-re. Derfor er det vigtigt, at kortsigtet driftsstøtte til markedsmodne teknologier suppleres med langsigtede inve-steringer i forskning, innovation og uddannelse, som det private hverken kan eller skal løfte alene. Gevinster ved forskning og udvikling er meget usikre og gavner samfundet generelt, og ikke kun de virksomheder, der finansierer det. OECD beregninger tilsiger at der er behov for 3-4 dobling af midler til forskning og udvik-ling over de næste 10-20 år. Indsatsen i Danmark er efter de sidst års stigninger nogenlunde på niveau med sammenlignelige lande men langt fra det krævede niveau. AnbefalingAnbefalingAnbefalingAnbefalingerererer: : : :

� Danmark skal også inden for VEDanmark skal også inden for VEDanmark skal også inden for VEDanmark skal også inden for VE----teknologierneteknologierneteknologierneteknologierne satser på at ligge højt i værdikæden og satser på at ligge højt i værdikæden og satser på at ligge højt i værdikæden og satser på at ligge højt i værdikæden og konkurrerer på de produkter hvor det danske omkostningskonkurrerer på de produkter hvor det danske omkostningskonkurrerer på de produkter hvor det danske omkostningskonkurrerer på de produkter hvor det danske omkostnings---- og kompetenceniveau giver rog kompetenceniveau giver rog kompetenceniveau giver rog kompetenceniveau giver re-e-e-e-elle muligheder for at deltage i den internationale konkurrence.elle muligheder for at deltage i den internationale konkurrence.elle muligheder for at deltage i den internationale konkurrence.elle muligheder for at deltage i den internationale konkurrence.

� En flerdobling af det danske innovationsbudget på energiområdet pEn flerdobling af det danske innovationsbudget på energiområdet pEn flerdobling af det danske innovationsbudget på energiområdet pEn flerdobling af det danske innovationsbudget på energiområdet på omkring 1 mia.kr. å omkring 1 mia.kr. å omkring 1 mia.kr. å omkring 1 mia.kr. over de næste 10 årover de næste 10 årover de næste 10 årover de næste 10 år

1.1.2. Prioritering af de næste 10 års innovationsmidler på energiområdet

Prioritering af den danske innovationsindsats må tage udgangspunkt i analyserne: det vil sige lige vægt på at fremme effektivisering af energiforbrug og styrkelse af langsigtet udvikling af VE-teknologier, og den hertil hørende infrastruktur. Hertil kommer de konkrete udfordringer det danske energisystem står overfor med integration af stærkt sti-gende og volatil vindmøllebaseret strøm i det nordvesteuropæiske elsystem: ikke kun Danmark bygger ud, det gør Tyskland, UK, Ireland, Frankrig og det øvrige Skandinavien også. Endeligt må indsatsen fokusere på danske styrkepositioner: et land på 5 millioner mennesker kan ikke være førende på samtlige områder. Indenfor energieffektivisering er det f.eks. kraftvarmeproduktion, termostats-teknologi(Danfoss), isolering(Rockwoll), fjernvarmerør (Løgstør) og energirådgivning(Rambøll og energi-selskaberne). For energieffektivisering er udfordringen ofte at realisere rentable projekter: nye projektmodel-ler til håndtering af forskellige risici og barrierer har vundet indpas i udlandet og på vej til Danmark. Der


6

ligger et betydeligt potentiale for energirådgivning på dette område som også kan kommercialiseres uden for Danmark. Indenfor VE er vindenergi, bioetanol, biomasse og på længere sigt brændselsceller velkendte styrkepositioner. Endeligt peger nogle analyser også på at finansiering for fremtidige lovende og rentable teknologier kan væ-re en bremse. Her er ikke så meget behov for støtte: pointen er at projekterne som udgangspunkt er meget tæt på at være rentable på markedsvilkår: det offentlige kan være fødselshjælper ved at bidrage til kapitelind-skud som en investor, der får del i både underskud og overskud. Vækstfonden med finansiering fra det of-fentlige bygger på denne forretningsmodel. Anbefalinger:Anbefalinger:Anbefalinger:Anbefalinger:

� Demonstrationsprojekter Demonstrationsprojekter Demonstrationsprojekter Demonstrationsprojekter indenfor havvindmøller, bioethanol, elbiler og affald (både huindenfor havvindmøller, bioethanol, elbiler og affald (både huindenfor havvindmøller, bioethanol, elbiler og affald (både huindenfor havvindmøller, bioethanol, elbiler og affald (både hus-s-s-s-holdninger og husdyr) med det sigte at nholdninger og husdyr) med det sigte at nholdninger og husdyr) med det sigte at nholdninger og husdyr) med det sigte at nedbringe omkostninger ved produktion af VEedbringe omkostninger ved produktion af VEedbringe omkostninger ved produktion af VEedbringe omkostninger ved produktion af VE

� Samspil mellem teknologierSamspil mellem teknologierSamspil mellem teknologierSamspil mellem teknologier: bedre lagrings: bedre lagrings: bedre lagrings: bedre lagrings---- og styringssystemer for elforbrug og og styringssystemer for elforbrug og og styringssystemer for elforbrug og og styringssystemer for elforbrug og ––––produktion, f.eks. i form af smartproduktion, f.eks. i form af smartproduktion, f.eks. i form af smartproduktion, f.eks. i form af smart----grid og akkumulationsmuligheder i transportgrid og akkumulationsmuligheder i transportgrid og akkumulationsmuligheder i transportgrid og akkumulationsmuligheder i transport---- og varmog varmog varmog varme-e-e-e-sektorensektorensektorensektoren

� Støtte til Støtte til Støtte til Støtte til udvikling og udspredniudvikling og udspredniudvikling og udspredniudvikling og udspredning af nye projektmodeler for energibesparelserng af nye projektmodeler for energibesparelserng af nye projektmodeler for energibesparelserng af nye projektmodeler for energibesparelser � Forskningsmidler til materialeteknologiForskningsmidler til materialeteknologiForskningsmidler til materialeteknologiForskningsmidler til materialeteknologi (vigtigt for udvikling af vindmøller) og brændsel(vigtigt for udvikling af vindmøller) og brændsel(vigtigt for udvikling af vindmøller) og brændsel(vigtigt for udvikling af vindmøller) og brændsels-s-s-s-

celler (også vigtigt for lagringceller (også vigtigt for lagringceller (også vigtigt for lagringceller (også vigtigt for lagring af elektricitetaf elektricitetaf elektricitetaf elektricitet) samt 2 og 3 generations biobrændstofsteknol) samt 2 og 3 generations biobrændstofsteknol) samt 2 og 3 generations biobrændstofsteknol) samt 2 og 3 generations biobrændstofsteknolo-o-o-o-gier med det sigte at gier med det sigte at gier med det sigte at gier med det sigte at gøre projekter og prototyper mere markedsmodnegøre projekter og prototyper mere markedsmodnegøre projekter og prototyper mere markedsmodnegøre projekter og prototyper mere markedsmodne

� EkstraEkstraEkstraEkstra kapitalindskud til Vækstfondenkapitalindskud til Vækstfondenkapitalindskud til Vækstfondenkapitalindskud til Vækstfonden for at skaffe kapital til lovende Cleanfor at skaffe kapital til lovende Cleanfor at skaffe kapital til lovende Cleanfor at skaffe kapital til lovende Clean----Tech virTech virTech virTech virk-k-k-k-somheder på energiområdet.somheder på energiområdet.somheder på energiområdet.somheder på energiområdet.

1.2. DE INTERNATIONALE RAMMEVILKÅR Danmark er en lille økonomi med et relativt begrænset hjemmemarked. Derfor er der ekstra store perspek-tiver for Danmark ved at udnytte potentialet ved et fælles EU-hjemmemarked også i relation til grøn inno-vation. Der er tre elementer i denne del af strategien:

1.2.1 Begyndende harmonisering af VE-støtten i EU.

Europæiske og ikke mindst danske virksomheder har behov for et stort og stærk indre marked for nye VE-teknoliger som basis for konkurrence med hastigt voksende konkurrenter fra store økonomier som USA, Indien og Kina. Mens EU's kvotesystem har været en stor succes i relation til at få udviklet fælles omkostningseffektive rammevilkår for VE-teknologierne, så halter EU's VE politik bagefter. Problemet er, at EU's VE politik er baseret på 27 forskellige nationale støttesystemer, der er forskellige både i valg af støtteinstrumenter og niveauet af støtten til de forskellige teknologier, der i øvrigt justeres med jævne mellemrum. Det giver usikre investeringsvilkår, nationalt opdelte markeder og risiko for betydelige


7

forvridninger i konkurrencevilkår. Det skal ses i forhold til at investeringer i VE kapacitet i de kommende 10 år meget vel kan løbe op i et beløb svarende til det danske bruttonationalprodukt og vil være helt domi-nerende i energisektoren i forhold til investeringer i fossile værker. Investeringerne vil være særligt store indenfor områder med betydelige danske kompetencer og interesser som vindmøller og biomasse.. Imidlertid synes udviklingen af dette marked i Nordvesteuropa særligt for havvindmøller at være præget nok så meget af forskelle i støttesatser som af potentialet i de havområder, møllerne skal opstilles i. Det gør dette marked meget sårbart for ændringer i regulering og risikerer at skabe betydelige forvridninger i konkurrencen til ugunst for stærke danske virksomheder i denne branche. Den nationale tilgang i stedet for en indre markeds tilgang er også dyr: den kan komme til at koste årligt 60 til 100 milliarder kroner på EU plan i ekstraudgifter til at nå EU's VE-målsætninger; midler der i stedet kunne fokuseres på at løfte den langsigtede innovation i energisektoren. EU’s direktiv om støtte til VE, som har ”tilladt” den meget nationale tilgang, skal først til revision i 2014. Den politiske indsats fra dansk side må derfor de kommende år baseres på praktiske tiltag, der sikrer en me-re omkostningsorienteret tilgang og mindre forvridende vilkår for dansk afsætning af energiproduktion og udstyr. Anbefalinger for dansk indsats:Anbefalinger for dansk indsats:Anbefalinger for dansk indsats:Anbefalinger for dansk indsats:

� Overfor EUOverfor EUOverfor EUOverfor EU----Kommissionen og vigtige partnere ikke mindst i NordvesteuropaKommissionen og vigtige partnere ikke mindst i NordvesteuropaKommissionen og vigtige partnere ikke mindst i NordvesteuropaKommissionen og vigtige partnere ikke mindst i Nordvesteuropa betoning af betoning af betoning af betoning af behobehobehobehovet for en VEvet for en VEvet for en VEvet for en VE----indsatsindsatsindsatsindsats baseret på det indre marked og langsigtet fokus på innovation. baseret på det indre marked og langsigtet fokus på innovation. baseret på det indre marked og langsigtet fokus på innovation. baseret på det indre marked og langsigtet fokus på innovation.

� Arbejde for at der etableres et fælles EU regime for udvikling af støtte til havvindmøller evt Arbejde for at der etableres et fælles EU regime for udvikling af støtte til havvindmøller evt Arbejde for at der etableres et fælles EU regime for udvikling af støtte til havvindmøller evt Arbejde for at der etableres et fælles EU regime for udvikling af støtte til havvindmøller evt med start imed start imed start imed start i et samarbejde i Nordet samarbejde i Nordet samarbejde i Nordet samarbejde i Nord---- og Østersøen.og Østersøen.og Østersøen.og Østersøen.

� TilslutninTilslutninTilslutninTilslutning til det norskeg til det norskeg til det norskeg til det norske----svensk samarbejde om etablering af et fælles marked for at nå VE svensk samarbejde om etablering af et fælles marked for at nå VE svensk samarbejde om etablering af et fælles marked for at nå VE svensk samarbejde om etablering af et fælles marked for at nå VE målene evt. sammen med andre lande og dermed være frontløber i skabelse af det indre målene evt. sammen med andre lande og dermed være frontløber i skabelse af det indre målene evt. sammen med andre lande og dermed være frontløber i skabelse af det indre målene evt. sammen med andre lande og dermed være frontløber i skabelse af det indre marked. marked. marked. marked.

1.2.2 Opstramning af EU's klimapolitik og fokus på globale instrumenter

EU's kvoEU's kvoEU's kvoEU's kvotesystem tesystem tesystem tesystem har været en entydig succes ved at skabe ensartet beskatning på EU-niveau af udlednin-ger af CO2 udledninger fra elproduktion og stærk energiintensive industrivirksomheder. Det giver samti-dig tilskyndelse til udbygning og innovation af low carbon teknologier herunder ikke mindst vedvarende energi. Set fra et langsigtet investeringsperspektiv er der imidlertid to hovedsvagheder ved det nuværende set-up. For det første er kvoteprisen ganske lav set i forhold til de langsigtede krav om betydelige reduktion af driv-husgaser: Den økonomiske krise har således ført til en nedjustering af den forventede kvotepris i 2020 i for-hold til forventningerne fra vedtagelse af EU's klima- og energipakke i 2008: fra 225 kroner per tons til 120 kr. per ton. Det er langt under det niveau, der er krævet med EU's målsætninger om at reducere drivhusgas-ser med 80 til 95 procent i 2050. Det betyder at tilskyndelser til energibesparelser er svækkede og giver et øget behov for statslige midler til at realisere EU's målsætninger om 20 procents vedvarende energi. For det andet øges investorers afkastkrav af den store usikkerhed om den fremtidige kvotepris: i nogle studier op til 40 procent for meget langsigtede og kapitaltunge projekter.


8

AnbefalingAnbefalingAnbefalingAnbefalingerererer::::

� EU bør gå videre med overvEU bør gå videre med overvEU bør gå videre med overvEU bør gå videre med overvejelser om at gå efter en 30 procents reduktion i 2020, som i ejelser om at gå efter en 30 procents reduktion i 2020, som i ejelser om at gå efter en 30 procents reduktion i 2020, som i ejelser om at gå efter en 30 procents reduktion i 2020, som i EUEUEUEU----Kommissionens nye beregninger fra 2010 giver en kvotepris på 225 kroner svarende det Kommissionens nye beregninger fra 2010 giver en kvotepris på 225 kroner svarende det Kommissionens nye beregninger fra 2010 giver en kvotepris på 225 kroner svarende det Kommissionens nye beregninger fra 2010 giver en kvotepris på 225 kroner svarende det oprindelige skøn fra 2008 ved en 20 procents reduktionoprindelige skøn fra 2008 ved en 20 procents reduktionoprindelige skøn fra 2008 ved en 20 procents reduktionoprindelige skøn fra 2008 ved en 20 procents reduktion

� EU bør undersøge modeller hvor der ligges bund og lEU bør undersøge modeller hvor der ligges bund og lEU bør undersøge modeller hvor der ligges bund og lEU bør undersøge modeller hvor der ligges bund og loft over kvoteprisen for at give stabile oft over kvoteprisen for at give stabile oft over kvoteprisen for at give stabile oft over kvoteprisen for at give stabile vilkår for investeringer, der på energiområdet har en meget lang tidshorisontvilkår for investeringer, der på energiområdet har en meget lang tidshorisontvilkår for investeringer, der på energiområdet har en meget lang tidshorisontvilkår for investeringer, der på energiområdet har en meget lang tidshorisont

Udbredelse af klimapolitik til det globale planUdbredelse af klimapolitik til det globale planUdbredelse af klimapolitik til det globale planUdbredelse af klimapolitik til det globale plan er både nødvendigt for klimapolitikken og indebærer et betydeligt markedspotentiale for danske virksomheder. I dag er Kyoto-protokollen baseret på opdeling af landene i to grupper, nemlig lande med og uden konkrete forpligtigelser til at reducere udledninger. Denne politik er uholdbar: allerede i 2030 vil udledningerne fra Kina, Indien og andre lande uden forpligtigelser i dag overstige loftet for de samlede mængde af udledninger, der er foreneligt med 2 graders målsætning fra COP15. Der er (heldigvis) et stort potentiale for energieffektivisering i mange af disse lande, fordi man ofte historisk har ført socialpolitik via energipriserne, og derfor har været vant til meget lave energipriser. Kyoto-aftalen indeholder allerede mekanismer – JI og CDM – som både giver de rige og de fattige lande incita-menter til at deltage i teknologioverførselsprojekter. COP15 efterlod imidlertid stor usikkerhed om hvor-dan sådanne mekanismer kommer til at se ud efter 2012. Det bremser i betydelig grad interessen for sådan-ne projekter og svækker eksportpotentialet for danske virksomheder. Uafhængig af klimaforhandlinger kan der imidlertid forventes en selvstændig interesse i en række udviklingslande for energieffektivisering. Anbefaling:Anbefaling:Anbefaling:Anbefaling:

� EU skal prioritere afklaring af vilkår for brug af (reviderede) CDM/JI mekanismer i foEU skal prioritere afklaring af vilkår for brug af (reviderede) CDM/JI mekanismer i foEU skal prioritere afklaring af vilkår for brug af (reviderede) CDM/JI mekanismer i foEU skal prioritere afklaring af vilkår for brug af (reviderede) CDM/JI mekanismer i for-r-r-r-handlingerne om den globale klimapolitik, hvilket også vil styrke dansk eksportpotentiale handlingerne om den globale klimapolitik, hvilket også vil styrke dansk eksportpotentiale handlingerne om den globale klimapolitik, hvilket også vil styrke dansk eksportpotentiale handlingerne om den globale klimapolitik, hvilket også vil styrke dansk eksportpotentiale særligt for energibesparelser og de mest rentable VEsærligt for energibesparelser og de mest rentable VEsærligt for energibesparelser og de mest rentable VEsærligt for energibesparelser og de mest rentable VE----teknologierteknologierteknologierteknologier

1.2.3 EU’s 8 rammeprogram skal prioritere energiforskning

Meget af den innovation, der skal skabes i de kommende årtier kræver stor kritisk masse og vil kunne give gevinster langt ud over de lande, der gennemfører forskning. Det taler for en betydelig styrkelse af EU's fremtidige forskningsprogram for perioden efter 2012. Danske virksomheder og forskningsinstitutioner er dygtige til at udnytte det eksisterende forskningsprogram for EU midler – returprocent på Danmarks finan-sieringsandel er på 400 procent på energiområdet – hvilket giver et godt udgangspunkt for Danmarks nytte af nye midler. Danmark bør arbejde for at de prioriterede områder også tilgodeser danske styrkeposition jf. ovenfor og at de danske forskningsinstitutioner er organiseret så de effektivt sammen med private danske virksomheder kan indgå i internationale forskningsprojekter. Der kan arbejdes for at placere EU viden-centre på energiområdet, hvilket vil kunne styrke dansk energiinnovation. AAAAnbefalingnbefalingnbefalingnbefaling

� Løft af EU's F&U program foLøft af EU's F&U program foLøft af EU's F&U program foLøft af EU's F&U program for energi fra i dag(2010) på godt 2r energi fra i dag(2010) på godt 2r energi fra i dag(2010) på godt 2r energi fra i dag(2010) på godt 2 mia.kr. til mindst mia.kr. til mindst mia.kr. til mindst mia.kr. til mindst 5555 mia.kr. mia.kr. mia.kr. mia.kr. i 2020 herunder prioritering også mod danske styi 2020 herunder prioritering også mod danske styi 2020 herunder prioritering også mod danske styi 2020 herunder prioritering også mod danske styrkepositionerrkepositionerrkepositionerrkepositioner


9


10

Der er særligt i OECD landene en bred principiel enighed om behovet for omlægningen af energipolitikken drevet af to grundlæggende politiske prioriteter. Den første består i behovet for at imødegå en truende op-varmning af kloden drevet af stigende udledning af drivhusgasser. COP15 indebar tilslutning til en langsig-tet reduktion forenelig med en begrænsning af jordens temperatur med 2 grader. Den anden prioritet er ønsket om at undgå ”for” stor afhængighed af import af gas og olie fra potentielt ustabile områder i verden. Det skal ses i sammenhæng med officielle fremskrivninger fra f.eks. IEA, hvor oliereserverne i stigende grad koncentreres i mellemøsten i takt med at oliedepoter i OECD lande forventes at blive tømt. Både før, under og også efter COP15 er der betydelig uenighed om dels byrdefordeling af de omkostninger, der følger af dette krav, dels den mere konkrete praktiske indretning af de instrumenter, der skal levere den langsigtede reduktion af drivhusgasserne. Tilgangen i dette studie er, at der i de kommende år materialiserer sig en praktisk politisk vilje til at leve op til løfterne, også globalt set, og at de europæiske lande er i front. En faktisk levering på løfterne kræver meget betydelige reduktion i udledninger af den største drivhusgas nemlig CO2. På den korte bane har EU forpligtiget sig til at reducere drivhusgasserne med 20 procent i 2020 i forhold til 1990, medens Danmark som led i den interne fordeling af kravene har forpligtiget sig til den samme reduktion. Globalt set vil selv betydelige nye tiltag føre til yderligere udledninger i de kommen-de år, som følge af stærkt stigende energiforbrug i lande som Kina, Indien mv: faktisk svarer en begrænsning af stigningen i udledningen i 2020 til 17 procent som i tabel 2.1 til meget betydelig reduktion i forhold til ”at-lade-stå-til”. På lang sigt skal udledningerne imidlertid begrænses på globalt plan med omtrent 50 pro-cent i forhold til 1990 og langt mere i rige regioner som EU, herunder Danmark.

Tabel 2.1 Målsætninger og krav i Danmark, EU og Globalt, 2020 og 2050

2020 2050

DK EU Globalt DK EU Globalt

GHG- reduktion i procent -20 -20 17 -80 -80 -50

VE mål, procent af energi-forbrug

30 20 - - - -

Energi-

22 20 - - - -

Kilder i appendiks

Vejen frem til at opnå de to mål – reduktion af drivhusgasser og bevaring af en strategisk forsyningssikker-hed – er per definition øget energieffektivitet og anvendelse af low carbon teknologier brændstoffer, ikke mindst vedvarende energi. EU har i den sammenhæng forpligtiget sig selv til at øge andelen af vedvarende energi til 20 procent af det samlede energiforbrug, hvor den nationale andel for Danmark er 30 pct.. Der er tilsvarende (mindre) bindende målsætninger for at reducere energiforbruget for EU, medens der i dansk sammenhæng, som led i energiaftaler, er fastsat nationale mål. I tilgift til klimapolitik er der to nok så stærke underliggende faktorer, der kommer til at dominere udvik-lingen. For det første er de reale energipriser, efter mange års fald eller i hvert fald stagnation, steget betyde-

Kapitel 2 DE GLOBALE RAMMER


11

ligt siden 2003-2004 og forventes at holde sig på et højt niveau i de kommende årtier jf. figur 2.1. Det skyldes først og fremmest betydelig økonomisk vækst udenfor OECD området, hvor energiforbruget i sti-gende grad sker.

Figur 2.1 Reale energipriser, 1974-2020

Kilde: IEA data services og IEA World Energy Outlook fremskrivninger

Den anden faktor er en bevægelse mod afvikling af store energisubsidier. Særligt i mellemindkomstlande i Asien samt Kina og Indien har der været en tradition for at forbrugere og industri betalte energipriser som lå betydeligt under de reelle produktionsomkostninger. Der har i de senere år både i Asean-lande, Kina og Indien startet en gradvis, begyndende reduktion af disse subsidier, som ud over at belaste miljøet, er en gan-ske betydelig omkostning for disse lande. OECD studier viser at en total afvikling af disse subsidier på glo-balt plan vil kunne reducere udledningen af CO2.betydeligt. Disse faktorer vil således sætte sit præg på verdenens energiforbrug og politik i de kommende årtier. Frem-skrivningerne fra det Internationale Energi Agentur som er forenelige med de langsigtede målsætninger fra COP15 tegner et billede af en global energisektor under ganske betydelig forandring. Indenfor de kommende ti år dvs. frem til 2020, forventes energibesparelser globalt set at levere hovedparten af CO2-reduktioner jf. figur 2.2. Efter 2020 kommer en stigende del fra vedvarende energi samt atomkraft og Carbon Capture and Storage(CCS) baserede kul- og gasværker. Mindre brug af atomkraft og CCS stiller højere krav til VE og energibesparelser. Der er klare regionale forskelle i realisering af disse mål på flere planer. På globalt plan handler det om en reduktion i tilvækst i energiforbrug og emissioner medens det i meget rige regioner som EU handler om at accelerere besparelser. På globalt plan står energieffektivitet for en noget større andel end i EU, hvor energi-effektiviteten ofte er høj og hvor VE implementering fylder en større rolle, jf. figur 2.2.

0

100

200

300

400

500

600

1975

1980

1985

1990

1995

2000

2005

2010

2015

2020

Index, 1995 = 100

Oliepris Kulpris Gaspris


12

Figur 2.2 Bidrag til CO2 reduktioner, Globalt og EU Globalt EU

Kilde: IEA (2009) World Energy Outlook

Udviklingen af det danske erhvervspotentiale for at deltage i en global grøn innovations proces på energi-området må tage afsæt i denne diagnostik på flere planer. Danske erfaringer med en effektiv og velreguleret energisektor kan bruges som udgangspunkt for salg af energieffektive løsninger af kul og gas baserede løs-ninger. Den danske energisektor omfatter en bred palet af ydelser, spændende fra energiproduktion (kul-kraftværker, vindmøller biomasse, effektive kraftvarme løsninger), rådgivning om implementering af energi-teknologiske løsninger og fremstilling af en række produkter (vindmøller, termostater mv) med en samlet beskæftigelse på 75.000 og en årlig værditilvækst på omtrent 35 mia.kr jf. figur 2.3. Indenfor fremstilling af energiudstyr fylder pumper og kompressorer samt elektroniske komponenter og plader samlet mere end vindmølle orienteret beskæftigelse, hvilket viser betydningen af en bred tilgang til det erhvervspolitiske po-tentiale. Blandt de 28 sværvægtere indenfor dansk energiteknologier fylder producenter med speciale i ener-gieffektivisering ligeså meget som producenter af VE-udstyr1.

Figur 2.3 Beskæftigelse, værditilvækst og eksportandele i energibranchen, 2006-2007

Kilde: Copenhagen Economics på baggrund af Brøndum (2009) og Eurostat, Comext / Procdom data

1 Brøndum et al (2009), figur 21

26

28

30

32

34

36

38

40

42

2007 2020 2030

Gt CO2

Effektivitet

Vedvarende energi

Nukleare

CCS

2

2,2

2,4

2,6

2,8

3

3,2

3,4

3,6

3,8

4

2007 2020 2030

Gt CO2

0 5 10 15 20

0 10000 20000 30000 40000

Tilbehør til ledninger og kabler

Elektroniske komponenter og plader

Vindmøller og dele hertil

Pumper og kompressorer

Andre brancher i

energiproduktionen

Energiproduktion og - distribution

Værdtilvækst,

mia. Dkr.

Beskæftigede

Value added og beskæftigede i energibrancher, 2007

Beskæftigede Værditilvækst

Kilde: Eurostat

0 50 100

Tilbehør til ledninger og kabler

Elektroniske komponenter og

plader

Vindmøller og dele hertil

Pumper og kompressorer

Andre brancher i

energiproduktionen

Energiproduktion og - distribution

Eksportandel for energibrancher, 2006


13

Virksomhederne er meget internationalt orienterede og er ved at omstille sig til at klimateknologi ikke bare er dansk/nordisk/europæisk. Eksportandele er typisk over 60-70 procent inden for energiudstyr (fremstil-ling) jf. figur 2.3 og for de større virksomheder (over 500 ansatte) er dette tal for over halvdelen af virksom-heder større end 75 procent2. Hidtil har nærmarkederne i Norden og det øvrige Vesteuropa været dominen-rende, men virksomhederne har klare forventninger om at den fremtidige vækst i højere grad kommer uden for Europa jf figur 2.4. Samlet set forventes markederne uden for Europa fremadrettet at blive de vigtigste markeder for satsninger. Det betyder også at markedsvilkår udenfor EU og konkurrencen fra lande uden for EU bliver vigtigere. Det er ikke noget tilfælde, at Kina i dag er blevet en meget stor spiller både på sol-celler og i stigende grad på produktion af turbiner til vindmøller. Det skyldes massive satsninger på dette område i offentlig og privat regi de allerseneste år.

Figur 2.4 Betydningen af forskellige eksportmarkeder: i dag og forventning til udvikling

Kilde: Brøndum & Fliess(2009)

Solceller er mere generelt et godt eksempel på hvor hurtigt markedspositioner kan ændres sig globalt og hvor vanskeligt det kan være at bygge varige konkurrencefordele gennem høje tilskud til udrulning i et lo-kalt marked. Tyskland har brugt 10 år på gennem meget høje tilskud til tilslutning af solceller at udvikle en solcelleindustri. Konsekvensen er at Tyskland i dag står for knap halvdelen af den installerede globale kapa-citet. Modsat er deres markedsandel i produktionen faldet til under 10 procent globalt set og falder fortsat med Taiwan og Kina stående for over halvdelen af produktionen ikke mindst fordi de er i stand til at sælge til lavere priser i de vigtigste nye markeder, både i USA og Kina selv. Den samlede produktion af solceller i Tyskland målt i kroner og øre har faktisk været faldende på trods af et stadig stigende total marked samtidig med at tysk eksport også viger jf. figur 2.5 panel

2 Brøndum et al (2009) side 43.

0 50 100 150 200 250 300

Norden

Vesteuropa, eksl. Norden

USA

Østeuropa, ekskl. Rusland

Øvrige Asien

Kina

Canada

Rusland

Japan

Øvrige

Antal virksomheder

Forventninger til fremtiden I dag


14

Figur 2.5 MarkedsIndikatorer for tysk solcelle industri

Note: For solceller er vist Tysk solcelle eksport, import samt produktionsværdi. Vindmøller er før 2007 ikke præcist defineret i pro-duktionsstatistik på europæisk niveau i værdi. For vindmøller er derfor vist den årligt installerede kapacitet i Europa og resten af verden mod Europas eksport samt import i værdi.

Kilde: Eurostat, Comext, Prodcom samt EWEA

Sat på spidsen kan man sige at tyske elforbrugere gennem højere elpriser har betalt for opbygningen af en relativt effektiv kinisisk solcelleindustri, som leverer produkter til lavere priser end tyskerne selv kan produ-cere til. Den kraftige udbygning af kinesisk produktion er således stort set skabt via eksport ikke mindst til Tyskland, Spanien og USA med en meget begrænset produktion til hjemmemarkedet indtil nu. De miljømæssige gevinster hidtil har samtidig været begrænsede i forhold til omkostningerne. Samlet set har tyske elbrugere betalt skønsmæssigt knap 400 mia.kr. over de sidste 10 år for en forholdsvis begrænset fortrængning af CO2. Skønsmæssigt betaler tyske elbrugere omtrent 5000 kr. for hver tons CO2 der for-trænges, ca 30-40 gange kvoteprisen i ETS3. Den forholdsvis simple lære i denne historie er, at omkostningseffektivitet i miljøpolitikken og innovations- og erhvervspolitik følges ad. Som det diskuteres i næste kapitel er det afgørende at offentlige midler til fremme af nye teknologier afpasses til deres modenhed. Tyskland har haft meget ringe afkast af at forcere udrulning af en meget umoden, dvs. meget dyr teknologi, på det tyske hjemmemarked: de havde stået bed-re hvis bare en beskeden andel af de 400 mia.kr. havde været brugt på målrettet støtte til innovation, her-under af solcelleteknologi.

3 RWI(2009)

0

1

2

3

4

5

6

7

8

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

Mia. euro

Tyskland, produktion Tyskland, import Tyskland, export


15

3.1. STYRKELSE AF MIDLER TIL FORSKNING, UDVIKLING OG INNOVATION Carbon pricing har betydeligt potentiale for at drive udrulning og udvikling af low carbon teknologier i de kommende år. ETS systemet kan således drive en betydelig del af deployment af eksisterende VE teknologi-er i de kommende år, det gælder særligt ved et krav om 30 procents reduktion som argumenteret ovenfor. OECD studier tilsiger4 at carbon pricing foreneligt med globale klimaambitioner vil gøre en 3-4 dobling af FU midler rentable, herunder for private virksomheder. Faktiske og lovede energiskatter kan imidlertid ikke alene levere det kvantespring i den teknologi udvikling, der er nødvendig for at nå de globale ambitioner af tre årsager. For det første betyder den langsigtede og meget usikre natur, der er forbundet med megen energirelevant forskning, at private investorer ikke vil in-vestere i en række forskningsprojekter på trods af et betydelig samfundsøkonomisk potentiale. For det andet vil der ofte være en række samfundsmæssige gevinster særligt ved grundforskning, som ikke høstes af inve-stor, men tilfalder samfundet mere generelt (”spill-over effekter”). For det tredje vil private investorer også være usikre på om regeringerne globalt set er villige til at løfte beskatningen af CO2 til det niveau, der skal til: det forudsætter at lande som USA, Canada, Austrilien, Kina, Indien bevæger sig fra en meget begrænset beskatning, for ikke sige tilskud til, forbrug af fossile brændstoffer til et beskatningsniveau, der ligger måske 10 gange højere end dagens pris på kvoter i ETS systemet. Internationale studier5 peger derfor også på behovet for en meget betydelig styrkelse af offentlige midler til energiforskning i det kommende årti. Olieprisstigningerne i 1970erne igangsatte en ganske betydelig stig-ning i de offentlige midler både som andel af BNP og som andel af de offentlige forskningsmidler medens de stagnerende/faldende reale oliepriser særligt fra 1980-1985 indledte en lang periode frem til 2005-6 med et betydeligt lavere niveau jf figur 3.1 De senere år har ført til opjusteringer igen af offentlige FU-midler – drevet af højere energipriser og klima- og energipolitiske målsætninger – men stadig betydeligt under ni-veauet fra slutningen af 1970erne. Et studie fra OECD indikerer imidlertid at behovet for egentlige kvante-spring i teknologier snarere end marginale forbedringer kræver en meget betydelig forøgelse af offentlige forskningsinvesteringer frem mod 2020, svarende til 0,12 procent af globalt BNP (stiplede linie i figuren ) mod i dag 0,04 procent. Given den forventede BNP vækst i de kommende 10 år så svarer det til 3-4 dob-ling af de offentlige FU-midler i faste priser.

4 OECD(2009). 5 OECD(2009a,b,c)

Kapitel 3 INNOVATIONSPOLITIKKEN I DANMARK: TRE FOKUSOMRÅDER


16

Figur 3.1 Offentlige RD investeringer i energi: andel af BNP, 1974-2020

Kilde: IEA data services, Bosetti et al (2009), EIA hjemmeside og egner beregninger

Der skal derfor opereres med en bred palet af instrumenter for at understøtte virksomhedernes innovation. Det afgørende er at støtten skal strikkes sammen sådan at den understøtter udvikling af nye produkter helt fra idestadiet til afsætning i et almindeligt kommercielt marked jf. figur 3.2:

� Grundforskning finansierer udvikling af nye ideer og koncepter. I mange tilfælde f.eks. støtte til udvikling af nye materialeteknologier vil der ofte være mange anvendelsesmuligheder af ny viden på området både indenfor og udenfor energiområdet

� Hernæst kommer en fase, hvor ideer flyttes fra forskningsverden ud i udviklingsafdelinger, hvor der er behov for de første testforsøg. Offentlig støtte knyttes til opbygning af kapacitet på området og målrettede testning af produkter. Anden generations biobrændstoffer er et eksempel

� Dernæst kommer en fase, hvor produktet skal testes i den virkelige verden, det kan f.eks. være nye generationer af mere teknisk avancerede havvindmøller: Støtten knyttes til udvikling af teknologi-versioner, der har et realistisk potentiale for at blive omkostningseffektive til de priser på fortræng-ning af CO2, som gælder om 5-10-20 år. Støtten kan afgrænses til demonstrations- og skalaforsøg og bør ikke udstrækkes til en generelt støtte per produceret enhed: 1000 anlæg giver ikke nødven-digvis megen mere ”læring” end 10 i tidlige stadier6.

� De tyske erfaringer med solceller viser farerne både erhvervspolitisk og miljøpolitisk med massiv driftsstøtte til en meget umoden og dyr solceller teknologi jf. kapitel 2.7 I vores tilgang vil støtte til

6 En meget betydelig af de faktiske teknologiske fremskridt skyldes forsknings- og udviklingsaktiviteter nok så meget som historisk produktion. En række studier advarer derfor mod en for ukritisk bruge af ”læringsgevinster” som argumenter for at udrulle umodne teknologier (oversigt f.eks. i Bosetti et al (2009). 7 RWI(2009)

0.00%

0.10%

0.20%

0.30%

0.40%

1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 2018

OECD, RD&D i energi, andel af BNP, procent

RD&D i energi, procent af BNP, historisk

RD&D i energi, procent af BNP til opfyldselse af 2 degree scenarie


17

solceller (PV) således skulle begrænses inden for en samlet pulje og med fokus på udvikling jf. fi-gur 3.2.

� Endeligt kan beskatning af CO2 (egentlige skatter eller EU's-kvotesystem) suppleret eventuelt af teknologineutrale støttesystemer bidrage til udrulning af VE-teknologier som er (meget) mar-kedsmodne, dvs. (næsten) konkurrencedygtige med gældende CO2-priser under forudsætning af et forholdsvis moderat pristillæg. For denne type af støttesystemer er det særligt væsentligt jf. den indledende diskussion at det foregår på internationalt harmoniserede rammevilkår.

Figur 3.2 Fra forskningsideer til marked i udviklingen af nye energiteknologier

Kilde: Baseret på en række publikationer fra International Energy Agency (se litteraturliste)

Med udgangspunkt i denne model og en evaluering af den nuværende indsats vurderes det at den danske innovationsindsats overfor energisektoren kan udvikles med to fokusområder:

• Skubning af den fremtidige indsats mod øget støtte til egentlig innovation relativ til driftsstøtte • fokus af støtte på områder med danske styrkepositioner i internationalt perspektiv

3.2. ØGET FOKUS PÅ INNOVATION OG FORØGELSE AF OFFENTLIGE FU-MIDLER Den nuværende danske støttesats til fremme af nye energiteknologier på godt 4 mia. kr. er overvejende af driftsstøtte karakter og meget nationalt orienteret. Støtteindsatsen er renset for støtteværdi af CO2-afgifter og EU's- kvotesystem, jf. figur 3.3. Således går omtrent 80 procent til produktionsstøtte i form af nedslag i energiafgifter samt direkte støtte til produktion af biomasse, biogasser, vindmøller mv. Den resterende del bruges på grundforskning (10 procent), mere anvendt støtte til forskning og udvikling (5 procent) samt markedsmodning. Den internationale indsats er først og fremmest EU's forskningsprogrammer, som er voksende, men fortsat ubetydelig samlet set.

Grundforskning � FoU � Demonstration �Markedsmodning � Kommercialisering

Tid

Markedsudvikling

Prototype og demonstrationsfase teknologier (Fx 2. generations biobrændstof)

Kontinuitet, Forskning og Kontinuitet, Forskning og Kontinuitet, Forskning og Kontinuitet, Forskning og udvikling, skabe attraktivt udvikling, skabe attraktivt udvikling, skabe attraktivt udvikling, skabe attraktivt

markedmarkedmarkedmarkedKapital omkostnings

incitamenter; investeringsskat kredit; rabatter, lånegarantier etc.

Høj omkostnings-gap teknologier

(Fx PV)

Lav omkostnings-gap teknologier

(Fx Landvindmøller)

Færdigudviklede teknologier(Fx hydro)

Teknologineutral konkurrenceTeknologineutral konkurrenceTeknologineutral konkurrenceTeknologineutral konkurrenceTGC; Handel med kulstof (EU ETS)

Stabilitet, lavStabilitet, lavStabilitet, lavStabilitet, lav----risikoincitamenterrisikoincitamenterrisikoincitamenterrisikoincitamenterPrisbaseret: FIT, FTP

Mængdebaseret: tilbud

Påtvunget markedsrisiko, Påtvunget markedsrisiko, Påtvunget markedsrisiko, Påtvunget markedsrisiko, garanteret men aftagende garanteret men aftagende garanteret men aftagende garanteret men aftagende

udbytteudbytteudbytteudbyttePrisbaseret: FIP

Mængdebaseret: TGC med teknologi sammenslutning

Stimulere market pullStimulere market pullStimulere market pullStimulere market pullFrivillig (grøn) efterspørgsel


18

Figur 3.3 Fordeling af energipolitiske midler på to dimensioner:

(1) fra ide til marked og (2) national og international

I alt 4300 mio. kr. I alt 80 mio. kr. I alt 34 mio. kr.

Kilde: Copenhagen Economics baseret på række kilder (Forskningsministeriet, Energiministeriet mv.)

Den nødvendige stramning over de kommende årtier af klimapolitikken med markant højere beskatning af CO2 og andre drivhusgasser vil nærmest per automatik føre til mindre behov for specifik produktionsstøtte: f.eks. vil højere priser på CO2 kvoter føre til tilsvarende mindre behov for at støtte biomasse- og vindmøl-leproduktion jf. diskussion i kapitel 2. I fravær at sådanne stramninger vil det være nødvendigt med betyde-lige driftsstøtte for at nå de EU fastlagte mål for VE andele i den samlede energiproduktion: måske i størrel-sesordenen knap 600 mia.kr. på EU niveau i 20208. Tilsvarende må offentlige midler til udvikling vokse: siden 2004 har midlerne som andel af BNP været voksende i Danmark, men fra et lavt niveau sammenlig-net med andre lande og fortsat langt under det krævede langsigtede niveau jf. OECD analyserne fra kapitel 2 og figur 3.4.

Figur 3.4 Dansk indsats voksende, først nu på niveau med sammenlignelige lande

Note: Den højre figur stopper i 2007 skyldes manglende data Kilde: IEA Data Services, BNP fra Verdensbanken

Der er i dag en betydeligt række puljer og aktører involveret i tildelingen af danske midler til innovation på energiområdet gennem den innovationspolitiske fødekæde. Det Strategiske Forskningsråd (DSF) fokuserer primært på anvendt forskning og udvikling jf. figur 3.5. Midler fra energinet.dk (ForskEL og ForskVE),

8 Euroelectric (2008)

0

0,01

0,02

0,03

0,04

0,05

0,06

0,07

0,08

2004 2005 2006 2007 2008

Andel af BNP, procent

Japan

USA

Verden

EU

0

0,01

0,02

0,03

0,04

0,05

0,06

0,07

0,08

2004 2005 2006 2007

Andel af BNP, procent

Danmark

Frankrig

Tyskland

Norge

Sverige

EU


19

Klima- og Energiministeriet (EUDP), Dansk Energi(Elforsk) og Højteknologifonden bruges på anvendt forskning, udvikling og demonstration. Fra 2010 fået Green Lab og Fornyelsesfonden, der fokuserer på demonstration og markedsinformation. Endelig er der Vækstfonden der med et offentligt baseret finansie-ring leverer forskellige former for kapitalindskud til private virksomheder for at de kan få finansiering til at få rulle produkterne ud i markedet. Sådan finansiering kan finansiere egentlige produktionsudstyr, mar-kedsføringsudgifter mv. Vækstfonden er aktiv på alle innovationsfronter men fokuserer også en del på ener-giområdet. De største midler uddeles af DSF og EUDP (de tykkeste pile i figur 3.5)

Figur 3.5 Instrumenter og puljer fordelt på skalaen: fra ide, demonstration til markedsudvikling (tyk-

kelse af pil afspejler størrelse af program)

Kilde: Energi 2010, årsrapport om de danske energiforskningsprogrammer

Givet behovet for at tilføre betydeligt øgede midler til energiforskningen i de kommende år bør strukturen i det offentlige bevillingssystem og støttemodellerne ses efter. Der bør særligt ses på behovet for at udvikle nye modeller til at støtte lovende, men endnu ikke markedsmodne teknologier, hvor opgaven er få nedbragt omkostninger ved produktion af energi før en egentlig kommercialisering. Et OECD studie har på basis af erfaringer fra blandt forskningspolitikken i US foreslået modeller, hvor man gør tildeling af midler til udvikling mere målorienteret og mindre fokuseret på individuelle støtteord-ninger til forskellige teknologier9. Det vil sige at man definere nogle konkrete krav til forventninger til ud-vikling på mere generelt plan f.eks. at man kan producere el eller anden energi til forventet niveau under et kritisk niveau indenfor en bestemt årrække. Alle typer af teknologier – energispareprojekter såvel som nye VE-projekter – kan byde ind. Det offentlige kan vælge at gå med som investor for at dele både gevinster og tab.

9 OECD(2009)

DSF

ForskEL

ForskVE

Elforsk

EUDP

Højteknologifonden

EU: FP7-Energy

Green lab

Fornyelsesfonden

Vækstfonden


20

Et kerneområde for styrkelse vil være den del af innovationsfødekæden, der ligger mellem grundforskning og udrulning af markedsmodne teknologier. Det vil sige i fasen hvor produkter skal testes af og afprøves i stadig større skala. Det kræver dels en styrkelse af kapacitet på området f.eks. som oprettelse af testcentre for vindmøller. Det offentlige leverer finansiering til den grundlæggende forskningsmæssige infrastruktur me-dens de private virksomheder (brugerne) i betydelig grad må betale for anvendelsen. Det svarer f.eks. til principperne, der anvendes for GTS institutionerne samt de nye Green Labs. Den konkrete grad af bru-gerfinansiering må afhænge af, hvor tæt på markedet produktet er og i hvilket omfang forskningsresultater mv. er offentlige tilgængelige for andre aktører. Jo større andel af gevinsten ved en aktivitet, der tilfalder den private virksomhed, jo større privat finansiering må der leveres. Der kan også være behov for styrkelse af venturekapital markedet. Det centrale er her projekterne som ud-gangspunkt skal have et sandsynligt privatøkonomisk potentiale: problemet er at finde investorer som tror på potentialet. En række undersøgelser viser, at en velfungerende infrastruktur for test og demonstrations-forsøg er afgørende: det giver finansielle investorer muligheder for at vurdere det tekniske potentiale i nye produkter og processer10. Tilbageslaget i den internationale økonomi og svækkelsen af den finansielle sek-tor11 samt usikre resultater fra COP15 har imidlertid gjort det sværere at skaffe kapital. En mulighed ville være at styrke Vækstfondens pulje for investeringer virksomheder med innovationspotentiale for energitek-nologier. Pointen ved at lægge indsatsen netop i Vækstfonden er, at det offentlige får del i både gevinster og tab idet kapitaltilførsel ofte foregår via delejerskab.

3.3. FOKUS PÅ DANSKE STYRKEPOSITIONER I INTERNATIONALT PERSPEKTIV En dansk analyse af de fremtidige behov for nye energiteknologier indeholder en bred palet af mulige ind-satsfelter jf. tabel 3.1. På produktionssiden peges der, ud over en række kendte VE-teknologier, også på kraftvarmeproduktion (medens Carbon Capture and Storage teknologier måske lidt overraskede er fravæ-rende). Nok så vigtigt udpeges energidistribution og energiforbrug som samtidige vigtige indsatsområder. For energidistribution peges ikke mindst på behovet for teknologier, der effektivt kan integrere mere volatil vindenergi i elsystemerne (infrastruktur til elbiler, biologiske batterier mv.). Hertil kommer effektiviserings-indsats ikke mindst overfor opvarmning af bygninger.

10 Vækstfonden 11 Brøndum et al (2009) peger på at manglen på ekstern finansiering er en hovedbarriere og knytter det delvist sammen til den finan-sielle krise.


21

Tabel 3.1: Centrale fremtidige teknologifelter med betydeligt potentiale Energikilder & produktion Energidistribution Energiforbrug

Kraftvarme Effektiv infrastruktur til eldrevne biler Integreret intelligent energistyring i bygninger

Vind Biologiske batterier Industrialiseret lavenergibyggeri

Ocean energi Integration af alternativ energi Energieffektivisering i bygninger

Vandkraft Brugeradfærdspåvirkning, bevidstgø-relse af energiforbrug og miljøpåvirk-ning

Brint Teknologi til fremtidens forsyningssy-stemer

Energieffektivisering i industri og ser-vicesektoren

Pålidelig Elforsyning Metrologi – Energimåling dvs. intelli-gente målesystemer til effektiv udnyt-telse af energi

Biomasse energi Håndteringen af biomasse (går fra dyrkningssystemer, teknologier til for-arbejdning dvs. konvertering fra bio-masse til energimedie)

Nye lavenergiapparater og systemer (fx til belysning)

Solenergi Vindmølleteknologi i en international kontekst

Kilde: GTS 2015 – ”Katalog over teknologiske indsatsområder”, side 8-9 og Energi 2010(2010)

Det danske innovationsmiljø kan ikke være ledende på samtlige fronter i udviklingen af nye teknologier. In-ternationale studier udviklet ikke mindst i OECD- og EU-regi peger på behovet for en regional specialise-ring baseret på de styrkepositioner lande, regioner, forskningsinstitutioner og FU-intensive virksomheder har udviklet over tid. Strategien bygger ikke på en ”pick-the-winners” tilgang dvs. definering af nogle virk-somheder eller teknologier som globalt set er ”de bedste”. Det er mere en snusfornuftig erkendelse af, at der er en betydelig akkumuleret høj kompetence indenfor nogle rimeligt afgrænsede felter af forskning og ud-vikling, som skal prioriteres i tildelingen af offentlige støttemidler. Der er også en vis koncensus om, hvor sådanne styrkepunkter befinder sig i Danmark. På VE området peges der særligt på vindkraft, bioteknologi, elektrisk belysning, motorer og pumper og ud-nyttelse af affald og materialeteknologi 12. Det sidste område er ikke mindst interessant i sammenhæng med udvikling af (hav)vindmøller, hvor behovet for reduktioner i produktionsomkostninger blandt andet er knyttet til udvikling af mere robuste og mindre dyre konstruktioner af fundamenter, vinger mv. Indpasning af vind i det danske og nordvesteuropæiske elsystem vil være en særlig udfordring jf. afsnittet om behovet for udvikling af infrastruktur i EU. Men der er også et innovationspotentiale på dette område. Det knytters sig til udvikling af elsystemer og lagring og brændselsceller jf. også tabel 3.1. Her er et område hvor Danmark både vil have særlig fordele af teknologiske fremskridt på grund af det store vindpotentiale og hvor danske innovationspotentiale ligger på et højt niveau. Tilsvarende må der ske en styrkelse af indsatsen for energibesparelser. Indsatsen må koncentreres om to om-råder. Det første er en egentlig styrkelse af forskningsmidler til området, der er blandt peget på øgede mid-

12 Se f.eks. Ministeriet for Videnskab, Teknologi og Udvikling (2009), side 55-56,


22

ler til udvikling af afprøvning af nye metoder og materiale for at nedbringe energi til opvarmning, der står for 40 procent af det danske energiforbrug13. Det andet område er effektiv implementering af allerede rentable energieffektive løsninger, dvs. løsninger hvor (det forventede) prisniveau er tilstrækkeligt højt til at retfærdiggøre omlægning af produktion og køb af nyt udstyr. Snarere end tilskud til rentable investeringer skal der udvikles nye modeller, hvor risici ved og finansiering af sådan projekter splittes op på en ny og intelligent metode blandt de forskellige aktører. De såkaldte ESCO-modeller med nye måde at fordele ansvar og risiko har vundet udbredelse i en række lande blandt USA, England og Sverige. De allerseneste år har sådanne modeller også fået en vis bevågenhed i Danmark med nogle af de første gennemførte forsøg. Kernen i sådanne projekter er følgende elementer:

� Manglende likviditet:Manglende likviditet:Manglende likviditet:Manglende likviditet: hvis husholdningen/virksomheden ikke har likviditet må banker mv. leve-re midler: det vil særlig gøre hvis der kan ”garanteres” en besparelse jf. nedenfor

� Usikkerhed om energipriserUsikkerhed om energipriserUsikkerhed om energipriserUsikkerhed om energipriser: hvis investeringer ikke foretages fordi investor ”frygter” at de frem-tidige energipriser falder i forhold til prospektet kan sådan risiki afdækkes af enten finansielle in-stitutioner og energiselskabernes distributionsselskaber

� Usikkerhed om reelle energibesparelserUsikkerhed om reelle energibesparelserUsikkerhed om reelle energibesparelserUsikkerhed om reelle energibesparelser: Hvis der er usikkerhed om investeringen ikke leverer den lovede reduktion i energiforbrug må leverandører påtage sig denne risiko. Det kan være en-trepenører, distributionsselskaber mv.

Støtte bør derfor gives til udvikling og aftestning af sådanne modeller gerne i en skala og med en sådan kva-litet i dokumentation at det kan virke som en driver for energieffektivisering i den private og offentlige sek-tor.

13 Energi2010(2010), side 16.


23

Danske virksomheder med tilknytning til energisektoren vil entydigt have gavn af at denne omstilling til en mere grøn økonomi tager udgangspunkt i regulerede og ensartede globale rammevilkår.regulerede og ensartede globale rammevilkår.regulerede og ensartede globale rammevilkår.regulerede og ensartede globale rammevilkår. Danmark har konsekvent over årtier været en af de mest stabile fortalere i internationale fora for frihandel og ensartede konkurrencevilkår. Det gælder såvel ift. EU’s indre marked som liberalisering i WTO sammenhæng. Det afspejler den simple virkelighed skitseret ovenfor med eksport som kernen for danske energiorienterede virksomheder. Denne rapport opstiller fire prioritetsområder for internationale rammevilkårfire prioritetsområder for internationale rammevilkårfire prioritetsområder for internationale rammevilkårfire prioritetsområder for internationale rammevilkår

� Videreudvikling af EU's kvotesystem og udbredning af klimapolitik til globalt plan � Begyndende harmonisering af EU's støtteordninger til vedvarende energi � Øgede rammer til EU's budget til energirelevant forskning � Øget liberalisering af elmarkeder og investeringer i grids

4.1. EU'S KVOTESYSTEM OG GLOBAL KLIMAPOLITIK EU’s fælles kvotesystem (ETS) er et kerneeksempel på vellykket international rammeregulering. EU’s fælles kvotesystem (ETS) er et kerneeksempel på vellykket international rammeregulering. EU’s fælles kvotesystem (ETS) er et kerneeksempel på vellykket international rammeregulering. EU’s fælles kvotesystem (ETS) er et kerneeksempel på vellykket international rammeregulering. Efter reformerne vedtaget i de sidste par år vil det ETS der træder i kraft i 2013 indebære/muliggør en substantiel beskatning af CO2 fra el, fjernvarme og de mest energiintensive industrielle processer på EU plan. Det giver dels tilskyndelser til energibesparelser og vil samtidig kune gøre betydelige mængder af vedvarende energi-kilder direkte rentable. Indførslen af ETS betyder også at kun beslutninger på EU niveau kan føre til reduktion af udledning af drivhusgasser fra kvotesektoren. Blev der f.eks. indført en beskatning af fossile brændstoffer for kul og gas til dansk elproduktion ville det alene føre til en reduktion i drivhusgasser direkte fra dansk territorium efter-fulgt en præcis øget udledning fra resten af EU samt en lavere kvotepris, der svækker tilskyndelser til ener-gibesparelser i resten af EU. Det skyldes at det samlede udledninger alene styres af det samlede antal kvoter: Nationale tiltag kan alene påvirke den fælles kvotepris i EU. Givet den højere danske energieffektivitet i ren el- og kraftvarmeproduktion vil ensidige danske stramninger samtidig fører til reduceret energieffektivitet i elproduktionen i EU14. I forhold til investeringsklimaet for udrulning og udvikling af energisparende udstyr og vedvarende energi har kvotesystemet været udsat for to (sammenhængende) kritikpunkter: Kvoteprisen er for lav og dernæst at den har været ganske svingende. Det første skyldes fundamentalt set, at kravet til reduktioner i CO2-udledninger indenfor kvotesektoren frem til 2020 har været ret beskedne samt at den økonomiske krise fører til en forventet samlet mindre ud-ledning af CO2 frem til 2020 end oprindeligt forudset. Ved vedtagelsen af EU's klima- og energipakke i 2008 forventedes en kvotepris på omkring 225 kroner per ton ved en 20 procents reduktion af udledning fra kvotesektoren.

14 CE(2008)

Kapitel 4 INTERNATIONALE INDSATSOMRÅDER: FIRE FOKUSOMRÅDER


24

Den økonomiske krise har imidlertid ført til et betydeligt fald i energiforbrug og efterspørgsel efter kul og gas til elproduktion: derfor venter EU-Kommissionen nu kun en kvotepris på 120 kr. per ton15. Det bety-der at en betydelig mindre mængde af VE er rentabel i fravær af specifik støtte. EU-Kommissionen anslår nu (maj 2010) at man med overgang til en reduktion på 30 procent for drivhusgasser nu ender med den samme carbon pris på 225 kroner per ton, som oprindeligt forudsat med EU's klima og energipakke fra 2008. Kvoteprisens effekt på VE-tilskyndelser er anskueliggjort i figur 4.1, som viser produktionsomkostninger for forskellige VE-kilder startende med de billigste og sluttende med de dyreste såsom solceller. Pointen med udbuds med en kvotepris på 210 kroner vil kun omtrent yderligere 300 Kr. Pr. MWh. (benævnelse på ak-sen) blive rentabel medens en kvotepris på 225 fører til et yderligere løft til godt 400 kr. pr. MWh. Det for-ventes at el og fjernvarmesektorens bidrag til VE kræver en forøgelse af produktion på op mod godt 1100-1200 TWh. ift. 2005 niveau for at EU kan nå 2020 mål. Det betyder selvsagt også, at der skal ganske bety-delige tilskud oven i kvoteprisen ved en 30 procents målsætning for at nå dette mål. Modsat vil de kvotepri-ser som forventes nødvendige for at nå fremtidige klimamålsætninger efter 2020 samtidig med teknologiske fremskridt for VE og forventelige højere priser på kul og gas kunne gøre ganske meget VE rentabel på lang sigt uden driftsstøtte.

Figur 4.1 Stiliseret udbudskurve for ekstra vedvarende energi til el og fjernvarmeproduktion i EU ift. til

2005

Note: Elpris ved kul i 20 og 30 procent reduktion af drivhusgasser er baseret på en kvotepris på henholdvis 225 og 120 kroner per ton. Det er så omregnet til en et tillæg til elpris via antagelser om omkostningseffektivitet på et marginalt kulkraftværk

Kilde: EU-Commission (2010), Eurelectric(2008) og egne beregninger

Variation i og manglende forudsigelighed for den fremtidige kvotepris vil også i selv kunne dæmpe investe-ringer i low carbon teknologier fordi det skaber større usikkerhed om fremtidig rentabilitet. Nogle studier

15 EC(2010a)

0

500

1000

1500

2000

0 200 400 600 800 1000 1200 1400 1600 1800 2000

Biomasse varme

Geotermisk

Hydro

Varmepumpe

Landvind

Bioaffald

Biomasse

Havvind

Biogas

Solceller

Tidevand

Bølge

Elpris ved kul, 20% reduktion

Elpris ved kul, 30% reduktion

Twh.

Kr. pr. MWh.


25

har anslået at for investeringer med meget lang tidshorisont kan sådan usikkerhed øge kravet til rentabilitet med op til knap 40 procent.16 Der er således samlet set en række gode argumenter for EU vælger at gå efter 30 procents løsningen og/eller at der lægges en bund under kvoteprisen. For det første svarer dette til den kvotepris man forventede i 2008 i forbindelse med vedtagelse af EU's klima- og energipakke. For det andet vil det i et vist omfang aflaste be-hovet for driftsstøtte for at nå VE målet på de 20 procent. For det tredje vil det styrke tilskyndelser til ener-gibeparelser, som har et betydeligt og omkostningseffektivt potentiale for at reducere drivhusgasser og bi-drage til øget forsyningssikkerhed for EU. For det fjerde vil EU's langsigtede ambitioner om reduktion af udledning af drivhusgasser på 80 til 95 procent i 205017 fører til krav om priser på globalt plan for udled-ning af CO2, der er også ganske hurtigt efter 2020 vil ligge betydeligt over de priser der forventes i de kommende 10 år selv med en 30 procent reduktion i EU.18 For det femte kan det også styrke en ganske be-tydelig innovation inden for energiteknologier. Ny undersøgelser tyder på at en stigning i energiprisen på 10 procent fører til en stigning i virksomhedernes patentansøgninger på 5 til 20 procent over en 5-10 år pe-riode19. Det skyldes ganske enkelt at med højere energi og CO2 priser øges forbrugeres og forbrugeres ge-vinster ved at købe low carbon løsninger. Det ved producenterne og derfor afsætter de forskningsmidler til at udvikle sådanne produkter. Samtidig skal de markedsbaserede mekanismer og klimapolitikken udbredes globalt.Samtidig skal de markedsbaserede mekanismer og klimapolitikken udbredes globalt.Samtidig skal de markedsbaserede mekanismer og klimapolitikken udbredes globalt.Samtidig skal de markedsbaserede mekanismer og klimapolitikken udbredes globalt. Andre industrire-gioner som USA, Australien, Japan har allerede overvejelser om at indføre cap-and-trade system for elsekto-ren. I næste række kunne de knyttes sammen på OECD plan som også lagt op til i flere OECD-studier. Det vil give danske virksomheder gode fælles rammebetingelser for at markedsføre og sælge produkter i ho-vedparten af deres nuværende markeder. Hertil kommer at lande som Kina, Indien og Brasilien må inddrages. Den globale klimapolitik har længe været baseret på opdeling af kloden i to regioner: (1) Udviklede lande med reduktionsforpligtigelse (anneks 1) (2) resten af verden har blødere mål. Det giver ingen mening fremadrettet: Kina, Indien med flere står for en større og større del af verdens udledninger og i 2050 forventes deres udledninger alene at overstige det krævede globale loft for udledninger jf figur 4.1. COP15 bevægede sig i retning af at anerkende dette faktum.

16 Blyth et al (2007), ”Investment risks under uncertain climate change policy”, Energy Policy 35. 17 European Council (2009), 18 Over 500 kr. per tons i 2030 i Bosetti et al(2009), side 39. 19 Endnu ikke publiceret rapport fra Copenhagen Economics


26

Figur 4.1 Fordeling af drivhusgasser på forskellige globale regioner, 1990-2030

Kilde: IEA (2009), Höhne et al (2009), Copenhagen Economics

Der ligger en betydelig udfordring men også betydelige gevinster ved at sammenstykke en politik overfor udviklingslande, der giver dem tilskyndelser til at deltage effektivt i den globale klimapolitik. Reformerede Clean Development Mechanisms (CDM)20 kan billiggøre de rigere regioners klimaforpligtigelser, overføre teknologi og åbne nye markeder og ikke mindst levere øget energieffektivitet til udviklingslande, som ofte har ineffektive og subsiduerede energisystemer som beskrevet ovenfor. Givet den danske energibranches be-tydelige kompetencer indenfor dette område som beskrevet i kapitel 2 ligger der et betydeligt eksportpoten-tiale på dette område. Det er derfor vigtigt at få klarhed om hvorledes dette marked kommer fungere efter 2012. COP15 bekræf-tede at markedsbaserede systemer for teknologioverførsel skulle fortsætte, men ikke nødvendigvis i uændret form. Der har været en del kritik af at sådanne projekter reelt ikke leverer reelle reduktioner af udledninger af drivhusgasser. Der er imidlertid ikke opnået enighed om konkrete reformer, hvilket betyder at der meget betydelig usikkerhed om, hvordan projekter med levetid efter 2012 bliver krediterede i lande med reduk-tionsmål. Endelig skal handelsbarrierer for eksport af dansk energiteknologi reducerEndelig skal handelsbarrierer for eksport af dansk energiteknologi reducerEndelig skal handelsbarrierer for eksport af dansk energiteknologi reducerEndelig skal handelsbarrierer for eksport af dansk energiteknologi reducereseseses. Det drejer sig dels om at få udviklet standarder for energieffektivitet der er så globalt dækkende som muligt. Dels drejer det sig om at få reduceret egentligt handelspolitiske barrierer. En række af de mest hastigt voksende markeder for energi-teknologi som Kina, Indien og USA er præget af uformelle barrierer i forbindelse med offentlige indkøb og udbud (f.eks. Kina, Canada og USA) og samtidigt relativt høje tariffer på import f.eks. på turbiner til vindmøller (f.eks. Brasilien, Kina, Indien). Givet dansk erhvervsliv stærke positioner på disse områder bør reduktion af sådanne barrierer stå højt på listen over danske prioriteringer for EU's fælles handelspolitik overfor disse lande.

20 Meget forsimplet er indholdet i et CDM projekt, at en virksomhed i et land uden reduktionsmål gennemfører en reduktion af sin udledning af drivhusgasser relativt til en såkaldt ”baseline” dvs. i fravær effekterne af projektet. De nødvendige investeringer, der er forbundet hermed leveres af en virksomhed, der er underlagt EU's kvotesystem. Den pågældende virksomhed får således en kredit får den sparede drivhusgasudledning, der har samme værdi som at købe en kvote. Der er lagt begrænsninger på hvor mange kreditter der samlet kan købes.

0

10

20

30

40

50

60

70

80

1990 2005 2020 2030

GHG emissions (Gt)

Annex 1

CHN+IND

ROW

2 degrees


27

4.2. BEGYNDENDE HARMONISERING AF EU'S VE-POLITIK EU’s VE politik der desværreEU’s VE politik der desværreEU’s VE politik der desværreEU’s VE politik der desværre halter betydeligt bagefter udviklingen af kvotesystemethalter betydeligt bagefter udviklingen af kvotesystemethalter betydeligt bagefter udviklingen af kvotesystemethalter betydeligt bagefter udviklingen af kvotesystemet. EU opererer i dag med 27 nationale støttesystemer for VE med forskellige typer af støtteinstrumenter og -niveauer for de samme typer af teknologier jf boks 4.1 De justeres også løbende på nationalt plan, det gælder både i Dan-mark og de omkringliggende lande såsom UK og Tyskland.

Boks 4.1 Variationen af VE støttesystemer i EU's medlemslande Nedenfor er beskrevet nogle elementer i forskelligheden i støttesystemer i EU med fokus på hovedtræk i støttemeto-den. Men hertil kommer at der er forskelle i i de faktiske effektive støttesatser for samme teknologi og forskellige betingel-ser for VE producenters integrration i de fysiske elmarkeder (betaling for ledningsnet etc.) TerminologierTerminologierTerminologierTerminologier I et feed-in system får producenten en garanteret afregningspris for en tidsbestemt periode, som afhænger af den konkrete teknologi. Med en fast feed-in er afregningen uafhængig af prisen på engrosmarkedet for strøm medens der i systemer med pris-tillæg loves en fast tillæg oveni strømprisen i en tidsbestemt periode. Med grønne certifikater gives også et tillæg til engrosprisen men tillægget er ens for alle teknologier med det sigte at styrke konkurrence og fremme effektive VE-løsninger. En nærmere beskrivelse af fordele og ulemper findes i CE(2010) Feed in systemerFeed in systemerFeed in systemerFeed in systemer 18 lande ( Østrig, Bulgarien, Tjekkiet, Danmark, Tyskland, Grækenland, Frankrig, Ungarn, Irland, Italien, Letland, Litau-en, Luxembourg, Holland, Portugal, Slovakiet, Slovenien, Spainen) bruger differentierede feed-in tariffer og pris-tillæg.

- I de fleste tilfælde er feed-in tariffer tidsbesgrænsede, bortset fra Spanien (Laver priser efter 15 – 25 år) og Letland (For solceller). Tidsbegrænsningen er forskelligt udformet, f. eks. I Ungarn, hvor perioden er afgjort af energimyndighederne.

- Østrig, Spanien og Slovenien sænker raterne efter et fast antal år Cypern og Estland (For 12 år) bruger faste feed-in tariffer. Pris-tillæg bruges i Danmark (10 år) og i Holland. Tjekkiet, Estland, Slovenien og Spanien tilbyder valget mellem feed-in tariffer og pris-tillæg. Grønne certifiaterGrønne certifiaterGrønne certifiaterGrønne certifiater Grønne certifikater bruges i Belgien, Italien, Polen, Rumænien, Sverige og UK.

- Programmerne er tidsbegrænsede, bortset fra i Polen - Belgien sætter minimumspriser (forskelligt over regioner), Polen beregner en pris ud fra gennemsnitlig mar-

kedspris over foregående år) og i Rumænien må priserne maks falde 24 – 24 euro op til 2012. Kun Sverige og UK bruger ikke garanterede priser.

- Litauen har forpligtiget sig til at bruge grønne certifikater efter 2020 UdbudUdbudUdbudUdbud Danmark (Havvindmøller), Frankrig (Vind, biomasse og biogas), Letland (Vind over 0,25 MW) og Portugal (vind og bio-masse) Malta tilbyder investeringstilskud.

Kilde: EU Comission (2010b)

Det vanskeliggør langsigtet planlægning ikke mindst for danske virksomheder med behov for eksport. Manglen på et indre marked i EU kan blive et betydeligt handicap for konkurrencen med USA, Kina og Indien som hver for sig opererer med på meget store hjemmemarkeder og som allerede har opbygget en meget betydelig produktionskapacitet på både vindturbine og solceller og med betydelige forventede vækst-rater i de kommende år som diskuteret i kapitel 2.


28

Der er også betydelig ekstraomkostninger ved bruge rent nationale løsninger i stedet for at aktivere det in-dre marked. Flere studier tyder på at det kan komme til at udgøre 60 til 100 milliarder kroner om året på EU niveau jf. tabel 4.1. To studier fokuserende respektivt på det nordiske og det baltiske områder tyder til-svarende på betydelige gevinster ved regionalt samarbejder om implementering af VE-mål. Ekstraomkost-ninger ved en national tilgang i stedet for en EU/regional tilgang kan perspektiveres ved at sætte det i for-hold til midler afsat til langsigtet innovation: ekstraomkostningen ved at nå VE målene udgør således 2 til 4 gange de offentlige FU-midler til energiforskning i EU-landene. Tabel 4.1 Ekstraomkostninger ved VE-mål implementeret nationalt sat i forhold til offentlige FU mid-ler til energiområdet Region Offentlige FoU-midler til energi, 2008

Mia.kr.

Ekstraomkostninger ved produktion, ved

fravalg af indre marked for VE, per år,

mia.kr. i perioden frem til 2020

EU 25 60-100

Baltiske stater 10 4

Norden 2 1,5

Note: Den Baltiske region består af Danmark, Estland, Finland, Tyskland, Letland, Litauen, Norge, Polen, Rusland og Sverige. De totale omkostningsbesparelser (hvis der ses bort fra totale RES-E omkostninger, som er investeringsomkostninger, driftsom-kostninger såvel som brændstofsomkostninger og varmevederlag til biomasseanlæg) som følge af et skift fra BAU til harmonise-rede kvotesystem

Kilde: 1) EU Commission(2008), EWI (2009), p.112, Eurelectric (2008), p.5, 3) Nordisk råd (2008), 2) Ea Energianalyse (2009) 4) Eurostat

Der ligger således en meget stor udfordring med at få målrettet EU's og nationale støttesystemer. Det nu-værende VE direktiv, der har cementeret retten til at operere med nationalt orienterede støttesystemer, vil først bliver evalueret i 2014. På den kortere bane handler det derfor om at udnytte mulighederne i direktivet og udvikle samarbejdet. Sverige og Norge er i gang med at forhandle om at implementere EU's VE-direktiv med et harmonisret støttesystem baseret på såkaldt grønne beviser. Danmark kunne deltage i dette samarbejde med det perspek-tiv at lade det udvikle til en bredere kreds af lande interesseret i et sådant samarbejde f.ek.s i den Baltiske Region, hvor også Tyskland indgår. Et harmoniseret system for havvindmøller kunne være et andet fokuspunkt. Der er en betydelig risiko for at udrulningen af havvindmøller i de kommende årti i Nordsøområdet i højere grad styres af størrelsen af sub-sidier end de underliggende vindpotentialer: f.eks. for megen udbygning af kapacitet i Tyskland i forhold til Norden, Polen og Storbritannien21. En naturlig prioritet for Danmark med stort potentiale for både pro-duktion baseret på vind og produktion af vindmøller vil således være et mere harmoniseret system. Dan-mark bør presse på for at det indgår i EU-Kommisionens løbende evalueringsarbejde på implementering af VE-directivet.

21 EWI (2009)


29

4.3. STYRKELSE AF EU'S FORSKNINGSBUDGET FOR ENERGIRELATERET FORSKNING Der vil være betydelige gevinster ved at der afsættes flere midler til forskning i energirelevante temaer på EU's fælles budget. Det skal ses i forhold til at løsningen af de langsigtede ambitioner for klimapolitikken kræver en så markant reduktion af CO2-udledninger som ikke kan leveres af de eksisterende teknologier: det ville kræve dramatiske reduktioner af energiforbrug og tab af velfærd Danmark vil også vinde ved et styrket internationalt forskningssamarbejde på denne front både i nordisk, regionalt og EU regi. Danmark har også indenfor rammerne af det sidste 7. EU forskningsprogram fået en langt større andel af EU's midler på energiområder end hvad der svarer til Danmarks finansieringsandel: Danmarks bidrag til EU's budget er omtrent 2 procent medens vores andel af midler på energiområder er 8 procent, dvs. fire gange højere. Danmark bør allerede nu fokusere på, hvordan det 8. rammeprogram stykkes sammen både i indhold, stør-relse og på hvordan det fungerer i praksis. Åbenlyst vil det være i dansk interesse om nogle af de danske styrkepositioner og særlige udfordringer ved vindintegration fik en høj prioritet. Udover at styrke EU's forskningsindsats på dette område vil det også være væsentligt at sikre at danske forskningsmiljøer bliver stærkt forankret i det internationale samarbejde. En meget betydelig af denne op-gave vil være at danske forskningsinstitutioner, energivirksomheder mv. finder stærke partnere med samme fokusområder (vind, biomasse, brændselsceller mv). Det kan i sig selv være en forudsætning for at fastholde den stærke danske stilling i forhold til EU's forskningsprogrammer. Et særlig målsætning kunne være at sik-re EU placerer videnscentre i Danmark med fokus på en af disse styrkepositioner.

4.4. LIBERALISERING AF ELMARKEDER OG INVESTERINGER I GRIDKAPACITET Der er betydelige gevinster for Danmark og danske energivirksomheder ved en fortsat liberalisering af el-markeder og udbygning af EU's Grid system. Det vil have et dobbelt formål. For det første har Danmark har en meget effektiv elproduktion med mulighed for eksport til markederne omkring os. En sådan eksport forudsætter imidlertid tilstrækkelig kapacitet til nærområder og at elmarkederne omkring Danmark ikke mindst i Tyskland de facto tillader en effektive konkurrence fra danske værker. For det andet er der hensy-net til en stærk stigende forventet andel af vindbaseret elproduktion i hele Nordvesteuropa med vindandele i produktion allerede i 2020 på over 30 procent i Danmark og Irland og over 10 procent i Tyskland, Polen, Storbritannien og muligvis også Norge jf. figur 4.3. Det giver en betydelig mere volativ elproduktion og større risiko for mange perioder med betydelig overskud eller underskud af strøm i de enkelte lande og regi-oner i Nordvesteuropa i fravær af en betydelig udbygning af ledningsnet og markedsreformer, der kan sikre at strømmen bliver transporteret fra overskuds til underskudsregioner. Det vil sige dels indenfor ”vindzo-ne”-området dels væk fra Nordvesteuropa som indikeret ved den brede pil i figur 4.3 Det vil have to sam-fundsmæssige og erhvervsøkonomiske gevinster. El kan flyttes derhen hvor forbrugerne har mest brug for den og lokale energiproducenter i området bliver ikke straffet med meget lave priser ved overskud af el til ugunst for langsigtede investeringer i energisektoren, herunder opstilling af vindmøller.


30

Figur 4.3 Forventede vindandele i energiproduktion i Nordvesteuropa, 2020.

Kilde: EWEA (2009), DG TREN (2008), ENTSO-E (2010) og EWIS(2010)

> 25 %

10 – 25 %

0 - 10 %

Potentiel vindproduktion I andel af forbrug

Fremtidigprimærvindgrænse

37 %

Lavpris områdeLavpris områdeLavpris områdeLavpris område

HøjprisHøjprisHøjprisHøjpris områdeområdeområdeområde

21 %

14 %

9 %

31 %

13 %

23 %

2 %

8 %


31

Kilder til tabel 2.1 GHG emissions til omregning af base years: IEO (International Energy Outlook)

GHG i EU: Climate action tracker GHG I DK: Decision No 406/2009/EC of the European Parliament and of the Council of 23 April 2009 on the effort of Member States to reduce their greenhouse gas emissions VE i DK og EU: DIRECTIVE 2009/28/EC on the promotion of the use of energy from renewable sources Global GHG reduktion: CE baseret på UNFCCC: Stigning hvis OP15 pledges op-fyldes Energibesparelse I EU: Action Plan for Energy Efficiency: Realising the Potential Energibesparelse danmark: Energistyrelsen (2008), 'Fremskrivning af Danmarks energiforbrug frem til 2025 samt udledningen af drivhusgasser - og effekten af ener-giaftalen af 21. februar.', tabel 3.

Kilder til figur 3.3

Fase

Danske instrumenter

(årligt tilskud i mio. kr.)

Nordiske instrumenter


EU-instrumenter


Grundforskning Det strategiske Forskningsråd (160)

Basismidler (60)

FoU Energinet.dk ForskEL (131)

Nordisk Topforskning (80,0)

FP7-Energy, forskning (14)

DE Elforsk (25)

Højteknologifonden (64)

Demonstration EUDP (298)

FP7-Energy, demonstration (20)

Markedsmodning Energinet.dk ForskVE (26)

Driftsstøtte Skatteudgift til biomasse (1.300)1)

PSO-bidrag til finansiering af miljøven-lig el-produktion (2300)

Total Total Total Total 4365436543654365 80808080 34343434

1) Den konkrete beregning forudsætter en række antagelser om hvordan fordelen ved fælles produktion af el og varme

(kraftvarmefordel) fordeler sig på el- og varmeproduktion. Beregningen er baseret på standardantagelser brugt i officielle

danske beregninger, som formentligt henregner for meget af fordelen til varmeproduktion og dermed modsat også

overvurderer biomasseforbrug til elproduktion. De 1300 mill.kr er derfor nok et overkantsskøn for fordelen ved, at der

ikke beregnes energiafgift af biomasse til varmeproduktion.

APPENDIKS - KILDER TIL TABELLER OG FIGURER


32

Kilder samt forklaring af diverse energipolitiske midlers placering i modningsprocessen vil blive gennemgået nedenfor. Med mindre andet er angivet, stammer tallene fra 2009. Noter til tabel:

1. Det Strategiske Forskningsråd (DSF): � Kilde: Energinet.dk (2010), “Energi forskning, udvikling og demonstration 2010”,

Dansk Energi mm, Tabel 1.1, side 6 � Placering: Er placeret under baggrundsforskning i Energinet.dk (2010), “Energi forsk-

ning, udvikling og demonstration 2010”, Figur 2.1, side 19 2. Basismidler:

� Kilde: Ministeriet for videnskab, teknologi og udvikling (2009), ”Kortlægning af Klima-forskning i Danmark”, Figur 2.3, side 18

� Placering: Er placeret under baggrundsforskning i Energinet.dk (2010), “Energi forsk-ning, udvikling og demonstration 2010”, Figur 2.1, side 19

� Kommentar: Data er for 2007. 3. Energinet.dk ForskEL:

� Kilde: Energinet.dk (2010), “Energi forskning, udvikling og demonstration 2010”, Tabel 1.1, side 6

� Placering: Er placeret under forskning og udvikling i Energinet.dk (2010), “Energi forskning, udvikling og demonstration 2010”, Figur 2.1, side 19

4. DE Elforsk: � Kilde: Energinet.dk (2010), “Energi forskning, udvikling og demonstration 2010”, Tabel

1.1, side 6 � Placering: Er placeret under forskning og udvikling i Energinet.dk (2010), “Energi

forskning, udvikling og demonstration 2010”, Figur 2.1, side 19 5. Højteknologifonden:


� Placering: Er hovedsageligt placeret under forskning og udvikling i Energinet.dk (2010), “Energi forskning, udvikling og demonstration 2010”, Figur 2.1, side 19

6. EUDP: � Kilde: Energinet.dk (2010), “Energi forskning, udvikling og demonstration 2010”, Tabel

1.1, side 6 � Placering: Er hovedsageligt placeret under demonstration i Energinet.dk (2010), “Energi

forskning, udvikling og demonstration 2010”, Figur 2.1, side 19 7. Energinet.dk ForskVE:


� Placering: Er placeret under markedsmodning i Energinet.dk (2010), “Energi forskning, udvikling og demonstration 2010”, Figur 2.1, side 19

8. Skatteudgift til biomasse: � Kilde: Beregninger af Copenhagen Economics for 2010.


33

� Placering: Driftsstøtte er en ny kategori (ift. figur 2.1 i Energinet.dk (2010), “Energi forskning, udvikling og demonstration 2010”, side 19). Det er derfor en vurdering af Copenhagen Economics, at skatteudgift til biomasse skal placeres i kategorien ’driftsstøt-te’.

9. PSO-bidrag til finansiering af miljøvenlig el-produktion: � Kilde: Energistyrelsen (2009), notat ”Status for PSO-omkostninger til miljøvenlig elpro-

duktion” � Placering: Driftsstøtte er en ny kategori (ift. figur 2.1 i Energinet.dk (2010), “Energi

forskning, udvikling og demonstration 2010”, side 19 som ellers er blevet brugt til place-ring af diverse midler). Det er derfor en vurdering af Copenhagen Economics, at PSO-bidrag til miljøvenlig el-produktion skal placeres i kategorien ’driftsstøtte’.

� Kommentar: PSO-bridrag til miljøvenlig el-produktion er beregnet som et det gennem-snitlige bidrag fra 2005-2008. Bidraget dækker desuden også gasbaseret el-produktion.

10. Nordisk Topforskning: � Kilde: Ministeriet for videnskab, teknologi og udvikling (2009), ”Grøn Forskning - Sta-

tus og perspektiver”, side 25. � Placering: Fremgår af beskrivelsen af midlerne i den tilhørende kilde. Bliver sammenlig-

net med EU’s rammeprogram, hvor midlerne også går ind under kategorien ’FoU’ 11. FP7-Energy, forskning:

� Kilde: Energinet.dk (2010), “Energi forskning, udvikling og demonstration 2010”, side 144. Beregninger af Copenhagen Economics for 2010, baseret på at Danmarks andel af EU’s budget er 2,1 %.

� Beregning: Forskning og brændsels- og brintceller = 170 mio. EUR * 7,45 DKK/EUR * 0,021 = 27 mio. DKK.

� Placering: Er benævnt ’forskning’ i den tilhørende kilde, hvorfor den er placeret under FoU.

12. FP7-Energy, demonstration: � Kilde: Energinet.dk (2010), “Energi forskning, udvikling og demonstration 2010”, side

144. Beregninger af Copenhagen Economics for 2010, baseret på at Danmarks andel af EU’s budget er 2,1 %.

� Beregning: Demonstration = 130 mio. EUR * 7,45 DKK/EUR * 0,021 = 20 mio. DKK � Placering: Er benævnt ’demonstration’ i den tilhørende kilde, hvorfor den er placeret un-

der demonstration i tabellen.


34

Bosetti et al (2009), ”The role of R&D and technology diffusion in climate change mitiga-tion: New perspectives using the Witch model”, OECD, Working Paper

Brøndum & Fliess (2009), “Cleantech – guldægget i dansk økonomi”, rapport udarbejdet i samarbejde med DI, DI Energibranchen og Energistyrelsen

CE(2008), ”Energibeskatningens rolle: klima, energi og fordeling”, rapport fra Copenhagen Economics for Dansk Energi

Danmarks Forskningspolitiske Råd (2010), ”Dansk forskning – Store globale udfordringer og store globale muligheder”, Årsrapport

Doornbosch et al (2008), “Mobilising investments in low-emission energy technologies on the scale needed to reduce the risks of climate change”, OECD

Dræbye (2010), ”Grundlag for statistik for energierhvervene”, notat udgivet ved samarbejde mellem Energistyrelsen, DI Energibranche og Økonomi- og Erhvervsministeriet

EA Energianalyse (2009), “Sustainable energy scenarios – Energy perspectives for the Baltic Region”, rapport udgivet i samarbjde med Nordisk Råd og Baltic Development Fo-rum

EC (2008), “Annex to the Impact Assessment, Document accompanying the Package of Im-plementation measures for the EU’s objective on climate change and renewable ener-gy for 2020”, Working document

EC (2008), ”R&D specialisation: Strategic intelligence in priority setting”, rapport

EC (2009), “Decision No 406/2009/EC of the European Parliament and of the Council of 23 April 2009 on the effort of Member States to reduce their greenhouse gas emis-sions to meet the Community’s greenhouse gas emission reduction commitments up to 2020.”, EU lovgivning

EC (2009), “Directive 2009/28/EC on the promotion of the use of energy from renewable sources and amending and subsequently repealing Directives 2001/77/EC and 2003/30/EC”, EU direktiv

EC (2010a), “Support schemes for renewable electricity in the EU”

LITTERATURLISTE


35

EC(2010b),”Analysis of options to go beyound 20% greenhouse gas reductions and assessing the risk of carbon leakage”, SEC(1010)650)

EC(2010c), ”What is the growth potential of green innovation? A assessment of EU climate policy options”, Economic paper

Energi10 (2010),“ Årsrapport om de danske energiforskningsprogrammer”, www.elforsk.dk

Energinet.dk (2010), “Energi forskning, udvikling og demonstration 2010”, rapport udgivet i samarbejde med Energistyrelsen, EUDP-sekretariatet, Dansk Energi mm.

Energistyrelsen (2009), ” Status for PSO-omkostninger til miljøvenlig elproduktion”, notat

Eurelectric (2008), ”Reaching EU RES Targets in an efficient manner – benefits of trade”,

Eurostat, http://epp.eurostat.ec.europa.eu/portal/page/portal/eurostat/home

EWI (2009), ” European RES-E Policy Analysis - A model-based analysis of RES-E deploy-ment and its impact on the conventional power market”, draft final report

FORA (2009), “Kortlægning af Miljøteknologiske Virksomheder i Danmark”, rapport ud-arbejdet for Miljøministeriet

Forsknings- og Innovationsstyrelsen (2009), ”EU giver 242 milliarder til 10 udvalgte forsk-ningsområder – grib muligheden og få støtte til dit projekt”, rapport

Forsknings- og Innovationsstyrelsen (2009), ”Tænk internationalt – udbytte og udfordringer ved EU-projekter”, rapport

Forsknings- og Innovationsstyrelsen (2010), ”Erhvervslivets forskning, udvikling og innova-tion i Danmark 2010”, rapport

Forsknings- og Innovationsstyrelsen (2010), ”InnovationDanmark 2010-2013. Viden til virksomheder skaber vækst. Handlingsplan fra Rådet for Teknologi og Innovation”, rapport udgivet for Rådet for Teknologi og Innovation

Forsknings- og Innovationsstyrelsen (2010), ”Kortlægning af dansk deltagelse i EU’s fælles-skabsinstrumenter”, rapport


36

IEA (2009), “World Energy Outlook”, rapport

IEA (2010), “Opportunities to transform the electricity sector in major economies”, rapport for The Clean Energy Ministerial

IEA (2010), ”Transforming global markets for clean energy products”, rapport for The Clean Energy Ministerial

IEA data services, http://data.iea.org

Ministeriet for Videnskab, Teknologi og Udvikling (2009), ”Evaluering af Forsk 2015”, ud-arbejdet af Teknologisk Institut

Ministeriet for Videnskab, Teknologi og Udvikling (2009), ”Forskningsbarometer 2009 – Dansk forskning i internationalt perspektiv”, rapport

Ministeriet for Videnskab, Teknologi og Udvikling (2009), ”Grøn forskning – Status og perspektiver”, rapport

Ministeriet for Videnskab, Teknologi og Udvikling (2009), ”Kortlægning af klimaforskning i Danmark”, rapport lavet i samarbejde med Koordineringsenhed for Forskning i Klimatilpasning

Ministeriet for Videnskab, Teknologi og Udvikling (2010), ”Videnskabsministeriets hand-lingsplan for forskningsinfrastruktur”, rapport

Nordisk Råd (2008), “Promotion of renewable energy in the Nordic coutrnies”

OECD (2009), “The Economics of Climate Change Mitigation: Policies and options for Global Action beyond 2012.

OECD (2010), ”Interim Report of the Green Growth Strategy: Implementing our com-mitment for a sustainable future”, Meeting of the OECD Council at Ministerial Lev-el, rapport

Regeringen (2009), ”Erhvervsklimastrategi, globale udfordringer – Danske muligheder”, præsentation udarbejdet af Økonomi- og Erhvervsministeriet


37

Scientific & Technical Reference, http://re.jrc.ec.europa.eu/refsys/

World Bank (2010), ”Gross domestic product 2009”, hentet fra World Development Indi-cators database

Gerson Lehrman Group, http://www.glgroup.com/News/China--Taiwan-Market-Share-for-Solar-Cells--impact-on-U.S.-49273.html

WTRG Economics, http://www.wtrg.com/prices.htm

DG Taxation and Customs Union

INNOVATION OF ENERGY TECH-NOLOGIES: THE ROLE OF TAXES FINAL REPORT | 26 NOVEMBER 2010

Innovation of energy technologies: the role of taxes

2

COLOPHON

Author: Project manager Mr. Sigurd Næss Schmidt and Dr. Eske Stig Hansen, Mr. Jonatan Tops, Mr. Holger Nikolaj Jensen, Dr. Svend Torp Jespersen.

Client: European Commission DG TAXUD Date: 26 November 2010 ISBN: Contact: SANKT ANNÆ PLADS 13, 2nd FLOOR | DK-1250 COPENHAGEN

PHONE: +45 2333 1810 | FAX: +45 7027 0741 WWW.COPENHAGENECONOMICS.COM


3

Preface .................................................................................................................. 5

Chapter 1 Key findings .............................................................................................. 6 1.1. taxation of greenhouse gases and energy key driver for innovation ......................... 6 1.2. R&D policies as a supplement to taxation ............................................................. 8 1.3. Taxation needed to reap benefits of R&D policies ............................................... 10

Chapter 2 Energy taxation as a driver of innovation in energy technologies ........... 11 2.1. Reactions to environmental tax and price increases .............................................. 12

Can we estimate direct tax impacts? ........................................................................... 12 Medium to long run elasticities can be substantial ..................................................... 15

2.2. Energy prices and taxes effect on penetration of low carbon technologies ............. 19 Renewable energy sources for electricity ..................................................................... 19 Adoption of energy efficient household appliances ..................................................... 20 Fuel efficient cars ....................................................................................................... 21

2.3. Higher Taxes lead to green technology innovations ............................................. 24 Powerful effect on R&D spending and green technology diffusion ............................ 24 Technology and innovation decisions have a global scope .......................................... 26

2.4. New empirical results on innovation impacts ....................................................... 27 Energy technologies ................................................................................................... 28 Patents, taxes and energy prices .................................................................................. 30 Graphical inspection of patent data and relationship between innovation and prices .. 35

Chapter 3 Clarifying the role of taxation vis-a-vis direct innovation policies .......... 43 3.1. Different roles for different instruments ............................................................... 43

The double externality problem ................................................................................. 44 The long term credibility problem ............................................................................. 47 R&D crowding out ................................................................................................... 48 Command-and-control systems and innovation ......................................................... 49

3.2. The interaction between taxation and R&D support ........................................... 49 Support for green technologies, their deployment and rebound effects ....................... 50 Timing of R&D support and taxation ....................................................................... 51

3.3. Time horizon and stringency of policy targets ...................................................... 54 Time horizon ............................................................................................................ 54 Stringency of policy targets ........................................................................................ 56

3.4. The efficient policy mix of carbon pricing and R&D support .............................. 57

Appendix A: Empirical analysis of patent data .............................................................. 70 A.1 Methodological framework ...................................................................................... 70 A.2 Results ..................................................................................................................... 72

TABLE OF CONTENTS


4

Patenting in Lighting Technology ............................................................................. 72 Patenting in LED Technology ................................................................................... 74 Patenting in Boiler Technology ................................................................................. 76 Patenting in Biomass Technology in Buildings .......................................................... 77 Patenting in Ventilation Technology ......................................................................... 78 Patenting in Fuel Efficiency for Motor Vehicles ......................................................... 80 Patenting in Pulp and Paper Manufacturing .............................................................. 80

Appendix B: Description of CERIM ............................................................................. 82

Appendix C: Description of IPC codes ......................................................................... 85

Microsoft Office Excel 97-2003 Worksh ........................................................................................................ 85


5

The EU and its member countries have recently at the March 2010 European Council meet-ing reiterated their commitment to ambitious long term goals to deal with climate change and energy security. To this end, the EU is ultimately committed to take a decision whether to move to a 30% reduction by 2020 compared to 1990 levels as a first step to achieve the ultimate target of staying below a 2°C increase in global temperatures compared to pre-industrial levels. This will in practice require substantial reduction in energy related CO2 emissions and de-ployment and development of low carbon energy technologies. In this context, the study fo-cuses on the importance of taxation of carbon and energy as a spur for innovation in such technologies, containing two main elements:

A policy based literature review of specific and direct links between energy taxes and innovation and in that context reporting the results of a major new economet-ric study using micro and macro data.

A policy based literature review of the merits of taxation relative to innovation and

R&D policies in attaining long term climate goals.

PREFACE


6

1.1. TAXATION OF GREENHOUSE GASES AND ENERGY KEY DRIVER FOR INNOVA-TION

The advantage of using taxation to spur innovation in energy technologies is just a mirror image of the advantage of using taxation to abate emissions in general. By taxing directly the quantity – for example emission of CO2 – the same incentive across all fields of innovation will be made available in order to save energy and/or reduce CO2 emissions. Hence remov-ing the need for policy makers to “guess”, based typically on incomplete information, where innovation activities should be focused. The effects on innovation are of an “induced” nature containing three steps. First, appropri-ate tax regimes can make it more expensive for private and industrial consumer to use (fossil) energy sources. Second, this in turn increases the demand for technical solutions that either save energy or use low fossil content energy sources and thereby improving the economic vi-ability of such technologies. Thirdly, this (re)directs the innovation efforts of enterprises in that direction; this is what we term “induced” innovation. These effects are not just of a theoretical nature, but are demonstrated in many applications of energy use over many decades. Our review of the literature suggest that the long term ef-fects of capturing all the three effects imply that an increase in energy prices or taxes of 1 per cent often leads to a fall in energy use of 1 per cent of more (c.f. chapter 2.1 for details). In this study, we review the empirical literature on induced innovation effects, and we con-duct own empirical investigations of the relationship. We will highlight the three most important conclusions from the new empirical research in this study seen from a policy perspective (c.f. chapter 2.4 for details). First, substantial increases in energy taxation can drive forward very substantial increases in innovation. Looking at seven different technology classes, we find (statistically significant) positive impacts on patenting activities from energy taxation for five technologies. The two other technologies related to lighting have too small effects to be well determined statisti-cally. Our results suggest that a one percentage point increase in the tax share of total user costs induces a 0.3-2.4 percentage increase in patenting, c.f. Table 1.1. This is indeed quite substantial. Table 1.1: Price and tax effects for different technologies Technologies Lighting LED Biomass in

buildings

Heat boilers Ventilation Motor

vehicles

Paper and

pulp

Price effect 0 0 0.28 0 0 0 0.47

Tax effect 0 0 0.28 2.33 2.37 2.19 0.47

Note: Long run estimated elasticities significant at 5 % confidence level. Due to the estimation strategy, taxes will always ‘inherit’ the price effect, but will be flexible to deviate if statistically significant.

Source: Copenhagen Economics

Chapter 1 KEY FINDINGS


7

Second (and closely related to the first conclusion), the tax induced innovation is signifi-cantly higher than the price induced innovation. This highlights the role of taxation as a credible long term instrument sending the right investment signals to innovators. As such, this conclusion is well supported by the literature, see chapter 3.1-3.2. However, we have reasons to believe that part of the difference between taxes and prices in our empirical results may be attributed to methodological issues. Third, the speed and size of innovation effects from energy/carbon taxes depend on a num-ber of well defined characteristics of the products and processes affected by the tax which are discussed in more detail in the report, c.f. chapter 2.2-2.3. In particular:

Energy use is typically just one (cost) component of a broader service produced by different types of capital equipment: gasoline is inserted into a car to produce a transport service; electricity is inserted into a light bulb to allow it to light up rooms etc. The higher the cost share presented by energy costs of deployment of the energy consuming product, the bigger the relative effect on user costs from energy taxes. In short, a 20 per cent increase in energy prices will lead to larger in-creases in the costs of using a car than a computer. This implies that the choice of a computer will be less driven by its energy consumption than the choice of a car would be. In turn this implies that energy cost driven innovation will be focused – naturally – in areas where energy account for a large share of the costs.

The speed of effect: the time from increase in taxes to effective introduction of new technology depends very much on the production and innovation cycle in the particular industry. The lead time from a higher tax rate to a patent is 4-5 years on average and then it needs to be converted into real products that will be deployed in the market over time. We can expect quick effects with simple, though highly energy consuming, products like hair dryers (not investigated here); slower effects with cars; and longer term effects on, e.g., heavy industrial equip-ment used in paper and pulp production. See chapter 2.4 and 3.3 for more on speed and time lags.

In addition to providing support for these findings, our literature review provides the follow-ing two main conclusions useful in assessing the size and adequacy of price/taxation induced innovation: The first conclusion is that global or at least regional tax rates should have broader and stronger effects on innovation than isolated tax rates in a few countries. The basic reasoning is that innovation strategies will be driven by simple market size concerns: the larger the market affected by a tax on energy, the larger the incentives for firms to spend their scarce innovation resources on responding to such taxes. As a counterpart to this, the innovation gains that small countries can achieve by imposing unilaterally higher tax rates on their own consumers and industries will be limited by two types of “leakage”. First, firms may consider relocation rather than investments in abatement technology. Second, firms producing poten-


8

tial abatement technology may hold back on such investments because the local markets ac-count only for a fraction of their global sales. See chapter 2.3 for more details. The second and very important conclusion relates to specific policy design and is partly de-rived from the conclusions above:

Some patience is needed in reaping the benefits of tax driven innovation with the speed depending on the length of product and innovation cycles. These lead times are important to keep in mind when setting instruments for obtaining medium ver-sus long term climate targets.

However, even if contributions to year 2020 targets from taxation induced innova-tion may be limited due to lead times, taxation will still be an extremely important instrument as the immediate effects on primary consumption and shifts towards energy efficient equipment will be the main engine through which these targets can be achieved.

In addition, for innovation with long expected time lags, it is essential to establish a long term credibility of maintained high level of tax rates to fix incentives for in-vestment.. This is also discussed below.

1.2. R&D POLICIES AS A SUPPLEMENT TO TAXATION While taxation can be a very effective driver of innovation in energy technologies, there are two basic arguments suggesting that energy/carbon taxation needs to be complemented with public research grants and other technology policies supporting long term innovation. The first is the so-called double externality problem. Carbon pricing is imposed because the costs for the society of emissions exceed the costs of private or industrial consumer that emit it. Hence, in line with standard environmental policy principles, by imposing a carbon price, we at one and the same time make the polluter pay and reduce emissions. However, at the same time, we have a classical externality problem in the production of knowledge: the bene-fits to the society from particular basic science may well exceed the private benefits from producing it (see also chapter 3.1). The second argument is the long term nature of innovation efforts and, linked to this, the credibility problem policy makers are facing. Private firms will only invest in research now to reap future benefits if they believe that the policy framework in place when innovation ef-forts are turned into products and processes will reward them for their efforts. However, pol-icy makers will also know that once firms have spent billions of Euros on R&D, they will seek to bring the new products to the market, provided that the marginal revenues of doing so exceed marginal costs of production. So policy makers may promise high taxes on carbon forever but drop them once firms have made the irreversible R&D investments.


9

This conclusion though raises the question: How much can the costs of attaining climate change and energy policy objectives be reduced by supplementing taxation of emissions with direct public support? We will argue that it depends on two main issues. The first issue is classical within the field of R&D economics. Public research grants require public funding with resulting distortions from higher tax rates. (A typical estimate is that 1 Euro spent on R&D requires benefits equal to 1.20 Euro or more to compensate for distor-tions.). Furthermore, increased innovation driven by public funding in one field of eco-nomic activity tends to squeeze out other innovation activity including privately funded R&D. However, according to several contributions from the economic literature, R&D support is usually considered to bring forth more economic benefits than what it costs to tax-payers, at least in the dynamic context. This provides a clear trade-off. The benefits to be reaped by producing positive spill-overs from energy technologies need to exceed the costs of lost innovation elsewhere as well as dis-tortions from higher tax rates to fund R&D subsidies. By contrast, revenues from energy taxes can be recycled so as to neutralise their adverse effect on the labour market. The second issue is more directly related to the level of ambitions that the EU has commit-ted itself to, and the time frame for attaining these. To be very clear, the benefits that the EU can expect from new public research grants initiatives between 2010 and 2020 in meeting 2020 objectives should be relatively limited. As discussed above, the time lag from spending on R&D to results being deployed in new products and processes is often measured in dec-ades. Moreover, reducing CO2 and other greenhouse gases by 20 per cent – 30 per cent in the context of a wider global agreement – can largely and effectively be met by deploying ex-isting energy efficient technologies helped by carbon pricing. Moreover, substantial im-provements of these technologies will become economically viable for producers with rela-tively modest increases in carbon pricing. Moving beyond 2020 objectives, the picture becomes different. Reducing CO2 emissions with 50 to 80 per cent from 1990 levels while seeing continued growth will imply either massive reductions in energy intensity or the introduction of new low carbon technologies in a scale not seen before. The cost to consumers of such a drastic change in living pattern would be very large, hence increasing vastly the value of technologies that could provide low carbon energy. These findings are confirmed in a number of recent studies. However, the increasing weight of R&D support to attain long term goals in climate and energy policies does not imply that taxation becomes less relevant over time. Indeed, a num-ber of studies have shown that carbon prices will have to rise further beyond 2020 even with very ambitious R&D policies.


10

1.3. TAXATION NEEDED TO REAP BENEFITS OF R&D POLICIES Well targeted R&D policies focused on solving research externalities still need to be backed up by continued strong carbon pricing by way of taxes and/or cap-and-trade systems. There are three basic arguments. First, public R&D support to increase the energy efficiency of fossil fuel technologies – combustion engines etc – will lead to more energy efficient cars on the roads, but also to lower costs of driving. Recent research from Germany suggests that up to 60 per cent of the energy savings from more energy efficient cars are transformed into consumers driving longer distances and or buying cars with more performance, a pattern often called the re-bound effect. Secondly, for end-of-pipe technologies such as coal based Carbon Capture and Storage power plants, the benefits are exclusively CO2-savings, while the output – electricity – is ex-actly the same as for traditional fossil based power plants. So these plants will never be de-ployed unless they receive a premium when selling electricity: despite up-front subsidies to-tal costs per unit sold will exceed traditional power plants. It is the role of carbon pricing to deliver this premium. Thirdly, R&D policies supporting renewable energy may well lead to a reduction in demand for fossil fuel, but that will at the same time lead to a reduction of crude oil, coal and gas prices on a global scale, triggering higher second round demand for such fossil fuels. The only real response possible is higher carbon taxes at a global level including in the EU.


11

From an economic perspective, taxes are a cost-efficient instrument in climate change miti-gation. Taxes (as well as emission trading schemes) create clear economic incentives to re-duce pollution, whilst being easy to implement. Alternatives, such as technology standards, typically only create medium term effects, unless continuously updated. A key problem with standards is to ensure that they align abatement costs across sectors along the lines of taxa-tion systems.1 However, the full abatement effects from energy taxation will most often first be seen dec-ades after the introduction. Basically, this stems from the three step nature of the reactions of consumers and producers to change in economic incentives. To be more specific about the difference between the three effects, we need to make clear that energy demand is a derived demand, derived from the demand for the output of some processes, e.g., a car engine pro-viding transportation, c.f. Figure 2.1. The short run effect of a tax (price) increase on energy inputs is that we reduce the amount of transportation. The medium run effect is that we buy smaller and less energy consuming cars, while the long run effect is that we invent hydrogen cars. Note the interdependence amongst the effects. If consumers are not hurt by the tax in-crease in the short run in form of reduced transportation, they are not likely to change their behaviour in the medium run either, and so there is no market for new inventions. Figure 2.1: Three channels for reduction in energy consumption/CO2 emissions over time

short run

medium run

long run

EU p

ollu

tion

redu

ctio

n 20

20(K

yoto

: 20

% b

elow

199

0 le

vel)

reduce activities creating pollution (e.g. less transportation)

conduct activities in less polluting way using known technologies (e.g. smaller and more fuel efficient cars)

conduct activities in less polluting way using new technologies (e.g. hydrogen cars)

Note: The choice of the Kyoto reduction target only serves as an example Source: Copenhagen Economics 1Popp et al (2009)

Chapter 2 ENERGY TAXATION AS A DRIVER OF INNOVATION IN

ENERGY TECHNOLOGIES


12

In this study we focus on this third element, the longer term effect on innovation. But it is very important to understand that this innovation is an induced or derived effect from con-sumers’ and producers’ short, medium and long run reactions to price changes. If consumers do not react to prices by consuming less energy and/or change the composition of energy-consuming capital, then innovators will not put research funds into innovation.2 We will structure this chapter according to these three channels. First, in section 2.1, we briefly review evidence of how private and industrial consumers over time have reacted to changes in energy prices, including changes originating from taxes. Second, in section 2.2, we review more directly how penetration of low carbon technologies has been impacted by energy and tax prices. Thirdly in section 2.3, we measure how innovation activities as meas-ured by different indicators typically with medium to long term lags have responded to such consumer behaviour. Finally, in section 2.4, we present new econometric evidence of the ef-fect of energy prices and taxes in four areas of energy use, reviewing seven different tech-nologies.

2.1. REACTIONS TO ENVIRONMENTAL TAX AND PRICE INCREASES Taxes provide incentives for consumers and firms to reduce energy consumption with the existing holdings of electric appliances, cars, production machinery etc. However, as long as the capital stock is held fixed, the only possibility is to reduce the level of pollution creating activities. In the medium to long run, consumers and industries will also change the compo-sition of the capital stock towards environmentally friendly technologies thereby creating further reductions in pollution.

Can we estimate direct tax impacts? Direct and robust estimates of long term effects of taxes on energy consumption and devel-opment of new technologies are difficult to obtain as energy taxes historically have been rela-tively sparsely used. This is evident from Table 2.1 where we see that energy taxes typically only amount to a few percent of GDP in the USA, Japan and EU countries. Energy taxes are mainly focused on gasoline, with low taxation of inputs for heating (represented by oil).

2 Acemoglu et al (2009).


13

Table 2.1: The importance of energy taxes in EU, US and Japan, 2007

Electricity tax, euro /

GJ. Gasoline tax, euro / GJ Diesel oil tax, euro / GJ

Revenue from energy

taxes as share of GDP

EU, high (67 percentile) 26 26 21 3,5

EU, middle (33-67 percentile) 9 23 17 2,5

EU, low (33 percentile) 5 20 15 1,6

USA 0 3 3 0,8

Japan 2 17 7 1,7

Note: Energy taxes share of GDP includes all environmental taxes. These numbers are for 2007. The tax on elec-

tricity, gasoline and diesel consist of all taxes paid by the end-user, excluding VAT. Source: IEA Data services and OECD, http://www2.oecd.org/ecoinst/queries/TaxInfo.htm Despite the significant change in policy focus around the world between 1994 and 2007, environmental taxes do not seem to have become more important during this time period, c.f. Figure 2.2. Obviously, the tax revenue to GDP ratio is also influenced by more than the level of environmental taxes. For example, a high GDP growth and a diminishing public sec-tor could imply that the ratio would fall. However, in the period we consider, neither of these explanations seems to be of major importance. Figure 2.2: Environmental tax revenues, share of GDP, 1994-2007

Source: IEA Data Services

0

1

2

3

4

1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Pct.

EU High, average (DK, NL, PT, FI, CZ)EU Middle, average (SE, HU, LU, IE, AT, UK, DE, FR, SK)EU Low, average (BE, EL, IT, ES, PL)Japan United States


14

Hence, when seen from the macro level over the last three decades, energy taxes have had only a limited impact on the level and variation of end user prices relevant for decisions by private and industrial consumers in most EU Member States. However, some exceptions ex-ist, e.g., a few household end-user products such as gasoline and fuel oil, c.f. Figure 2.3. Figure 2.3: Energy taxes across products (excl. VAT), share of total end use prices in EU

Note: Simple average Source: IEA Data services Given the limited overall importance of environmental taxes, empirical work on the effects of taxation have to look at the historical effects from changes in energy prices as well as changes in tax rates3. A priori, we would expect demand reactions to be similar for a tax rais-ing the product price by 1 percent and a cost increase also raising the price by 1 percent. In-deed, the literature provides a large range of studies determining the behavioural response from general price changes including changes in tax rates.4 As an introduction to this literature, we need to underline the role played by demand elastic-ities, i.e., the percentage change in consumption by a percentage change in final price. In the next subsection, we will present ranges of elasticity estimates. If consumers do not respond to price increases in the short or medium run, i.e., we face zero price elasticity, then there is lit-tle role to be played by taxes in climate change mitigation. The behavioural change is neces-

3 See Killian (2007) for a good survey on impacts from energy price fluctuations. 4 Popp (2002) discusses this point for the direct price-patent relationship and concludes that price and tax move-ments will have similar impact. Flood et al (2010) uses a more elaborate political economy approach to examine dif-ferences in price and tax impacts on gasoline demand. They conclude that from a demand side perspective, which is relevant here, the impacts are equal. However, there seems to be a political response to tax levels from price fluctua-tions to some extent dampening the fluctuations.

0%

10%

20%

30%

40%

50%

60%

70%

80%

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

Gasoline, tax share of total end-use price

Fuel oil, tax share of total end use price

Electricity, tax share of total end use price


15

sary not only for the direct impact on pollution, but also on the economic incentives for conducting expensive R&D in green technologies. There are obvious methodological problems in getting price elasticities right. First, the de-mand for fossil fuel inputs is not only determined by price, but also by income. Richer na-tions use more fuels, and as the world economy grows, energy demand increases. Thus, esti-mating elasticities requires an adequate control for income effects. Second, equilibrium ef-fects must be taken into account. Observed data must be seen as both demand and supply responses, and disentangling demand from supply effects requires some methodological basis or identifying assumptions. Third, in the interpretation of elasticities it is important to dis-tinguish between pure consumption reduction and simple fuel substitution. For example, an elasticity of 1 may consist of 0.5 pure reductions in energy consumption and 0.5 substitu-tions to alternative energy sources. Obviously, when energy taxation hits the pollution di-rectly as in the case of carbon taxes, then substitution towards zero-carbon energy sources must be included as a potential beneficial response to taxes. But in the case of a specific fuel tax, say on oil, substitution towards natural gas is less beneficial. At this point, we should clarify that in this study, we are essentially looking at both substitution and pure reduction. Innovation can take place within low-carbon technologies and within energy-efficiency. Baring in mind these uncertainties, it seems relatively clear that long term effects from taxa-tion can be substantial. In the next paragraph we will show such demand responses in more detail.

Medium to long run elasticities can be substantial Looking across a very wide range of studies on the price elasticity of energy demand, there is a wide consensus that long term effects are two to three times higher than short term ones and are substantial, typically exceeding unity. This literature actually started as early as 1951 and despite the refinements in methodologies and broader availability of data, consistently with this contribution estimates are still found in the ballpark of 0 to -0.5 in the short run and -0.5 to -2 in the long run, c.f. Figure 2.4. The figure captures 67 studies with more 273 different elasticity estimates. Moreover, the figure demonstrates that measured as bands of estimates there is not that much difference between sectors and energy products. The only exception is residential demand which seems to have somewhat lower long run elasticity. However, it seems to be a reasonable assertion that long run estimates are typically not too far from -1.


16

Figure 2.4: Short and long run elasticity estimates across sectors and products

Note: The elasticities represented are based on a large number of empirical studies which are reviewed in the studies mentioned below as sources.

Source: Bohi and Zimmerman (1984), Bernstein and Griffin (2005), Dahl (1993), Newell and Pizer (2008), and Wade (2003).

-3

-2

-1

0

Commercial short run

Commercial long run

Industrial short run

Industrial long run

Residential short run

Residential long run

-3

-2

-1

0

Electricity short run

Electricity long run

Fuel oil short run

Fuel oil long run Natual gas short run

Natural gas long run


17

Table 2.2: Elasticity estimate averages across energy types and sectors Energy type Short run Medium run Long run

Coal 0.08 - 0.30

Electricity 0.23 0.49 1.15

Gasoline 0.22 - 0.65

Natural gas 0.28 1.15 1.72

Oil 0.15 0.66 1.73

Aggregate energy 0.23 0.40 0.63

Households 0.225 0.549 1.24

Commercial 0.256 0.26 1.57

Industrial 0.239 0.762 1.34 Sourc: Same sources as in figure 2.4 On the other hand, one may ask for the source of variation in these estimates. A few recent studies, in form of meta-analyses, have tried to assess this.5 The conclusions are as follows (further explanations follow):

The empirical methodology may influence the results significantly Elasticities change over time and with price levels Countries with similar economic structure may have quite different elasticities

Concerning the influence from the empirical methodology, meta-analyses show that some 40-50 percent of the variation may be attributed to methodological differences, implying that the remaining 50-60 percent must be attributed to real (economic) differences between samples.6 The most important difference arises from the data itself: Cross-country vs. time series esti-mations. Time series data have the advantage that the underlying economic entity with its basic structures is not changing between data points. Cross country data have the advantage that larger variations are typically present.7 When it comes to changes over time, a meta-analysis explicitly shows that electricity elastic-ities increased by 0.7 in absolute terms (that is, from, say, -1.0 to -1.7) immediately after the first oil price hike around 1973.8 Similar magnitudes are found for gasoline demand elastic-ities. We give three interpretations to this result.

5 See Espey and Espey (2004), Brons et. al. (2006), and Brons et al (2008). 6 See Espey and Espey (2004), Brons (2005), and Brons et et al (2008). 7 This touches a classic discussion in much applied analysis during the last decades. For an example directly discuss-ing the difference between data types and estimates, see Gardes et al (1996). 8 See Espey and Espey (2004).


18

The first interpretation concerns the ability of statistical methods in general to capture elas-ticities when there is little price variation; estimates simply become too low because the methods cannot distinguish the small variations from noise.9 From a very general point of view, econometrics is all about separating data signals from data noise, but this task becomes more and more difficult as the level of noise increases. When we give this interpretation of statistical difficulties, it will imply that the true elasticity values are about 0.7 lower in abso-lute terms than typical estimates (higher in numerical terms). The second interpretation concerns a change in attitudes and thereby in behaviour. The oil price hike brought another focus on energy savings and therefore consumers changed their behaviour. This interpretation is also backed up by a study from the US demonstrating that consumers reacted more strongly to energy prices when energy standards became obligatory thus allowing consumers to pick the products that were most energy efficient. 10 A third interpretation simply takes the result at face value and concludes that iso-elastic de-mand curves do not describe behaviour very well. Instead, elasticities are increasing in price levels.11 This is consistent with the assertion that elasticities are higher when energy costs are relatively high compared to output (GDP). Finally, we seem to find evidence that differences are substantial across countries. Some of this may be explained by economic and political structures,12 some of it by differences in technical structures. A good example where market design matters is the case of electricity re-tail prices. Historically, these have been determined ex post (in order to keep zero profits of regulated firms) in many countries. This completely eliminates any short/medium run re-sponses in demand. Moreover, some of the country differences can also be attributed to the level of attention / attitudes. Taken as a whole, the above discussion suggests that elasticities increase with price levels and with the general level of attention towards energy scarcity / environmental issues. Further-more, we can see that the rate structure can increase the long run elasticity by app. -0.5 for electricity when marginal (short run) rates vary with consumption. From a policy perspective the literature therefore provides three simple lessons:

(i) Energy tax policies work well when consumers are informed about products and alert about consequences.

9 It is standard knowledge in the econometric profession that low variation causes less precise estimates, c.f. Greene (2003). Moreover, low signal-to-noise ratios generally lead to downward biases of estimates, and low variation may cause simultaneity biases (relevant in non-system estimation of elasticities) to increase. 10 Gillingham et al (2006) 11 Bernstein and Griffin (2005), Espey and Espey (2004). 12 For example, France is known for a weak coupling between end user electricity prices, and wholesale market prices.


19

(ii) Energy tax policies work well when consumers are given the opportunity to re-act in the short run.

(iii) Stronger effect when overall energy costs are high. The first point suggests the use of information standards and campaigns,13 while the second suggests that policy makers should strive for transparent pricing systems. However, some-times politicians attempt to counteract price fluctuations by levying taxes and subsidies.14 In other words, the economic and informational context must be put in place.15 The third les-son is related to a specific market aspect, but it may suggest that taxation as an instrument has the advantage of increasing effectiveness as the level increases. However, the main finding here is that medium and long run elasticities seem to be eco-nomically significant. This strongly suggests that consumers and industries are willing to substitute towards green technologies, and this conclusion is extremely important for taxes to play a role in the deployment and development of low carbon technologies.

2.2. ENERGY PRICES AND TAXES EFFECT ON PENETRATION OF LOW CARBON TECHNOLOGIES

In this section we move the focus from pure consumption reductions by consumers and in-dustries to more fundamental changes in the way we use energy. Most of these effects con-cern medium run adjustments in capital holdings, and we will focus on three areas:

Renewable energy sources for electricity Energy efficient household appliances. Fuel efficient cars

Renewable energy sources for electricity A moderate carbon price will be sufficient to make profitable a substantial amount of non-fossil energy systems in the power system. Figure 2.5 shows the estimated global production costs of electricity for different renewable technologies compared to coal. It shows that with a carbon price of 30 euro / tonne, a range of renewable technologies come close to being competitive. That is, they have the same production costs as coal when including the 30 euro / tonne price. However, the figure also shows that without a carbon price, coal is gener-ally cheaper than most of the other existing technologies. While most of these technologies are today supported by direct subsidies, most estimates suggest that tightened climate poli-cies resulting in higher prices of ETS allowances would be sufficient to make substantially further amounts of renewable energy economically viable, c.f. Figure 2.5.

13 Campaigns serve to improve attention to environmental costs and to provide guidance to possible energy cutting. From a classical economics perspective, such campaigns are of little value since they provide no hard (financial) in-centives. Modern economic theory, however, acknowledges the 14 See Flood et al (2010). 15 See also Suslov (2008) for a general result on this point.


20

Figure 2.5: Projected cost ranges for power production across technologies, 2020-2030

Source: IEA, World Energy Outlook 2009 and CE calculations

Adoption of energy efficient household appliances The high average electricity price over the past years (2000-2004) seems to have led to con-sumption choices with a higher penetration of energy efficient appliances in the period of in-terest (2005).16 Hence in countries with high electricity taxes such as Denmark, The Nether-lands and Sweden, the share of highly efficient household appliances – those marked in en-ergy class A and A+ - is much more widespread.17 A more in-depth study for US showed that over a time span of 30 years, rising energy prices have led to innovation in energy con-suming household appliances, and that more energy efficient models were offered for sale (and actually sold). Here, it is also emphasised that additional information about energy use, directs consumer behaviour towards more energy efficient appliances.18 One should take this example to note the general point that the link between taxation and energy efficiency is in-directly through prices. Some countries may have high prices without having significant taxes and vice versa. However, the tax always adds to the price and therefore helps moving the adoption towards energy efficient appliances.

16 Conducting a simple regression between penetration and average electricity price results in a coefficient of 1.59 (i.e., a 1 $-cent/KWh increase leads to a 1.59 percent increase in penetration of energy efficient appliances) with a corresponding t-value of 2.16 being significant at the 5 percent level. 17 Bertoldi and Atanasiu(2007) 18 Newell et al (1999)

0 100 200 300 400

Coal with 30 euro …

Coal

Nuclear power

Large Hydro

Biomass for …

Onshore wind

Offshore wind

Biomass

Small hydro

CCS

Geothermal

Solar photovoltaic

Euro pr. MWh.


21

Figure 2.6: Penetration of energy efficient household appliances

Note: Electricity price including tax. We define energy efficient appliances as refrigerators, freezers, washing ma-chines, dishwashers and ovens rated A or better. The penetration rate shown in the graph is an average weighted with total sales.

Source: Copenhagen Economics based on Bertoldi, P. & Atanasiu, B., (2007) and IEA Energy price and tax data-base.

Fuel efficient cars The effects of energy prices and energy taxes on the fuel efficiency of cars deployed are evi-denced in a number of studies including our own simple graphical representation below in Figure 2.7. The difference in end user gasoline prices between countries seems to affect dif-ferences in average fuel efficiency. Higher gasoline prices provide incentives to improve the fuel efficiency of cars.19 The variations away from the trend line can, to a large extent, be ex-plained by historical and demographic differences. Considering the two largest EU Member States, Germany and France, it should be no surprise to find Germany above and France be-low the line. Germans have a tradition for larger and therefore less energy efficient cars, while the French typically tend to drive smaller cars. In Sweden, the low energy efficiency probably stems from the longer distances inducing higher driving comfort. Such conclusions are also reached in more elaborate modelling attempts.20 Other empirical analyses show that the fuel efficiency response to gasoline price are alike in different regions, whereas the final energy demand is influenced by different price elasticities when it comes to numbers of cars owned and demand for transportation.21

19 A simple linear regression with fuel efficiency as the dependent variable and gasoline price as regressor yields a t-value of -3.7 which is significant at the 1 percent level. This corresponds to a 2.5 liters/100 km increase in fuel effi-ciency if gasoline prices rose by 1 $/Ltr. 20 Eftec (2008). 21 Brons et al. (2006)

autbel

cze

dnk

deu

grc

espfra

irl

ita

lux

hun

nld

pol

prt

svkfin

gbr

rou

0.000

0.050

0.100

0.150

0.200

0.250

0% 10% 20% 30% 40% 50% 60% 70% 80% 90%

Penetration of energy efficient appliances 2005

Average electricity price (US$/KWh) 2000-2004


22

Figure 2.7: Gasoline price levels and fuel-efficiency; Average new gasoline LDV fuel effi-ciency and end-use gasoline prices, 200422

Source: IEA, Energy Technology Perspectives 2008 p. 436 and World Development Indicators, Table 3.12 We also attempt to show evidence that changes in gasoline prices over time within countries affect fuel efficiency. Below is shown the development of end-user price on gasoline for Germany, France, the UK, and the US compared to the development of fuel-efficiency on new gasoline LDV’s in the corresponding countries. The data only allows us to cover the pe-riod from 1980 onward. For the UK there is a sharp rise in fuel efficiency up to 1985, where it stabilizes in a period with falling and low gasoline prices. As the gasoline prices start rising again from 1991 onward, there is a following increase in the fuel efficiency, starting with some years lag. The delay can be explained by consumers waiting to see if a new higher gasoline price is just temporary: a more energy efficient, but perhaps also more costly and less attractive car in terms of performance, will only be chosen if the price change turns out to be of a more permanent character which may take some years to establish.

22 The above figure lists the fuel price for gasoline in the respective country vs. the average fuel efficiency for new gasoline driven cars in the respective countries. For Australia and Japan the average fuel efficiency is based on both diesel and gasoline cats, due to lack of data on the split between gasoline and diesel cars in the two countries.

United States

Australia

Canada

SwedenGermany

NetherlandsUnited Kingdom

Japan

FranceItaly

R² = 0.6688

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

5 6 7 8 9 10 11

Price ($/Ltr)

Litres pr. 100 km


23

Figure 2.8 UK gasoline prices and fuel efficiency for new gasoline LDV’s The UK Germany

France The US

Source: IEA data services for gasoline prices and IEA data from MoMo-model, upon written request A similar picture is seen below for France, though consumer behaviour tends to react a bit faster. For Germany the effect from gasoline prices on fuel efficiency is more moderate. This can be the result of the above mentioned fact that the Germans tend to drive larger and less energy-efficient cars possibly resulting from preferences for comfort and car size. US data is available for a longer period, and we see rising fuel efficiency up until 1985, as a lagged response to the second oil crisis. The new regime of falling oil prices are transmitted into a period with zero growth in fuel efficiency, then to recently beginning to rise as a re-sponse to the growing gasoline prices since 1999. A variety of literature studies examines the changes in gasoline prices over time and the effect on consumer choices. In a recent study for the US market, the market share of the 25 per cent most fuel efficient cars is estimated to rise by 20 percent in response to a 1 USD rise in gasoline prices.23 Moreover, the share of larger SUV’s is also estimated to be affected by the gasoline prices in the sense that the average fuel efficiency of new sold cars rises with ca. 0,2 – 0,4 km. / litres with a 1 USD rise in gasoline price.24 A study from the UK Department of Transport finds a fuel-price elasticity to the new car fuel efficiency of 0.2 %.25

23 Busse et al.(2009) 24 Klier(2008) 25 Eftec(2008)

60

65

70

75

80

85

90

95

100

105

110

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

200

0

200

2

200

4

200

6

Gasoline price, 2000 = 100

Fuel efficiency, (km. / liter) 2000 = 100

60

70

80

90

100

110

120

130

140

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

200

0

200

2

200

4

200

6



60

65

70

75

80

85

90

95

100

105

110

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

200

0

200

2

200

4

200

6



020406080

100120140160180200

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

200

0

200

2

200

4

200

6

200

8

USA, gasoline prices, 2000 = 100



24

2.3. HIGHER TAXES LEAD TO GREEN TECHNOLOGY INNOVATIONS Having established that price and tax increases can be highly effective in reducing energy demand and pushing demand in the direction of energy efficient variants, we are ready to discuss the long run implications for innovation in such technologies. As pointed out above, the presence of these behavioural changes is a prerequisite for the existence of a market for green innovations. In fact, the more expensive we make fossil fuel inputs, energy consump-tion, emissions and other forms of pollution, and the larger behavioural responses (demand and substitution elasticities), the greater incentives we find for investing in R&D. In other words, higher energy prices change the relative returns from longer term investment options to the benefit of energy and GHG displacing technologies.

Powerful effect on R&D spending and green technology diffusion Studies on the effect of policy or prices on innovation draw their motivation from the notion of induced innovation (or directed innovation), which recognises that R&D is a profit-motivated investment activity and that the direction of innovation likely responds positively in the direction of increased prices.26 Empirical studies on the effect of policy and prices on environmental innovation both support the conjectures of the induced innovation hypothe-sis and provide evidence of the magnitude of these. One branch of such empirical studies uses simulation models to assess the magnitude of in-duced innovation.27 A recent study carried out by the OECD28 finds that current and future expected carbon prices appear to have powerful effects on R&D spending and clean tech-nology diffusion. The study assumes a global carbon price reflecting the CO2 emission tra-jectories necessary to keep temperature increases below 2˚ Celsius. Under this scenario new technologies will contribute with ca. 50 percent decarbonisation where current rates are ca. 35 percent. These calculations are based on a detailed description of the energy sector (bot-tom-up) and the carbon markets combined with a general description of the global economy (top-down, CGE). A second branch applies econometric techniques to historical data and in this way attempts to assess the linkage between pollution prices (either in terms of energy input, consumption or emission prices) and targeted R&D performance.29 This is the approach we take in the empirical section below. The typical measure of R&D performance is the number (or num-ber of citations) of patents within a technology class, and the vast majority of studies either concern the US or Europe within the last 30-40 years. Early studies used pollution abate-

26 Hicks 1932, Binswanger and Ruttan 1978, Acemoglu 2002 27 E.g. Popp (2006), Acemoglu et al (2009), Fisher and Newel (2008). 28 OECD (2009), ”The role of R&D and technology diffusion in climate change mitigation: new perspectives using the WITCH model,” Working Papers No. 664. 29 E.g., Popp (2002), Brunnermeier and Cohen (2003), Hamamoto (2006), Johnstone et al (2009).


25

ment control expenditures (PACE) to proxy for environmental regulatory stringency since environmental taxes have typically been lower than such costs.30 In fact, recent studies have shown robust effects from energy prices on patenting on energy technologies.31 An important methodological advance in achieving this result concerns the use of disaggregated patent counts as it allows targeting the empirical analysis directly on relevant energy related patents.32 The general result from analyses exploiting disaggregated patent data is that induced innovation materialises quite clearly in patenting activities. This holds for several measures of environmental policies: PACE33, regulation34, and energy prices (taxes)35. For example, an increase in compliance expenditures of 1 percent typically lead to increases in R&D expenditure of 0.2 percent across countries and industries. This holds for any kind of instrument applied: be it taxes, allowances, or command-and-control. The esti-mate on energy prices is also economically and statistically significant. A one percentage in-crease in energy prices implies an approx. 0.4 percentage increase in energy technology pat-enting. The estimations also suggest that non-market policies are less effective; technology standards (command-and-control) create fewer patents with less environmental impact.36 The literature has acknowledged the presence of time lags in induced innovation.37 There are several reasons for time lags. First, firms must wait for clear price signals before they allocate resources to R&D – and in the case of energy prices, we typically see significant short run volatility blurring the price signal. Second, institutional (within-firm) barriers may cause slow adjustment. Along the same lines, research personnel may need some time to adapt to a changed focus. Third, patents are created on the basis of an innovation process which in it-self takes some time. The empirical literature typically comes up with a half-life of induced innovation in the ballpark 3-5 years, i.e., half of the induced innovations have occurred 3-5 years after the price/tax increase.38 Fourth, patents need to be converted into deployment of new technologies before having real effects on energy consumption. A deeper reading of this literature provides some further insights into the nature and neces-sary preconditions of innovation impacts. There is unfortunately very little to be said directly on how environmental taxes differ in impact from overall prices and/or other instruments. It is noteworthy that time and geography seem to influence the level of the impact, but in our context it is particularly interesting to look at policy conditions that favour innovation im-pacts. We find an overview of this in Table 2.3.

30 For example Jaffe and Palmer (1997) and Brunnermeier and Cohen (2003). 31 Popp (2006), Johnstone et al (2009). 32 Lanjouw and Mody (1996), Popp (2006) 33 Lanjouw and Mody (1996). 34 Popp (2006). 35 Newell et al (1999), Popp (2002), Johnstone et al (2009) (a). 36 Johnstone et al (2009). 37 See, e.g., Peeters and Surry (2000). 38 Johnstone et al (2009) (a), Popp (2010).


26

Table 2.3: Factors influencing the innovation impacts Factor influencing impact Effect Source Policies and instruments Domestic policies + Popp (2006) Information to consumers (labelling) + Newell et al (1999) Geography/economic development Developed countries + Lanjouw and Moody (1996) Developing countries - Lanjouw and Moody (1996) Time Recent years + Brunnemeier and Cohen (2003) Source: Copenhagen Economics The table demonstrates how a specific factor, say predictability and stability of policies, in-fluences the level of induced innovation. A “+” implies that the factor increases the innova-tion impact from prices and/or policies. Thus, the table conveys that predictable and stable policies induce more innovation than more ad-hoc policy solutions. Note that we have already touched upon the issue of labelling and consumer information when discussing elasticities above. In that (much broader) literature it was a standard finding that labelling and information increased responsiveness. In the case of induced innovation, according to the source references in the table above, domestic policies (a policy category used by Popp mainly including information campaigns) and labelling yield the same result. As the innovation impacts should work ‘through’ elasticities, it is comforting that we can point to some of the same factors. The table also demonstrates some differences over countries and time. The basic message from this part is that the more advanced level of development of the economies (looking along both the time and cross country dimension) leads to more responsive patenting – probably because general production of innovation is more important and more rent-seeking. This observation, however, is not very relevant from a policy perspective.

Technology and innovation decisions have a global scope In the above discussion of demand responses and innovation impacts, we have been rather vague in our definition of the relevant market. In the following paragraphs, we attempt to clarify the role of global markets – both concerning the medium run decisions on technology choice and long run decisions on induced innovation. The decision to change technology by profit seeking firms is based on a cost benefit analysis. As shown above, green technologies are typically more expensive up-front, but excel in their lower running costs. However, there may be other possibilities than changing technology. Firms may simply change location when environmental policies become too rigid in a spe-cific country. In the worst case, an environmental policy can trigger an outsourcing of energy intensive processes to regions with low or no taxes, but deploying highly inefficient tech-nologies. Thus, global emissions will increase and the policy will be counterproductive.39

39 There is a long literature on so called carbon leakage. Some of the most interesting references are Babiker (2005), Szabo et al(2006), Demailly and Quirion (2006), and Barker et al (2007).


27

Leakage will not arise mainly as a competition advantage of countries not participating in the emission reductions, but as an input price effect. Lower world demand for fossil fuels will lower prices and increase fossil-fuel consumption in these countries. Consumer markets and industry processes closest to the end market are less vulnerable. The same goes for prod-ucts that are costly to transport.40 There is a mirror image to leakage in production for ‘leakage’ in innovation. The point is that for geographically narrow defined taxes there is little incentive to engage in development projects that would lead to new technologies cutting these taxes back. Important innovations always have a market perspective that is broader than narrow national markets in order to repay the research investments. Formally, innovation has at least two global dimensions.41 First, innovation exploits global opportunities as just outlined. Second, geographically diversified innovation reduces invest-ment risks. Both dimensions clearly suggest that local taxes will only have limited impact on innovation. A study finds that EU27 is geographically insufficient to avoid carbon leakage for environ-mentally uncompromising tax policies.42 Thus, from a European Union point of view, it seems natural to mitigate carbon leakage by, at least, including all Member States under the allowance trading schemes and unifying current national energy tax schemes. The study clearly suggests that policy makers should opt for widest possible coverage – even if the out-come of the agreements may seem second-best compared to what is attainable. This is not to say that European taxes are inefficient, but simply that they are much more efficient if they are backed up on the international scene too. That implies as well that global carbon pricing has stronger effect on EU innovation than EU carbon pricing alone.

2.4. NEW EMPIRICAL RESULTS ON INNOVATION IMPACTS This section is devoted to the empirical analysis using European patent data to investigate the relation between price signals and induced innovation. The overall conclusion is that en-ergy price increases do induce more innovation. However, before turning to this result, we will first describe the technologies considered and describe their relevance as energy consum-ers. After a review of the main results from the empirical model, we will use graphical tools to illustrate some important points related to the results and applied methodology.

40 OECD (2008). 41 Hitt et al (1994). 42 OECD (2008), see also Bosetti et al (2009b).


28

Energy technologies A number of technologies exist which in some way relate to energy consumption. Here, we choose to look at seven specific technologies:43

Biomass for heating in buildings Boilers Ventilation in buildings Lighting Light emitting diodes (LED) Motor vehicle fuel efficiency Paper and pulp production

These are all characterised by a close link between energy consumption and the final output, e.g., oil consumed by a boiler to produce heated rooms. However, these technologies also possess characteristics distinguishing them from one another. We expect the different technologies to take different paths in their response to changes in energy prices. Cars, biomass in buildings, boilers and ventilation in buildings are characte-rized by high investment prices for the end-user. The percentage reductions in energy use may not be large, but the effect on household budgets could be significant as the cost of energy in these products is substantial. Lighting and LED technologies represent a small in-vestment for the end user, but with very large, percentage-wise, reductions in energy use. The energy cost of these products is typically small in terms of household budgets. Finally, the paper and pulp industry represents a class of its own, where innovations in energy effi-ciency lower the cost of producing a product that for the end user is unchanged.

43 The exact definition with respect to extraction from the EPO database is described in appendix C.


29

Figure 2.9: Energy budget (cost) shares for different processes

Note: For lighting, heating, ventilation, and motor vehicles, we present the budget share of total household energy costs. For paper and pulp we present the budget share of total production costs.

Source: Copenhagen Economics based on IEA, Eurostat, and specific studies The above differences in the technologies’ properties will imply that the innovation response to changes in energy prices will differ among the products. For biomass, boilers, ventilation and motor vehicles the penetration rate of the energy effi-cient technologies will be driven by longer term changes in energy prices, as these products

0

10

20

30

40

50

60

Lighting (Lighting and LED) Heating and ventilation Motor vehicles

Share of household energy

consumption

0

10

20

30

40

50

60

Paper and pulp

Energy costs share of total

production costs


30

have a lifespan of 5+ years44, and that their cost structure suggests that consumer choice of technology will be made upon time of replacement. Even though the effect may take time, changes in energy prices should have a high effect on consumer choices, as these products have a high energy cost as share of the household budget, so energy costs play a relatively high role compared to the agents other preferences for the products. The opposite holds for lightning and LED, partly because they exhibit shorter life spans, partly due to the fact that energy costs of these products play a minor part in the household’s budget. To a large extent, consumer taste/preferences will be determining for lighting source decisions. The paper and pulp industry’s choice of technology is a long term decision, where choices of production methods represent large costs and 10+ year lifespan. As paper and pulp producers should aim for profit maximization, their investment in energy efficient technologies will be solely driven of their own long term projection of energy prices and the cost of the actual in-vestment. This suggests that the adaption of energy efficient technology in the paper and pulp industry to a high degree will be driven by energy prices, but with a significant time lag.

Patents, taxes and energy prices In this section, we present the results from our econometric investigations on the link be-tween energy prices, taxes, and patents. Here we focus on the general messages, while a more thorough, methodological treatment is given in Appendix A. Looking across the results from different technologies, we find quite clear patterns, c.f. Table 2.4. The table demonstrates the sign and statistical significance of estimated coefficients across different estimations. Thus, a “+” refers to moderate evidence of a positive relation, whereas the double “++” implies highly significant estimates of a positive relation in most of the estimations.45 Similarly, “-“ and “- -“ will denote negative relations, while a “0” indicates a lack of statistical significance. For example, for ventilation technologies we see that public R&D together with (long run) taxes imply a highly significant positive impact on patenting activity.

44 Eurostat 45 There is no mathematical rule for the sign/significance assignment in the table. We start by classifying the regres-sions as to whether they produce meaningful results or not. Only thereafter, we compare the evidence across the remaining specifications.


31

Table 2.4: Sign and statistical significance of estimates Technology Public

R&D Legislation Patent

trend Price (long run)

Tax (long run)

Lighting + ++ ++ 0 +

LED + 0 0 0 0

Biomass 0 NA 0 ++ ++

Boilers 0 NA - 0 ++

Ventilation ++ 0 0 0 ++

Motor vehicles + NA 0 0 ++

Paper and pulp NA NA + ++ ++

Note: Signs represent the statistical size of impact across models. Roughly, a single sign represents cases with a coef-ficient being weakly significant in a few model estimations, whereas a double sign requires higher signifi-cance and agreement across estimations.

Source: Copenhagen Economics We complement this table with estimates from the estimation we prefer the most for each technology, c.f. Table 2.5. However, due to econometric issues, we are reluctant to put too much emphasis on a single specification and instead prefer to look at general patterns across estimations. One should therefore also be very careful when comparing the two tables. There are two specific caveats. One specific caveat is that the tax estimates comes on top of their implicit price impacts. Thus, if the price effect is significant, then the tax is also significant by default, but may have an additional effect which is tested by the specific tax coefficient. Thus, in Table 2.4 we can characterise, say, the long run tax impact as highly positive (“++”) despite the additional tax effect being slightly negative as long as the total long run price estimate is positive and sig-nificant. The other caveat concerns the difference between short and long run effects. A short run ef-fect is estimated directly as a parameter in the econometric specification, while the long run effect must be calculated based on the dynamic structure of the specification. For example, if the model is given by

then the short run effect is given by β, i.e., the immediate impact from last year’s tax in-crease, while the long run effect will be calculated as β/(1-α), i.e., the accumulated effect over a number of years (approximated by the infinite horizon).


32

Table 2.5: Selected coefficient estimates Technology Public

R&D Legislation Patent

trend Price short run

Price long run

Tax short run

Tax long run

Lighting -0.163 0.251*** 0.425** -0.163 -0.579 1.526* 0.496

LED 0.062 -0.041 NA 0.493 -0.636 -0.174 0.416

Biomass -0.020 NA 0.555 0.719*** 0.283** -1.168 1.074

Boilers 0.000 NA -0.348 -0.140 0.227 0.951*** 2.332***

Ventilation 0.130*** 0.111 -0.336 -0.314 -0.531 0.718 2.369***

Motor vehicles 0.046 NA NA 0.392 0.081 1.110 2.192**

Paper and pulp NA NA 0.165 0.188 0.467*** -0.851 -0.288

Note: * implies significance at the 10 pct level, ** at the 5 pct level, and *** at the 1 pct level Source: Copenhagen Economics From the above tables there seems to be three main messages. First, environmental taxes have much stronger and more lasting effects on patenting than short run price movements. This is evident from comparing the last two columns of Table 2.4. We attribute the credibil-ity and transparency of taxes compared to the volatility of energy prices to this difference.46 Second, public ‘institutions’ (such as public R&D and legislation) surrounding the private innovation environments do contribute positively to patenting. This is evident from looking at the first two columns. Third, technologies such as lighting and LED with small energy budget shares do not seem to experience much price/tax induced innovation. This is the message from the first two rows of the table. The first observation concerning the role of taxes for innovation leads to the question of the economic importance, i.e., to the size of the coefficient estimates. We illustrate the sizes of short and long run impacts in Figure 2.11. For the majority of industries, we observe a much higher impact from taxes than from prices. Long run elasticities are in the range 2-2.5, c.f. Figure 2.10, which is approximately five times larger than the estimated price impacts, see Figure 2.11 below. In both cases, one should not forget that we have focussed on significant – and thereby typically the largest - es-timates, but still we believe that there is strong evidence of very significant impacts.

46 In our data set, the tax variable is approximately 40-60 percent less volatile measured by (comparable) standard errors.


33

Figure 2.10: Innovation impacts from environmental taxes, significant estimates

Note: The significance criteria are consistent with the sign criteria in the previous table. Source: Copenhagen Economics There are two sectors where the prices (including taxes) turn out to be significant drivers of innovation; biomass in buildings and the paper and pulp industry. In both cases, the elastic-ity estimates of around 0.4 are very much in line with the results from the rest of the litera-ture (in the previous section we also reported 0.4 as the average estimate). Again, this implies that a 1 percent increase in energy prices leads to a 0.4 percent increase in patents. The esti-mates on biomass in buildings are characterised by a high short run impact and a more mod-erate long run impact (the latter being more in line with the literature.) However, the stan-dard error attached to the short run estimate is much higher, so the more plausible ranking of low short run and high long run is still within reasonable confidence intervals. It seems natural that the effect in the paper and pulp industry comes directly from prices as environ-mental taxes on industrial processes historically have been zero or very close to zero.

0

0.5

1

1.5

2

2.5

Lighting LED Biomass Boilers Ventilation Motor vehicles

Paper and pulp

Short run tax Long run tax


34

Figure 2.11: Innovation impacts from energy prices, significant estimates

Note: The significance criteria are consistent with the sign criteria in the previous table. Source: Copenhagen Economics The fact that innovation responds considerably more to taxes than to prices may have several explanations. The first explanation concerns the stability and credibility of taxes (we will also return to this point later when we discuss policy mix in chapter 4). Tax rates are imple-mented with long horizons and are seldom reset to lower levels. Thus, they send an unambi-guous signal to innovators that prices will remain at higher levels. In contrast, price changes may be caused by short run expectations, economic upturns etc. which does not provide long term incentives for R&D investments. Other types of explanations concern the econometric methodology and statistical behaviour, i.e. they imply that estimations contain some bias. The econometric methodology works un-der functional form approximations, c.f. the discussion in the appendix. These approxima-tions may cause some bias – especially for fuels where taxes form a relatively large part of end user costs, e.g., gasoline for motor vehicles. Moreover, the fact that tax rates change rather infrequently and typically in a unidirectional way suggests an easier identification of long run impacts for the class of estimators. Energy price movements are frequent and not unidi-rectional thereby giving rise to much short term noise easily influencing the econometric es-timates. Note that this latter explanation implies that price elasticities are estimated with a negative bias (they are too low compared to true effects), rather than tax estimates being too high. Looking at the above results from a slightly different perspective, we have three different classes of estimates/technologies. For lighting and LED technologies, we do not find evi-dence of the induced innovation hypothesis; neither from prices nor from taxes. In a second class, we find typical consumer good technologies (boilers, ventilation, and motor vehicles) where price signals play a minor role compared to tax counterparts. Estimates are both eco-nomically and statistically significant. Finally, biomass in buildings together with paper and

0

0.2

0.4

0.6

0.8

1

1.2

Lighting LED Biomass Boilers Ventilation Motor vehicles

Paper and pulp

Short run price Long run price


35

pulp technologies follow the induced innovation hypothesis directly from prices although at somewhat lower response estimates. A priori, we would assume that the biomass in buildings belonged to the second class (together with the related technologies boilers and ventilation), but this was not confirmed by the estimations. A second output of the estimations relates the time lags in the induced innovation hypothe-sis, c.f. the discussion above. Typical estimates of half lives were 3-5 years, where a half life is the time it takes from the price/tax signal to half of the induced patents to be filed. Thus, the first half of patents appears within 3-5 years, while the other half takes longer to materialise. Overall, we confirm this range despite most technologies falling in the lower end of this range, c.f. Figure 2.12.47 Common for both our estimates and those found in the literature is a large uncertainty attached to the estimates. The primary cause for this uncertainty is related to the methodological approach where the parameters capturing the dynamics of patenting almost completely govern the half life estimate. Despite the considerable uncertainty, we still believe that the estimate seems economically reasonable and intuitive. Figure 2.12: Half life estimates

Note: Half times are based on lag structure and estimated dynamic multipliers. Source: Copenhagen Economics

Graphical inspection of patent data and relationship between innovation and prices In this section we take a less sophisticated look at the data. Specifically, we use graphical techniques to investigate patents and energy prices and taxes. Due to the amount of data (30 years, 30 countries, 13 energy products, and 7 technologies) we focus on a few (more or less randomly) selected examples.

47 We have included half life estimates for lighting and LED even though the induced innovation hypothesis was not confirmed. Technically, half lives can be retrieved for statistically insignificant price/tax estimates without any problem.

0 1 2 3 4 5

Lighting

LED

Biomass

Boilers

Ventilation

Motor vehicles

Paper and pulp

years


36

The specific purpose of this section is to provide the reader with a flavour of the underlying relations between prices and patents and to introduce some of the econometric issues in the empirical analysis. Anticipating the conclusions from the latter issue, i.e., what a good em-pirical model will need to capture, we can state the main elements as follows:

Dynamic effects of patents. We show that a change in patenting activity created dynamic reactions several years after the initial change.

Common trends unrelated to energy prices. A significant portion of the trend throughout time seems to relate to changes in industry structure, developments of global markets, etc.

Country differences. Specifically, the model needs to account for the fact that large countries produce more patents than small countries.

Fuel / energy source. Several technologies can draw on different energy sources and we need the right match between patents and energy price.

Taxes are (sometimes) relevant as a separate component. We identified some cases where taxes constitute a large proportion of the end user price.

In the paragraphs below, we will demonstrate why we derived these conclusions. We will not go into further technical details of the applied methodology here, and we refer the interested reader to Appendix A. Patents We choose to let the number of patents within a certain technology class be a measure of in-novation. It is undisputed that a substantial part of private R&D is made with the purpose of issuing patents and subsequently earning economic rents on these patents. As such, the measure should be a valid indicator of private R&D output. In the literature, patents are clearly also the favoured measure of innovation, cf. our discussion above on innovation im-pacts. Yet, the literature also points to drawbacks and limitations of this measure. In the fol-lowing, we sum up the most important criticisms:

Patents are specific to some physical aspect of the world, not to working or organ-isational processes. In this way, we do not capture many relevant progresses in, say, the organisation of industrial production processes.

Patents may differ quite a lot in their quality/relevance. A good patent can contain more ‘innovation’ than hundreds of inferior patents.48

Important innovations may never be patented. Rather, the inventor may want to keep his invention secret or exploit the commercial value without a protecting pat-ent for various reasons.49

48 Popp (2006). 49 Levin et al (1987).


37

Leaving aside these concerns, we can take a look at patent characteristics. Within energy-related technology groups, patenting is most substantial for buildings and energy generation, c.f. Figure 2.13. We also observe that carbon storage together with paper making technolo-gies seem to trend in slightly different ways than the other technology classes. Note that we do not consider any technologies within carbon storage in the econometric work. Figure 2.13: Number of patents across technology classes, OECD countries

Note: The vertical axis measures the total number of patents Source: Copenhagen Economics based on EPO To obtain an impression of the development at country level, we illustrate the level of pat-enting across selected countries in Figure 2.14. We see that countries have quite different levels of patenting, e.g., Japan is a clear outlier due to formal reasons in the patenting system. However, also European countries show differences which are not merely related to the size of the economies. For example, Finland and Sweden produce marginally more patents than larger countries such as Italy and Spain. Thus, the suggested panel data approach controlling for country fixed effects seems reasonable.

0

5,000

10,000

15,000

20,000

25,000

30,000

35,000

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

Buildings

Energy Generation

Paper Making

Transport


38

Figure 2.14: Patents across countries

Note: The vertical axis measures the total number of patents Source: Copenhagen Economics based on EPO An interesting characteristic about patents concerns their dynamic behaviour. There are good economic reasons why patenting in one year is closely related to patenting in the previ-ous year. First, company level choices of investing in R&D are hardly ever a year by year dis-crete choice. Instead, money flows to R&D departments often have longer horizons which create dependence between years not only on the input side, but also on the output side. Second, patents will often feed into new patents within the same class. When researchers come up with important technical advances, these will spread out to all related fields in the next years. In Figure 2.15 we demonstrate the dynamic patterns. The figure shows the patenting activity for six EU Member States year by year as the relative activity compared to the country’s av-erage activity over the entire 30 years. (That is, a value of 10 percent in a given year tells us that patenting was 10 percent above the country’s average patenting level.) In this way, series that seldom cross the 0-line display significant dynamics; a high level in one period carries over to the next period and similarly for low levels.50 The figure clearly demonstrates that there are rather few intersections over time. Furthermore, the figure demonstrates a signifi-cant increase in volatility around 2002. One possible explanation is the political attention devoted to energy efficiency and bans of light bulbs over this period starting with Directive 2002/91/EC. This may have triggered some patenting; both patents that were simply wait-ing for the right moment to apply and patents following up on specific legal requirements. To the extent that on-the-shelf patents were issued, this would explain much of the volatility increase.

50 In a 30 years series completely without dynamics, we would in average see almost 15 crossings.

0

10,000

20,000

30,000

40,000

50,000

60,000

70,000

80,000

90,000

100,000

JP CN DE

US FR GB

RU FI CA AT

AU SE NL IT ES CH PL BE DK CZ HU

NO GR PT

457,664


39

Figure 2.15: Patenting dynamics for six EU MS, 1978-2007, lighting technologies

Note: The figure shows the number of patents relative to the country average over time. Source: Copenhagen Economics based on EPO Patents, energy prices and taxes The purpose of the econometric study is to test the impact from energy prices (and taxes) on patenting activity within various technology classes. In this section, we take a short look at this relation. In Figure 2.16 we see the development in transportation patents for four se-lected countries (all being motor vehicle producers) together with oil prices. We see that pat-ents often, though not always, covariate with prices. In particular, it is interesting to observe the differences in behaviour around the second oil crisis (1979-1980). For all four countries we see a clear upward trend in oil prices, with the high price remaining a few years into the 1980s. When it comes to patenting, however, German car producers hardly show any response to the new situation, whereas Japanese car manufacturers seem to respond quite fast. (UK and France do moderately increase patents for a short period.) The most straightforward interpretation of this difference relates to the different market segments. To put it bluntly, German cars were aiming at customers looking for comfort and robustness, while Japanese cars were sold to customers focusing on the price-quality dimension. Hence, Japanese car producers had to respond to higher gasoline prices to offer a cost-efficient product. Obviously, there may be several other factors explain-ing this difference, e.g., differences in expectations towards national policies.

‐150%

‐100%

‐50%

0%

50%

100%

150%

200%

250%

300%

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

AT BE ES FR GB IT


40

Figure 2.16: Refined oil end user prices and patenting in motor vehicles, four large car-producers

Note: The left vertical axis displays dollar prices of oil. The right vertical axis displays the number of patents. Source: Copenhagen Economics based on EPO and IEA data. The car example can also demonstrate two other important points complicating the analysis between prices and innovation. The first point concerns the composition of patents. Even in the German case without much overall response to the second oil price crisis, there could still be a significant change in the number of patents directly related to fuel cost savings at the cost of other patents. We simply do not know.51 The second point is related to the choice of fuel price. In the figures above, we plot patenting activity against the price of oil , but cars consume gasoline or automotive diesel, not crude oil. And in the case of gasoline, the stan-dard type has changed from leaded to unleaded over the period of interest. The choice of the relevant fuel becomes particularly relevant when taxes start playing a more pronounced role. Turning to another example, we can look at the paper and pulp industry where the energy costs are incurred during the production of the goods, not during deployment. Figure 2.17 provides a simple cross-plot of patents and the relevant industry electricity price. Obviously, such a simple representation must not be taken as evidence for a strong positive relation be-tween prices and innovation. Yet, it does provide a first impression of the basic correlation to be more precisely determined in the econometric analysis. Moreover, Figure 2.17 demonstrates some of the challenges in working with a panel consist-ing of both a country and a time dimension. The figure “identifies” two paths in patenting: One path which clearly responds to underlying prices and another which is completely im-

51 Or to put it more precisely: it would require a case-by-case examination of all patents together with some scoring technique.

0

100

200

300

400

500

600

0100200300400500600700800900

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

Num

ber o

f paten

ts

US‐$ in 200

5 PP

P

Germany

oil price transportation patents

01020304050607080

0100200300400500600700800900

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

Num

ber o

f paten

ts

US‐$ in 200

5 PP

P

France


0

10

20

30

40

50

60

0100200300400500600700800900

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

Num

ber o

f paten

ts

US‐$ in 200

5 PP

P

UK


0

200

400

600

800

1000

1200

1400

0

200

400

600

800

1000

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

Num

ber o

f paten

ts

US‐$ in 200

5 PP

P

Japan



41

mune to price changes. A good econometric model will need to control for these two paths (much is already done by introducing country fixed effects thereby measuring characteristics such is size and industry starting point.) Figure 2.17: Electricity prices and patenting in the paper and pulp industry

Note: The figure is based on yearly data points for all countries in the sample. Source: Copenhagen Economics based on EPO and IEA statistics Finally, in specifically assessing the relationship between innovation and taxes, one must first ask for the historical importance of taxes. As shown in Figure 2.18, taxes have played a mi-nor role for most energy products, except gasoline and automotive diesel. In other cases, it is questionable how much taxes can contribute to the trends in patenting. However, the figure hides the fact that the data may contain significant cross-country and time series variation even when the average level is low. Furthermore, as explained elsewhere in this report, there are good reasons why tax changes may create larger innovation effects than similar price changes (due to non-tax changes.) In fact, part of the estimation strategy is to estimate both in levels and in differences in order to address this issue.

0

200

400

600

800

1000

1200

1400

0 0.05 0.1 0.15 0.2 0.25

Num

ber o

f paten

ts per year

Electricity price, industry, PPP US‐$ per KWh


42

Figure 2.18: Tax shares for different fuels, OECD averages 1978-2007

Note: The tax share is calculated as the $ amount of taxes relative to the final end user price (in $). Averages are simple, not weighted.

Source: Copenhagen Economics based on IEA

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%Tax share of end

user price, average

197

8‐20

07

Household

Industry


43

The previous chapter substantiated that energy and CO2 taxes are likely to have considerable effects on the innovation of energy technologies; nevertheless more direct public support for climate friendly technologies is also needed. This chapter is divided into four sections. The first section 3.1 describes the individual tasks to be fulfilled by taxation on the one hand and R&D support on the other, i.e., what is the division of labour? The second section 3.2 then takes a closer look at the interaction between the two instruments. That is, we address the question: to what extent in particular are the benefits of energy related public RD support depending on the tax environment? The third section 3.3 reviews how the mix between tax and R&D depends on the stringency and time horizon of policy ambitions. Finally, the fourth section 3.4 reviews some empirical studies on optimal policy mix between energy taxes and RD policies in delivering on global climate policy objectives. Throughout the chapter we will repeatedly draw on results from a simple simulation model, the CERIM (Copenhagen Economics Renewables Innovation Model) in order to illustrate important points. An introduction to the CERIM is given in Box 3.1 below, while a formal model description can be found in Appendix B.

3.1. DIFFERENT ROLES FOR DIFFERENT INSTRUMENTS In this section we address the different roles economic theory attaches to environmental taxation and R&D support. We will focus on the following four issues:

The dual externality problem The long term credibility problem R&D support and crowding out The effect on innovation played by non market-based mechanisms

Chapter 3 CLARIFYING THE ROLE OF TAXATION VIS-A-VIS DIRECT

INNOVATION POLICIES


44

Box 3.1. Introduction to the CERIM Copenhagen Economics Renewables Innovation Model is based on the formulation in Fisher and Newell (2008). The focus is on the US electricity sector, and the model divides production in four technologies: nu-clear, coal, natural gas, and renewables. Since the establishment of nuclear plants is more closely linked to political processes than market signals (and since it is non-emitting), we assume an exogenous path for this technology. The model contains two periods each containing several years. The first period starts from the current situation (parameter estimates from around 2008) and covers n1 years while the second period starts immediately after (and contains n2 years.) The renewable sector has the potential for technological advance depending on how much resources are al-located to R&D in the first period. The benefits, in terms of lower renewable production costs, are reaped in the second period. The policy maker has two instruments at his disposal: a carbon emission tax and an R&D subsidy. In this way, he can influence the production and emission path of the electricity sector by setting taxes and subsidies in both the current and in the future period. The model is based on firm profit maximisation and consumer utility maximisation. The outputs of the models are multiple: carbon emissions, renewable shares, welfare, etc.

Source: Copenhagen Economics

The double externality problem Up to this point, we have only addressed the problem of missing prices for pollution. As should be clear by now, environmental taxes will – at least partially – be able to solve this problem. But when it comes to innovations of green technologies there is an additional problem. A new innovation may create positive spillovers to other firms and the rest of the economy since innovations can be improved, standardized and create the basis for new technology classes. But these positive effects, which may exceed the direct profit creating effects to the company by several factors, are not fully appropriated by the company financing the re-search. Thus, the investments in company R&D are not sufficient when compared to the societal gains they create. In other words, we again see a missing price (payment in this case) for the effects created by an economic activity, R&D52. Below, in Figure 3.1, we illustrate the double externality problem graphically. In the left panel, we see the environmental externality as a difference between social and private mar-ginal cost curves (SMC and PMC respectively) – social marginal costs include the cost of climate change and other pollution damages and are therefore higher. A good environmental tax would add to private marginal costs so that these become exactly aligned with the social counterpart. That is, the tax should equal the vertical distance between the two cost curves; this distance is denoted by ‘A’ leading also to a reduction in emissions.

52 “All private sector innovation suffers from market failures. These are even more acute in the case of climate change, as environmental market failures compound the problem. Thus, policy plays a key role in shaping both the direction and magnitude of climate-friendly technological change”, Popp (2010).


45

Figure 3.1: Representation of double externality problem

CO2 emissions Research market

Sustainable consumption

SMC

PMC

Unsustainableconsumption

SMU

PMU

Private R&D Optimal R&D

A

B

C

Source: Copenhagen Economics The right panel illustrates the knowledge spillovers in the market for research. We assume as standard that return for research in energy research shows declining returns, hence a falling demand curve, while costs of supplying (quality adjusted) research is increasing, hence an upward sloping supply curve. Here, it is the private value of R&D which is below the socie-tal value, since private companies do not care about the positive spill over effects from R&D. This means that the private demand (PMU) lies below the societal optimal demand (SMU). Now, given that we impose a tax on pollution, we will raise the private value of innovation in pollution-reducing technologies, and we illustrate this shift with the dashed line in the middle of the panel. The upward shift in demand for research amounts to the distance de-noted ‘B’ in the figure. However, there is still a research spill-over failure. This can be ad-dressed by an R&D subsidy (of some sort) that increases the demand for research corre-sponding to the distance C leading to a further increase in R&D corresponding to this dis-tance. From a policy perspective, the main question concerns the mix between environmental taxes that internalise the environmental externality and public funding / subsidies of R&D inter-nalising the innovation spillovers. It is clear from the previous discussion that in a first best world each externality requires its own instrument. Economics literature provides two main conclusions in this respect. First, if the economy is in equilibrium, the tax and support instruments should be directly targeted to the corre-


46

sponding externality, c.f. the discussion above. Second, if the knowledge stock on environ-mentally friendly technologies is too low (i.e. there is disequilibrium), then both instruments could be used more aggressively than simply by addressing the equilibrium externalities. Ad-aptation of new policies that dramatically changes relative prices in the economy – i.e. price of carbon – is a clear example of how the stock of knowledge capital in a sector can change from being in equilibrium in view of the old policy stance and become far too low relative to future needs with new policy stance. Below, we discuss each of these arguments in turn. Starting with the argument focussing on equilibrium externality effects, we find that policies to address knowledge spillovers are more effective if they address all knowledge spillovers, rather than focusing exclusively on R&D pertaining to alternative energy.53 Similarly, the lit-erature suggests that environmental taxes should directly target what they are meant for – the externality.54 Not surprisingly, technology subsidies alone have a smaller environmental im-pact than policies that directly address the environmental externality. Yet, one can play with the idea of setting a high energy tax to attain both objectives.55 If the tax is sufficiently high, one could induce the level of innovations in green technologies that are optimal from a societal point of view. This corresponds to setting a high tax that shifts the dashed line in the right panel of Figure 3.1 sufficiently to reach the line for societal op-timal demand. However, the tax will necessarily be distorting. First, it will over-internalise the environmental externality and therefore distort consumption choices. Second, it will in-duce innovations in green technologies only, and will therefore not solve the general knowl-edge spill-over externality that exists for all types of R&D. In fact, the studies playing with this idea come to the conclusion that it is in-optimal from an environmental point of view.56 Turning to the case where we assume a current disequilibrium in the knowledge stock or similar, the previous conclusion will be challenged. For example, we can imagine that the green technologies R&D stock has been neglected for decades (say due to missing price sig-nals) such that the additional societal value of R&D in these technologies largely exceeds that of other technologies.57 In such cases, we need a push in green technology innovations. It is therefore argued that environmental taxes may play a role in achieving this.

53 Schneider and Goulder (1997). 54 Popp et al (2009). 55 Popp (2006), Hart (2008). A secondary argument relates to the use of revenues from taxes on energy and carbon. Potentially, the labour distortions resulting from energy taxes may be lower than income taxes. Hence a switch from energy taxes to income taxes may improve labour market functioning. However, this argument is tricky: energy taxes may have less of an impact on labour market because energy demand is less income elastic than other con-sumer goods and hence have a lower impact on marginal tax rates than for example VAT. But that is reflected by definition in a redistribution of net income from low to high income families. In effect any policy that accepts such a redistribution can be used to finance lower marginal tax rates. For this and other reason, the very extensive Mirrlees review (2010) undertaken by the Institute of Fiscal Studies,UK, was very sceptical about using labour mar-ket arguments for raising energy or other environmental taxes. 56 Hart (2008), Popp (2006), Greaker and Pade (2008). 57 See Acemoglu et al (2009) for an example of this argument.


47

A similar argument is based on rising societal cost of emissions (i.e., temperature increases become more and more expensive to deal with.)58 In this case, the spillovers from emissions-savings knowledge will again be more valuable than spillovers from other innovations, justi-fying a temporary increase in the optimal emissions tax as well as R&D support to account for differences in the social benefits of spillovers across technologies.59 Another type of exception – still arising from economic disequilibria – is if patent policy is weak. Then additional environmental taxes are justified as a second-best policy for address-ing the knowledge market spill-over.60 Still, we emphasise that the solution is second-best, especially as we would distort innovations towards green technologies (now assuming that this is not necessary.) Thus, our reading of the literature on solving the double externality problem suggests that both an environmental tax and R&D support is the only way of adequately addressing the problem.

The long term credibility problem R&D is a risky investment that, when yielding new profit opportunities for private compa-nies, will pay off in a distant future. Cost-benefit analyses of various research projects must therefore include the risk that pollution prices are not predictable far in the future.61 Since many research projects in green technologies can move from the green to the red zone for small variations in, say, emission prices, it is extremely important that long term tax policies are well-defined, credible, and demonstrate a high degree of continuity over time. The literature has recognised that it is the expectations of future policies that motivate R&D, and that emission caps put in place before innovations resulting from R&D can be deployed have no effect as incentives.62 Indeed, the literature emphasises the ‘announcement effect’ of future carbon limits.63 In the case of emissions prices, studies point to the large uncertainty attached to future commitments and allocation of allowances.64 The literature suggests that high volatility in prices of CO2 considerably reduce willingness to make early investments in low carbon power generation and carbon and capture storage (CCS) technologies.65 Such volatility sig-nificantly increases investment risk and cost of capital which makes it profitable to postpone investments. So CO2 price volatility may hamper the investments that climate policy is at-tempting to encourage. Uncertainty in climate policy contributes to volatile CO2 prices and

58 See, e.g., Fankhauser (1993). 59 Hart (2008). 60 Greaker and Pade (2008). 61 Baker and Adu-Bonnah (2008). 62 Yang et al (2008). 63 Montgomery (1972), Montgomery and Smith(2007). 64 DEFRA (2008). 65 Blyth et al (2007), Celebi and Graves (2009), Weber and Swider (2004).


48

therefore long-term policy certainty is vital to minimise investment risks in low carbon tech-nologies. The example of carbon prices, therefore, fits quite well with the notion of dynamic inconsis-tency.66 Carbon prices will need to be high to create additional R&D investment possibili-ties, but even if the policy makers announce future emission levels that create such an incen-tive, the government will prefer reneging on this level once the technology is developed. To sum up, when policy makers opt for more aggressive environmental taxes, it is of utter importance that these policies are credible and communicated in a convincing way. Any le-gal manoeuvre that will bind future policy makers to stringent carbon taxing will be good for current R&D investments.

R&D crowding out Crowding out of R&D is an issue when we attempt to emphasise innovations of green tech-nology, irrespective of our instruments. It is therefore related to the discussion of using taxes to equilibrate imbalances in the current state of the economy. Thus, if we assume a current disequilibrium in research efforts of energy technologies, the entire discussion of crowding out becomes much less important: one unit of more research in energy technology has a higher value to a society than 1 unit of research in other fields of research. In the relevant case of targeting R&D support in an economy with no knowledge stock imbalances, the ba-sic message from this subsection is that R&D externalities from all technology branches should be equilibrated. We should note two things from the beginning of this discussion. First, we note that R&D crowding out is unavoidable in reaching the goal of more research in green technologies – human and physical resources must be drawn from somewhere in the economy. Yet, in the typical understanding of the word, crowding out specifically refers to a reduction of research (spending, employment, patents, or similar) in another technology branch to substitute the increased research in green technologies. Second, we note that both higher taxes as well as larger targeted R&D subsidies lead to crowding out. However, while the two instruments are formally equivalent, there exists an important dif-ference in the labour market distortions created by each instrument. Where taxes will help collect revenue that can be recycled to neutralise any adverse effect on labour supply, R&D subsidies will have the opposite effect since gathering of tax revenue is a prerequisite for funding expenditure. So gains from spillovers must exceed tax induced distortions from RD funding.

66 Montgomery and Smith (2007).


49

In addition, our reading of the empirical literature on induced innovation tells us that crowding out seems to exist. 67 Impacts on aggregate R&D expenditures and patenting are much smaller than for specific, smaller technology classes. Thus, there seems to be smaller, if any, effects on total innovation stock, so the effects obtained at disaggregated level must stem from reallocation of R&D efforts.68

Command-and-control systems and innovation In this section we address the sufficiency of market based systems. That is, we ask if envi-ronmental taxes and R&D subsidies need to be complemented by command-and-control policies to boost innovation. We have already touched upon the question in section 2.1 above stating that campaigns together with information and technology standards seem to create larger consumer responses (thereby creating more innovation.) Evidence on the efficacy of command-and-control mechanisms is mixed as regards effect on innovation. Some studies point to the basic problem that there is no incentive to innovate beyond the current technology standard.69 Also, command-and-control is typically designed to punish underperformers, while over-performers – those being ahead of the standard – are not rewarded. Other studies, which look directly at patents related to a specific technology standard, find quite impressive innovation effects.70 Thus, command-and-control may under certain circumstances be efficient in reaching medium run climate targets while they do not seem to provide long run solutions. Critics of pure market-based systems also see a role for command-and-control mechanisms in introducing close-to-market technologies that would see long lead times due to uncertainties.71 Summing up, market-based mechanisms are likely to solve most of the double externality problem as long as the technologies of interest are not too costly and risky. Publicly funded research must step in to assure that the entire research portfolio is sufficiently diversified to reach the climate targets.

3.2. THE INTERACTION BETWEEN TAXATION AND R&D SUPPORT Given that taxation and R&D have different roles in promoting environmental friendly in-novation, we must consider how they interact under certain circumstances. We focus the discussion on the following two issues:

Innovation and rebound effects The proper timing of support and taxation

67 Popp (2004), Gerlagh (2008). 68 Hamamoto (2006), Brunnermeier and Cohen (2003). 69 Popp (2010), Magat (1978), Milliman and Prince (1989). 70 Popp (2006) 71 Montgomery and Smith (2007).


50

The discussion below examines these two points.

Support for green technologies, their deployment and rebound effects Any policy boosting technological advance in a particular area will drive consumer demand in this direction. If we support technologies that lower the costs of using energy, then con-sumers will respond by increasing the level of the energy consuming activity, c.f. the above discussion on energy price elasticity. This is the so called rebound effect. A study measuring the rebound effect from fuel efficiency innovations finds that 60 percent of the improve-ments is lost again due to the rebound effect.72 This estimate seems to represent similar stud-ies quite well, c.f. Table 3.1. Table 3.1 Overview of studies of rebound effects Author/Date Region Efficiency improvements Estimated rebound effect Semboja, 1994 Kenya Improvements in both production

and consumption sectors >100% in both cases

Dufournaud et al, 1994 Sudan 100-200% improvement in effi-ciency of in heating stoves

47-77%

Vikstrom, 2003 Sweden 15% in production sectors and 12% in energy sectors

50-60%

Washida, 2004

Japan

1% all sectors

53% in base case

Grepperud & Rasmussen, 2004 Norway Doubling of historical growth rate of electricity productivity for four sec-tors, and doubling of growth rate of oil efficiency for two sectors

Small for oil but >100% in some cases for electricity

Glomsrod &Taoyuan, 2005 China Deregulation of coal cleaning indus-try, lowering price and increasing supply of clean coal

>100%

Hanley et al, 2005 Scotland 5% for producers (including energy supply)

>100%

Allan et al, 2006 UK 5% for producers (including energy supply)

37% in base case

Frondel et al, 2007 Germany Historical fuel efficiency improve-ments in the transport sector

57-67%

Source: Copenhagen Economics based on Allan et al. (2007) In this respect, it is important to make a distinction between “dirty” and “clean” technolo-gies.73 If we promote zero carbon technologies, rebound is not so problematic (apart from energy efficiency goals), while support for, e.g., clean coal or new diesel engine design is more problematic. When we promote the latter type of innovations, then we need at the same time to ensure that economic incentives intended to save CO2 are set sufficiently high. We need to see this as a complementary policy design where taxes are raised and R&D pro-jects with high spill-over effects are supported at the same time. Moreover, with end-of-pipe technology focusing on removing/reducing the polluting quan-tity, while leaving the basic service unchanged, taxation/carbon pricing is required for any deployment to take place. Within the area of energy, Carbon Capture and Storage (CCS) technologies are a clear example. They will add substantially to the costs of producing energy

72 Frondel et al (2007). 73 Aghion et al (2004).


51

while producing the same good, namely electricity. Thus even huge amounts of subsidies for development of CCS will come to nothing if CO2 is not priced. To put this concisely: re-search may take place, but there is no diffusion into the economy due to lacking incentives. At the other end of the spectrum, support for innovation in energy efficiency, lowering the costs of energy efficient products will have inherent value to consumers, thus leading to more deployment of these products even in the absence of carbon pricing. This brings us back to the discussion of the rebound effects stated above. Thus, support of innovation in both “dirty” and end-of-pipe technologies clearly requires taxation/carbon pricing to put a price on emissions.

Timing of R&D support and taxation In this subsection, we look at issues related to timing of tax rates and R&D subsidies. The basic insight goes as follows: assuming that R&D subsidies today can induce technologies that lower abatement costs in the future, then it will be societal optimal to have a rising tax profile.74 To understand this argumentation, we must first recall that CO2 pollution is basi-cally a stock, we can add to over time, for example over the two time periods: “today” and “tomorrow”. When additional CO2 cannot exceed a certain level after tomorrow, then the question arises of how to divide emissions between now and tomorrow. And it is obviously better to abate more, when abatement costs are lowest. Thus, a rising tax profile will ac-commodate an intelligent division between abatement today and tomorrow given that inno-vation actually lowers tomorrow’s abatement costs. This insight can easily be demonstrated by simulations with CERIM. The model can show the necessary carbon taxes to achieve a certain emission target, say a 10 percent reduction over a 20 year period. Maximising social welfare (in the US electricity industry) will suggest a rising tax profile over the model’s two base periods as shown in 3.2. The figure shows the tax related consumer welfare losses for a base year on both periods of the model.75 Moreover, the figure does so for both a constant tax scheme (top) and the optimal rising tax scheme (bottom). Take the case of the constant tax scheme. Without the tax, the electricity market price would have been 81.8 $/MWh, while the after-tax price (including equilibrium effects) becomes 87.3 $/MWh. This corresponds to a welfare loss (Harberger triangle) of $ 173 mill. In a second period year, this will be only 128 $/MWh thanks to the progress in renewables. If there are equally as many years in both periods (and no discounting), we can calculate the welfare loss as the simple sum of welfare losses amounting to 301 mil $. In contrast, imple-menting the optimal rising tax scheme (bottom of figure) yields a welfare loss of just $ 220 mill.

74 Fisher et al (1999). 75 Technically, we need to refer to welfare losses since we do not explicitly evaluate the pollution costs. The basic idea is that we give up something in the product market (a welfare loss) to achieve something else in the pollution market (a welfare gain due to cleaner air and less climate change). When we fix the gain to a 10 percent reduction, we can focus entirely on minimizing the welfare losses.


52

Figure 3.2: Welfare loss for constant and rising tax rates when achieving a 10 percent reduction in CO2 emissions, CERIM simulations

Note: The tax profile is calculated as the endogenous carbon price in a version of the CERIM with fixed total emis-sions corresponding to a 20 percent reduction and maximised welfare.

Source: Copenhagen Economics However, the reverse profile is called for when learning costs are substantial.76 In that case, it is better to start abating today, since time is of the essence. Hence, there is a central differ-ence in appropriate policies depending on the nature of technological progress. Classical in-novations require a rising tax profile, while learning-by-doing implies a decreasing profile. It therefore becomes a central question to estimate the size of learning effects. The tradi-tional “learning literature provides estimates in the range 5 to 20 percent a year.77 In Figure 3.2 we demonstrate the difference in break-even allowance prices for a hypothetical technol-ogy close to the market, when learning effects are 5 and 20 percent pro annum respectively. Clearly, there is a substantial difference between the two extreme cases, since technologies can double their efficiency within just a few years in the case of 20 percent, while the same improvement takes around 15 years in the other case. Overall, our reading of the literature suggests that learning rates should be in the lower end of the scale. Once, it is recognised that actual technological progress for any given product is affected by both past investments in R&D and past deployment (“learning”), estimates of learning tends to be at the low end often with gains from R&D being dominant. Further-more, “learning” is not for free: increased innovation in any particular technology tends to

76 Popp et al (2009). 77 McDonald and Schrattenholzer (2000)

Period 1 Period 2

Constant tax

Rising tax

87.3 $/MWh

81.8 $/MWh

81.9 $/MWh

77.1 $/MWh

173 mio. $ 128 mio. $

85.4 $/MWh

81.8 $/MWh

60 mio. $

82.4 $/MWh

77.1 $/MWh

160 mio. $


53

crowd out partially other advances of technology in the section on that subject. In short, a lot of caution is required when putting forward learning costs as an argument for early de-ployment. In Box 3.2 we provide a further discussion of learning effects and estimates thereof. Figure 3.2: Development of break-even prices for different learning effects

Note: The figure shows the development in the lowest possible allowance price that would facilitate introduction of

a certain technology experiencing learning effects. Source: Copenhagen Economics

0

20

40

60

80

100

120

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Years from t0

Price

5 % learning effect

20 % learning effect


54

Box 3.2: Learning effect estimates

Source: Popp et al (2009)

3.3. TIME HORIZON AND STRINGENCY OF POLICY TARGETS In this paragraph we discuss how time horizon and stringency of policy targets influence the mixing of instruments. The basic insight here is that in the case of ambitious policies in a near future, we definitely need both taxation and R&D support. Less ambitious targets in a more distant future would mean that gradual improvements are sufficient and taxation itself will do a reasonably good job.

Time horizon A good way of shaping the discussion is to step back for a moment and consider Figure 3.3. The figure shows a standard innovation cycle from the initiation of R&D to final products that are sufficiently economical to penetrate the relevant market. The figure starts with the idea stage, leading approximately up to the patenting point, where researchers develop con-ceptual descriptions of new green technologies. We recall from the previous section that

Typically, studies on new energy technologies find faster learning for younger technologies, with esti-mates clustering around 15-20% for alternative energy sources such as wind and solar energy (McDonald and Schrattenholzer 2000).

Table 3.2: Learning rates Study Technology Diffusion Innovation

Criqui et al (2000) Wind 16 % 7 %

Jamasab (2007) Wind 13 % 26 %

Söderholm and Klassens (2007) Wind 3 % 13 %

Klassens et al (2005) Wind 13 %

Criqui et al (2000) PV 20 % 10 %

Jamasab (2007) Solar 2 % 5 %

Jamasab (2007) Nuclear 37 % 24 %

Jamasab (2007) CCGT 1 % 18 %

Source: Copenhagen Economics based on studies from the list One significant caveat with estimated learning rates is that they typically focus on correlations between energy technology usage and costs, rather than causation. Recent papers by Klaassens et al. (2005), Söderholm and Sundqvist (2007), and Söderholm and Klaassens (2007) attempt to disentangle the sepa-rate contributions of R&D and experience by estimating “two-factor” learning curves for environmental technologies. These two-factor curves model cost reductions as a function of both cumulative capacity (learning-by-doing) and R&D (learning-by-searching, or LBS). To be comparable with the notion of cumu-lative capacity, in these models R&D is typically aggregated into a stock of R&D capital. Thus, endogeneity is a concern, as we would expect both investments in capacity to be a function of past R&D expenditures and R&D expenditures to be influenced by capacity, which helps determine demand for R&D. Söderholm and Sundqvist address this endogeneity in their paper and find LBD rates around 5 percent, and LBS rates around 15 percent, suggesting that R&D, rather than learning-by-doing, contributes more to cost reduc-tions. However, these results are very sensitive to the model specification, illustrating the difficulty of sorting through the various channels through which costs may fall over time. To further address the problems associated with estimating and interpreting learning curves, Nemet (2006) uses simulation techniques to decompose cost reductions for PV cells into seven categories. Plant size (e.g. returns to scale), efficiency improvements, and lower silicon costs explain the majority of cost reductions. Notably, most of the major improvements in efficiency come from universities, where tradi-tional learning by doing through production experience would not be a factor. Learning from experience (e.g. through increased yield of PV cells) plays a much smaller role, accounting for just 10 percent of the cost decreases in Nemet’s sample.


55

price effects have around two to five year’s half-life on patents, implying that this phase in it-self is considerable when considering the induced innovation hypothesis. The next phase concerns prototype development. After this stage, the technologies are largely classified by their cost-effectiveness compared to current market technologies. Some innovations may climb up the ladder rather quickly while others require longer development times, e.g., due to complementarities with other technologies. Figure 3.3: Policy instruments during the innovation cycle

Mar

ketd

eplo

ymen

t

Idea stage

Prototype & demonstration stage

technologies(e. g. 2nd Generation

biofuels

High cost-gaptechnologies

(e.g. PV)

Low cost-gaptechnologies

(e.g. windonshore)

Maturetehcnologies(e.g. hydro)

Continuity, RD&D, create marketattractiveness

Stability, low-riskincentives

Technology-neutralcompetition

Stimulate market pull

Imposed market risk, guarenteed but declining

minimum return

Development Niche markerts Mass market Time

Source: Copenhagen Economics based on IEA (2008) There are four policy issues emerging from the figure. First, there is the question of the cy-cle’s time length – what is realistically attained within a 10, 20 or 50 year timeframe? As just mentioned, the literature (including our own investigations) estimates that half of induced patents are accepted within ca. 2-5 years after the price or tax change. The other half takes longer. And since patents are just the first stage on a long journey towards final products, we must accept that we often work with very long time leads. Unfortunately, the empirical literature is not very helpful in determining normal time frames for the remaining stages of the innovation cycle. The obvious interest of policy makers is the accumulated time frame from tax intervention to market penetration of green technologies, so we may draw on some historical examples. This is the content of Table 3.3. The table shows the time from conception, typically measured by the patent application, to market maturity (production) for five selected green technologies. We draw on some quite old and some newer examples. In most cases, we find innovation cycles of +20 years, while only Warren Johnson’s room thermostat had a very short lead time from patent in 1883 to large-scale production in 1885.


56

Table 3.3: Examples of innovation lead times Technology Short description Conception Market matur-

ity Motor vehicle technologies Energy Saving Module Ensure that compressors work at maximum ef-

ficiency 1983 Post 2000

EFI Electronic fuel injection 1952 1982 End-of-pipe technologies Carbon capture and storage Post-combustion CO2 is captured and stored

geologically 1977 Estimated 2008

Room heating technologies CO2 heat pump system Used for heating 1995 not yet

Mineral wool Used for thermal insulation 1840 1871 Room thermostat Regulate temperature of room or system 1883 1885 Source: Copenhagen Economics When lead times are significant, there are obvious implications for policy ambitions at dif-ferent time horizons. We simply cannot expect R&D subsidies to contribute much to the agreed 2020 Kyoto targets (as seen from today.) However, in reaching much more substan-tial reductions in 2050, it is much more reasonable to assume that innovation will play a role as discussed below.

Stringency of policy targets The higher the policy ambition, the higher is the level of optimal private and public spend-ing on climate related technologies. It should be clear from the previous discussions that low policy ambitions, e.g., low emission reduction targets, will imply relatively low value of R&D for the abaters: options based on existing technologies are cheaper. Instead, setting a high level of ambitions will imply a relatively high value of R&D, since current technologies are insufficient. A basic question is then how much more is optimal. For example will going from a 20 per cent cut in emissions to 40 per cent emissions in the same time period, require more or less than a doubling of optimal RD support levels? The size of the optimal increase is driven by factors pulling in different directions. On the one hand, the larger the ambitions, the larger the distortions to consumer welfare by using only existing technologies for abatement. This goes back to simple welfare economics as il-lustrated in figure 3.2 above: the additional loss to society of accepting in period 1 a more ambitious target is 113(=173-60) which is substantially larger than accepting the lower level of ambition which equals 60 despite a less than proportional increase in the emission cut. So the potential value to society of R&D that can replace existing technology is an increasing function of the levels of ambitions. This suggests that doubling the level of ambitions re-quire more than a doubling of R&D expenditures. On the other hand, the marginal return on a net basis of yet more R&D efforts going into the energy sector is falling as the quality of each new additional research project. Moreover,


57

the costs of producing quality-adjusted research is going up for example as the marginal costs of loosing innovation and production elsewhere in the economy as more resources are pulled into energy research are rising. The magnitude also plays a role for the way subsidies to R&D should be assigned. Simplify-ing quite a bit, we can state that subsidies to industrial R&D primarily obtain improvements in known technologies, and therefore primarily serve to shorten innovation cycles.78 In con-trast, subsidies to public fundamental research will serve to start entirely new inventions. So aiming for ambitious targets of 50-80 percent reduction suggests a larger fraction of R&D subsidies to public fundamental research. Consider, for instance, the case of solar energy. Despite research efforts that began during the energy crises of the 1970s, solar energy is still only cost competitive in niche markets, such as remote off-grid locations. This leaves a potential role for government-sponsored R&D to fill in the gaps, particularly in the case of climate change, where a diversified energy portfolio will be necessary to meet currently proposed emission reduction targets. So the bottom line is: the more stringent the targets the larger the role of R&D support to support innovation focusing more on technology leaps as opposed to marginal improve-ments of existing technologies where carbon pricing should do the main job.

3.4. THE EFFICIENT POLICY MIX OF CARBON PRICING AND R&D SUPPORT In this final section of the chapter, we will attempt to summarise the policy findings from above. Based on our previous results, we conclude that R&D support is particularly impor-tant if:

knowledge spill- over externalities are high, crowding out is limited, high imbalance between desired and actual research stock for exampling resulting

from adaptation of very ambitious long term climate and energy policies, public costs of funding are low, results from policy induced innovation come quick enough to help compliance

with policy objective.

Similarly, taxes are particularly important if:

pollution externalities are high, long term price signals are missing, end-of-pipe technologies are cost-efficient solutions to abatement,

78 Popp (2010)


58

strong rebound effects are present (higher demand for fossil fuels induced by tech-nology progress).

According to the above discussion, the two main ingredients in defining the efficient policy mix are the respective externalities from knowledge and from pollution which need to be de-fined in a long term perspective. The literature seems to suggest that pollution externalities are larger than knowledge externalities79. Indeed, while all such calculations are very sensitive to parameter assumptions as well as the policy goals to be reached, a number of recent em-pirical studies confirm the primacy of taxation and equivalent instruments in reaching long term climate and energy policy goals while also underlining the very useful role that direct RD support policies can deliver.80

• A OECD study suggests that carbon pricing consistent with ambitious 2050 global goals could induce a “four-fold” increase in energy R&D expenditure while public R&D policies could most productively be focused on “major” technological break-throughs” rather than marginal innovations81. It also concludes that even a 30-fold increase in energy related R&D would be insufficient to stabilise emissions which is linked to some of the issues discussed in this study such as rebound effects.

• An EU study suggests that, in addition to much higher carbon prices, substantial, frontloaded R&D support is needed to adjust to ambitious long term climate goals82. Other subsidy mechanisms are also investigated, but the combination of tightening emissions-caps (rising carbon prices) and up-front R&D support for green technologies yields the most favourable economic outcome. The study also concludes, similar to our previous findings, that R&D support must not favour green technologies in the long run and therefore suggests a phasing out of R&D support for green technologies by spreading it to all sectors of the economy. The results are based on a forward-looking, general equilibrium model of the European economy where R&D and innovation is specifically modelled.

• A study focusing on US compliance with climate policy objectives finds that car-bon taxes alone achieves 95 per cent of the welfare gains compared to the first-best case of both an optimally-designed carbon tax (one equating the marginal benefits of carbon reductions with the marginal costs of such reductions) and optimally de-signed R&D subsidies. By contrast, working with an optimal R&D subsidy alone attains just 11% of the welfare gains.

• A study on climate policies directed at the US electricity sector finds that the rank-ing of potential policy instruments is roughly as follows: (1) emissions price/tax, (2) emissions performance standard, (3) fossil power tax, (4) renewables share require-ment, (5) renewables subsidy, and (6) R&D subsidy. Nonetheless, an optimal port-

79 Popp (2006), Fisher (2008). 80 Popp (2010). 81 Bosetti et al(2009) 82 ECFIN(2010)


59

folio of policies – including emissions pricing and R&D – achieves emission reduc-tions at significantly lower costs than any single policy.


60

Acemoglu, 2002, “Directed technical change.” Rev. Econ. Stud. 69, 781-809

Acemoglu, Aghiom, Bursztyn, and Hermous, 2009, “The Environment and Directed Technological Change,” NBER Working Paper, No. 15451

Aghion et. al., 2004 “Competition and Innovation: An Inverted-U Relationship”, The Quarterly Journal of Economics, Vol. 120, No. 2, Pages 701-728

Aghion et. al., 2009, “No green growth without innovation”, Bruegel policy brief, issue 7 Aldy and Pizer, 2008, “Issues in Designing U.S. Climate Change Policy”, Discussion Pa-

per, RFF DP 08-20 Aldy et. al., 2009, “Designing Climate Mitigation Policy”, Discussion Paper, RFF DP 08-

16

Babiker, 2005, “Climate change policy, market structure, and carbon leakage”, Journal of International Economics 65, 421– 445

Barker et. al., 2007, “Carbon leakage from unilateral Environmental Tax Reforms in Europe, 1995–2005”, Cambridge Econometrics, Covent Garden, Cambridge, CB1 2HS, UK

Baker and Adu-Bonnah, 2008, “Investment in risky R&D programs in the face of climate uncertainty”, Energy Economics, 30, 465-486

Baker, Clarke and Shittu, 2008, “Technical change and the marginal cost of abatement”, Energy Economics, 30, 2799-2816

Bernstein and Griffin, 2005, “Regional Differences in the Price- Elasticity of Demand For energy”, RAND Technical Reports

Bertoldi and Atanasiu, 2007, “Electricity Consumption and Efficiency Trends in the Enlarged European Union”, Institute for Environment and Sustainability, EUR 22753 EN

Binswanger and Ruttan, 1978, “Induced Innovation: Technology Institutions and Devel-opment.” Johns Hopkins University Press, Baltimore, MD.

Blyth et. al., 2007, “Investment risks under uncertain climate change policy”, Energy Pol-icy 35, 5766–5773

Bohi and Zimmerman, 1984, “An update on econometric studies of energy demand behav-

REFERENCES


61

ior”, Ann. Rev. Energy, 9, 105-154

Bond, 2002, “Dynamic panel data models: a guide to micro data methods and practice”, Institute for Fiscal Studies, 7 Ridgmount Street, WC1E 7AE London, UK

Borenstein and Shepard, 1996, “Dynamic pricing in retail gasoline markets”, RAND Jour-nal of Economics, vol. 27, no. 3

Bosetti et. al., 2009, “The role of R&D and technology diffusion in climate change mitiga-tion: New perspectives using the WITCH model”, Economics department working papers, no. 664

Bosetti et al., 2009b, “Climate Policy after 2012,” CESifo Economic Studies, Vol. 55, No. 2, pp. 235-254.

Bramoullé and Olson, 2005, “Allocation of pollution abatement under learning by doing”, Journal of Public Economics, 89, 1935-1960

Brenton, 1997, “Estimates of the demand for energy using cross-country consumption data”, Applied Economics 29, 851-859

Brons et. al., 2006, “A Meta-analysis of the Price Elasticity of Gasoline Demand. A System of Equations Approach”, Tinbergen Institute Discussion Paper, 106/3

Brons et. al., 2008, “A meta-analysis of the price elasticity of gasoline demand. A SUR ap-proach”, Volume 30, Issue 5, September 2008, Pages 2105-2122

Brunnermeier and Cohen, 2003, “Determinants of environmental innovation in US manu-facturing industries”, J Environ Econom Manage 45, 278-293

Bruno, 2005, “Approximating the bias of the LSDV estimator for dynamic unbalanced panel data models”, Economics Letters 87, 361-366

Burniaux et. al., 2008, “The economics of climate change mitigation: Policies and options for the future”, Economics department working paper, no. 658

Busse et. al., 2009, “The Differential Effect of Usage Cost on New and Used Durable Goods Markets: Evidence from the Automobile Industry” manuscript, Northwest-ern University and NBER.

Celebi, M. and Frank Graves (2009), Volatile CO2 prices discourage CCS investments, mimeo.


62

Castelnuovo, Galeotti, Gambarelli and Vergalli, 2005, “Learning-by-Doing vs. Learning by Researching in a model of climate change policy analysis”, Ecological Economics, 54, 261-276

Dahl, 1993, “A survey of energy demand elasticities in support of the development of the NEMS”, MPRA Paper No. 13962

David, Hall, and Toole, 2000, “Is Public R&D a Complement of Substitute for Private R&D? A Review of the Econometric Evidence," Research Policy, 29, 497-529.

DEFRA, Department for Environment, Food and Rural Affairs, 2008, “An analysis of the technical feasibility and potential cost of a personal carbon trading scheme”, www.defra.gov.uk

Demailly and Quirion, 2006, “Leakage from climate policies and border tax adjustment: lessons from a geographic model of the cement industry”, Working Papers halshs-00009337 version 1, HAL

EFTEC, Economics for the Environment Consultancy Ltd., 2008, “Demand for Cars and their Attributes”, report prepared for the Department of Transport, UK, London, January.

EC, European Commission, 2008, “Proposal for a Directive of the European Parliament and of the Council amending" Directive 2003/87/EC so as to improve and extend the EU greenhouse gas emission allowance trading system, COM(2008), Brussels, 23.1.2008.

ECFIN, 2010, “What is the growth potential of green innovation? An assessment of EU climate policy options,” Economic Papers, No. 413.

Espey, 1998, “Gasoline demand revisited: an international meta-analysis of elasticities”, Energy Econimics, 20, 273-295

Espey and Espey, 2004, “Turning on the Lights: A Meta-Analysis of Residential Electricity Demand Elasticities”, Journal of Agricultural and Applied Economics , 36,1, 65-81

Fankhauser, 1993, “Evaluating the social costs of greenhouse gas emissions”, CSERGE Working Paper GEC 94-01

Fischer, 2008, “Emissions Pricing, spillovers, and public investment in environmentally friendly technologies”, Energy Economics, 30, 487-502

Fischer, 2009, “The Role of Technology Policies in Climate Mitigation”, RFF, Issue Brief


63

no. 08-09

Fischer and Newell, 2008, “Enviromental and technology policies for climate mitigation”, Journal of Environmental Economics and Management, 55, 142-162

Fischer, Parry, and Pizer, 1999 “Instrument choice for environmental protection when technological innovation is endogenous”, Discussion Paper, Resources for the Fu-ture, Washington, DC.

Fischer, Parry and Pizer, 2003, “Instrument choice for environmental protection when technological innovation is endogenous”, Journal of Environmental Economics and Management, 45, 523-545

Flood et. al., 2010, “Are demand elasticities affected by politically determined tax levels? Simultaneous estimates of gasoline demand and price”, Applied Economics Letters, Volume 17, pages 325 – 328.

Frondel, Horbach, and Rennings, 2007, “End-of-pipe or cleaner production? An empirical comparison of environmental innovation decisions across OECD countries”. Busi-ness Strategy and the Environment 16, 571-584.

Gardes et. al., 1996, “Cross-section versus time-series income elasticities”, Economics Let-ters 51, 169-175.

Gerlagh, 2008, “A climate-change policy induced shift from innovations in carbon-energy production to carbon-energy savings”, Energy Econ. 30, 425-48

Gillingham, Newell and Palmer, 2006, “Energy Efficiency Policies: A Retrospective Ex-amination”, Annual Reviews

Gillingham, Newell and Palmer, 2009, “Energy efficiency economics and policy”, NBER working paper series, 15031

Gillingham, Newell and Pizer, 2008, “Modeling Endogenous technological change for cli-mate policy analysis”, Energy Economics, 30, 2734-2753

Goel, 1994, “Quasi-experimental taxation elasticities of US gasoline demand”, Energy Economics, 16 (2), 133-137

Goolsbee, 1998, “Does Government R&D Policy Mainly Benefit Scientists and Engi-neers?" American Economic Review, 88(2), 298-302.

Goulder and Schneider, 1999, “Induced technological change and the attractiveness og


64

CO2 abatement policies”, Resource and energy Economics, 21, 211-253

Greaker and Pade, 2008, “Optimal CO2 abatement and technological change: Should emission taxes start high in order to spur R&D?” Discussion Papers No. 548, Statis-tics Norway, Research Department.

Greene, 2003, “Econometric Analysis”, fifth edition, Prentice Hall Upper Saddle River, New Jersey 07458

Hamamoto, 2006, “Environmental regulation and the productivity of Japanese manufac-turing industries.” Resource Energy Econ. 28, 299-312

Hart, 2008, “The timing of taxes on CO2 emissions when technological change is endoge-nous”. Journal of Environmental Economics and Management 55(2), 194-212.

Hascic et. al., 2008, “Environmental Policy Stringency and Technological Innovation: Evi-dence from Patent Counts”, June 15

Hicks, 1932, “The Theory of Wages Macmillan”, London

Hitt et. al., 1994, “A mid-range theory of the interactive effects of international and prod-uct diversification on innovation and performance”, Journal of Management, Vol-ume 20, Issue 2, 297-326

House of Lords, 2005, “The Economics of Climate Change”, Select Committee on Eco-nomic Affairs

International Energy Agency, 2009, “Ensuring Green Growth in a Time og economic Cri-sis”, OECD/IEA, G8 Siracusa 22-23-24 Aprile

Jaffe and Palmer, 1997, “Environmental regulation and innovation: a panel data study” Rev Econ Stat 79(4), 610–619

Johnstone et. al., 2008, “Renewable energy Policies and technological innovation: Evidence based on patent counts”, NBER working paper series, 13760

Johnstone et. al., 2009 (a), “Renewable energy Policies and technological innovation: Evi-dence based on patent counts”, Springer, Environ Resource Econ

Johnstone et. al., 2009 (b), “Invention and Transfer of Climate Change Mitigation Tech-nologies on a Global Scale: A Study Drawing on Patent Data”, Fondazione Eni En-rico Mattei, Nota Di Lavoro, 82


65

Kaldellis and Gavras, 2000, “The economic viability of commercial wind plants in Greece, A complete sensitivity analysis”, Energy Policy, 28, 509-517

Kilian, 2007, “The Economic Effects of Energy Price Shocks”, University of Michigan and CEPR

Kiviet, 1995, “On bias, inconsistency, and efficiency of various estimators in dynamic panel data models”, Journal of Econometrics 68, 53-78

Klaassen et. al., 2005, “The impact of R&D on innovation for wind energy in Denmark, Germany and United Kingdom”, Ecological Economics, 54, 227-240

Klier, 2008, “The price of gasoline and the demand for fuel efficiency: Evidence from monthly new vehicles sales data”, Federal Reserve Bank of Chicago.

Knigge and Görlach, 2005, “Effects of Germany’s Ecological Tax Reforms on the Envi-ronment, Employment and Technological Innovation”, Ecologic, Institute for In-ternational and European Environmental Policy

Koetse et. al., 2008, “Capital-energy substitution and shifts in factor demand: A meta-analysis”, Energy Economics, 30, 2236-2251

Kydes, 1997, “Sensitivity of Energy Intensity in U.S. Energy Markets to Technological Change and Adoption”, Office of Integrated Analysis and Forecasting, U.S. De-partment of Energy

Kydland and Prescott, 1977, “Rules Rather than Discretion: The Inconsistency of Optimal Plans”, The Journal of Political Economy, 1977, vol. 85, no. 3

Labandeira et. al., “Effects of green tax reforms in Spain. A new analytical approach inte-grating micro and macro-economic models”

Lanjouw and Mody, 1996, “Innovation and the international diffusion of environmentally responsive technology”, Res Policy 25, 549–571

Lanoie et. al., 2007, “Environmental Policy, Innovation and Performance: new Insights on the Porter Hypothesis”, CIRANO

Levin, Klevorick, Nelson, and Winter, 1987, “Appropriating the returns from industrial re-search and development”, Brookings Papers on Economic Activity, 3, 783–820.

Liu, 2004, “Estimating Energy Demand, Elasticities for OECD Countries, A Dynamic Panel Data Approach”, Statistics Norway, Research Department, Discussion Papers


66

no. 373

Magat, 1978, “Pollution control and technological advance: A dynamic model of the firm.” J. Environ. Econ. Manage. 5, 1-25

Martin, 2009, “The Impacts of the Climate Change Levy on Business: Evidence from Mi-crodata”, Centre for Climate Change Economics and Policy, Working Paper No. 7, Grantham Research Institute for climate Change and the Environment, Working paper No. 6

Martiskainen, 2007, “Affecting consumer behavior on energy demand”, Sussex Energy Group, SPRU - Science and Technology Policy Research, University of Sussex

McDonald and Schrattenholzer, 2000, “Learning rates for energy technologies.” Energy Pol. 29, 255-61

Milliman and Prince, 1989, “Firm incentives to promote technological change in pollution control.” J. Environ. Econ. Manage. 17, 247-65

Mirrless Review(2010), Reforming the tax system for the 21th century”, Institute for Fiscal Studies, Oxford University Press

Montgomery, 1972, “Markets in Licenses and Efficient Pollution Control Programs.” Journal of Economic Theory 5, 395–418.

Montgomery and Smith, 2005, “Price, Quantity, and Technology Strategies for Climate Change Policy”, CRA International

Montgomery and Smith, 2007, “Price, quantity and technology strategies for climate change policy”. In: Human-Induced Climate Change: An Interdisciplinary Assess-ment. Cambridge University Press, Cambridge.

National Science Board, 2006, “Science and Engineering Indicators 2006”, National Sci-ence Foundation, vol. 1

National Science Board, 2006, “Science and Engineering Indicators 2006”, National Sci-ence Foundation, vol. 2

Nemet, 2006, “Beyond the learning curve: factors influencing cost reductions in photovol-taics”, Energy Policy, 34, 3218-3232

Newell, 2009, “International Climate Technology Strategies”, Duke University, National


67

Bureau of Economic Research

Newell et. al., 1999, “The induced innovation hypothesis and energy-saving technological change”, The Quarterly Journal of economics, vol. 114, no. 3

Newell and Pizer, 2008, “Carbon mitigation costs for the commercial building sector: Dis-crete-continuous choice analysis og multifuel energy demand”, Resource end En-ergy Economics, 30, 527-539

Newell and Rogers, “The Market-Based Lead Phasedown”, edited by Freeman and Kolstad, 2007, “Moving to markets in environmental regulation, lessons from twenty years of experience”, Oxford University Press, 173-195

Noailly et. al., 2008, “The impact of environmental policy on energy-efficient innovations in buildings: The case of the Netherlands”, CPB Netherlands Bureau for Economic Policy Analysis

Nordhaus, 2002, “Modeling Induced Innovation in Climate Change Policy”, Resources for the Future Press, chapter 9

OECD, 2009, “Working Party on National Environmental Policies, Indicators of Innova-tion and Transfer in Environmentally Sound Technologies: Methodological Issues”, ENV/EPOC/WPNEP1/FINAL

OECD, 2008, “Compendium of Patent Statistics”, OECD Peeters and Surry, 2000, “Incorporating Price-Induced in a Symmetric Generalised

McFadden Cost Function with several outputs”, Journal of Productivity Analysis 14, 53-70

Pizer and Popp, 2008, “Endogenizing technological change: Matching empirical evidence to modeling needs”, Energy Economics, 30, 2754-2770

Popp, 1999, “Induced Innovation and Energy Prices”, The University of Kansas

Popp, 2002, “Induced Innovation and Energy Prices”, Amer. Ec. Rev. 92 (1), 160-180

Popp, 2004, “ENTICE: endogenous technological change in the DICE model of global warming”, Journal of Environmental Economics and Management, 48, 742-768

Popp, 2006, “R6D subsidies and climate policy: Is there a “free lunch”?”, Climatic Change, 77, 311-341


68

Popp, 2010, “Innovation and Climate Policy”, NBER Working Paper Series, 15673

Popp and Newell, 2009, “Where does energy R&D come from? Examining Crowding Out from Environmentally-Friendly R&D.” NBER Working Paper, 15423

Popp et. al., 2009, “Energy, the environment, and technological change”, NBER Working Paper Series, 14832

Reichman, 2008, “Intellectual Property and Alternatives: Strategies for Green Innovation”, Chatham House, Energy, Environment and Development Working Paper: 08/03

Reinaud, 2008, “Climate Policy and Carbon Leakage, Impacts of the European Emissions Trading Scheme on Aluminium”, IEA Information Paper

Reinthaler and Wolff, 2004, “The Effectiveness of Subsidies Revisited: Accounting for Wage and Employment Effects in Business R&D”, ZEI Working Paper, B 21

Requate, 2005, “Timing and Commitment of Environmental Policy, Adoption of New Technology, and Repercussions on R&D”, Environmental and Resource Econom-ics, 31, 175-199

Rivers and Jaccard, 2006, “Choice of environmental policy in the presence of learning by doing”, Energy Economics, 28, 223-242

Roy et. al., 2006, “Substitution and price elasticity estimates using inter-country pooled data in a translog cost model”, Lawrence Berkeley National Laboratory, LBNL-55306

Schneider and Goulder, 1997, “Achieving low-cost emissions targets”, Nature, vol. 389

Smulders and Corrado Di Maria, 2007, “Endogenous Technological Change and the Cost of Environmental Policy”, February 7

Smulders and Nooij, 2003, “The impact of energy conservation on technology and eco-nomic growth”, Resource and Energy Economics, 25, 59-79

Stern, 2007, “The Economics of Climate Change: The Stern Review.” Cambridge: Cam-bridge University Press.

Stern, 2009, “Interfuel Substitution: A Meta-Analysis”, MPRA, paper no. 15792

Suslov, 2008, “Explaining Factors Affecting Energy Demand Elasticity: Does the Quality


69

of Institutions Matter?”, Faculty of Economics, Novosibirsk State University

Szabo et. al., 2006, “CO2 emission trading within the European Union and Annex B countries: the cement industry case”, Energy Policy, Volume 34, Issue 1, January, 72-87

Vollebergh and Kemfert, 2005, “The role of technological change for a sustainable devel-opment”, Ecological Economics, 54, 133-147

Wade, 2003, “Price Responsiveness in the AEO2003 NEMS Residential and Commercial Buildings Sector Models”, Energy Information Administration/Price Responsive-ness in the NEMS Buildings Sector Models

Weber, C. and D. J. Swider (2004), Power Plant Investments Under Fuel and Carbon Price Uncertainty

Wolff and Reinthaler, 2008, “The effectiveness of subsidies revisited: Accounting for wage and employment effects in business R&D”, Research Policy 37, 1403–1412

Yang et. al., 2008, “Evaluating the power investment options with uncertainty in climate policy”, Economics 30, 1933–1950.


70

A.1 METHODOLOGICAL FRAMEWORK In the multivariate work, we restrict attention to countries with at least 1000 patents in the technology areas covered by this study. That leaves us with 33 countries over 30 years (1978-2007). From these, we focus attention on countries with an average of at least 5 patents per year for each technology. Also, due to missing data in a number of the independent variables or the patent counts, the number of observations in the regression results will be substan-tially lower than the potential 990 observations. Countries may differ – due to cultural reasons and unobserved economic variations – in their patenting propensity. We follow Johnstone et al. (2008) and OECD (2009) in employ-ing quasi-static panel regressions with fixed-effects at the country level and, due to robust-ness concerns, extend the analysis to include dynamic panel regressions too. We specify our static regression equation as follows:

where k denotes the country and t denotes time (in years). Pt is the total number of EPO patent filings in year t. This variable serves to capture common trends in patenting which are unrelated to patenting incentives for environmentally relevant technologies. Rkt is govern-ment R&D in the technology of interest, and pkt is the price of a relevant energy input in country k and year t. The regressor Ikt reflects possible policy interventions, e.g., the an-nouncement or the enactment of a European directive. The regression is estimated separately for the patenting areas of interest, such as lighting, LEDs, etc. Note that in this log-log specification, θ indicates the relative change in the number of pat-ent filings in response to a relative change in the price level, i.e. an elasticity. More precisely, a unit coefficient would mean that for a positive (negative) change in price by ten percent, patent filings would increase (decrease) by approximately ten percent. For large changes in the price variable, this statement does not longer hold due to the non-linear nature of the re-lationship and an explicit calculation is required. The advantage of estimating elasticities, their interpretation apart, is that we are able to compare estimates across technologies. We estimate the static equation above using standard fixed effects estimators. Johnstone et al. (2008) estimate a similar regression in a negative binomial framework. When we employ such count data estimators, we achieve almost identical results to our fixed effects regres-sions. Therefore, we do not present them below. The dynamic model is specified as follows:

That is, we introduce dynamics in the endogenous variable, patent counts, and in the ex-planatory price variable(s). The advantage of this specification is twofold. First, we address

APPENDIX A: EMPIRICAL ANALYSIS OF PATENT DATA


71

an econometric issue concerning potential inconsistency of the static estimator when the data is actually dynamic in nature.83 There are valid reasons why patenting may happen in ‘waves’: ideas foster ideas and R&D expenditure is typically assigned over longer time spans. Second, the formulation allows us to test hypotheses about the way prices and taxes influ-ence patenting. If we simplify notation by joining all other explanatory variables in the con-stant term, then we can rewrite the model as:

∆ 1 1 ∆

This is the error correction representation of the model. The term in square brackets repre-sents the long run relation between patents and prices, while the remaining terms describe short run adjustments. In this way we can test hypotheses of differences in short vs. long term impacts from price changes to patenting. The dynamic panel regression is estimated by deploying two estimators. First, we use the bias-corrected LSDV estimator suggested by Kiviet (1995). The advantage of this estimator is that it works significantly well for small-N samples, c.f. Bruno (2005). Standard errors are found by bootstrapping. Second, we esti-mate using the standard Arellano-Bond estimator – the classic dynamic panel data estimator. After estimation we test the significance of both the short run effects, i.e., the t-values for , and the long run effects using Chi-square test for the long run coefficients in the square brackets. Note that we a priori value the static and the dynamic model equally. They are simply ro-bustness controls of each other. However, if the estimate of the lagged dependent variable is statistically significant and large, we favour the dynamic estimates, whereas small estimates suggest that the more parsimonious static model is preferable. Finally, we also attempt to include taxes in the analysis. Our data on prices and taxes have the following additivity characteristic:

Now, including taxes in the above regression would therefore cause multicollinearity prob-lems. However, given the logarithmic transformation, we would violate the additivity char-acteristic by using instead. Thus, we need a different approach. Assuming that taxes may have a different impact on innovation than prices, we can postulate the following rela-tion:

83 See Bond (2002).


72

where ct captures all other variables (including the residual) and φ measures the additional effect from taxes. Note that φ = 0 means that taxes have the same impact as prices. In logs, this is approximated by log βlnp + βln(1+ φtax/p) . The second term can be approximated by φtax/p. This would call for using lnp (where price is inclusive of taxes) and the share vari-able tax/p (the share of taxes in prices). This resolves the collinearity problem between prices and taxes – at least to some degree – and maintains a generalized form of additivity. The co-efficient of the share variable then has the usual interpretation of an excess effect (positive coefficients indicate that taxes work more strongly than prices, vice versa for negative ef-fects). Short and long run effects can now be calculated in the same way as shown for prices above.

A.2 RESULTS Below we present results from the econometric estimations for each technology apart. How-ever, the tables presenting regression output will have a similar structure across technologies:

1. Column (1) contains the coefficient estimates from the static model in its standard form. We include both residential and industry prices if relevant.

2. Column (2) contains the coefficient estimates from the static model including only the most relevant sector prices.

3. Column (3) contains the coefficient estimates from the dynamic model with price lags (leading to the error correction model). We include only the most relevant price series (residential or industry). This column is based on the Kiviet estimator.

4. Column (4) contains the estimates from the Arellano-Bond estimator for the same model as in column (3).

All models include the tax share variable as based on the previous discussion. Furthermore, the below tables will refer to lags. These are lags as explained in the discussion of dynamic models, i.e., the inclusion of last year’s value of a specific variable instead of the current value. Thus, the specifications (1) and (2) are static, while (3) and (4) are dynamic.

Patenting in Lighting Technology Lighting is a significant cause of energy consumption, yet it constitutes aa rather moderate share of consumer budgets. In order to explore the impact of energy prices on patenting in lighting technologies, we employ the country panel data and regress the logarithm of coun-try-level patent counts on energy prices and taxes (both residential and industrial), public R&D in the underlying technologies, and dummy variables indicating both the announce-ment and the enactment of the Directive on the Energy Performance of Buildings (2002/91/EC).


73

Determinants of Patenting in Lighting Technology (1) (2) (3) (4) Ln price electricity (residential) -0.397 (0.275) Tax share (residential) 1.010 (0.867) Ln price electricity (industrial) -0.418*** -0.225 -0.253 -0.163 (0.158) (0.227) (0.246) (0.188) Tax share (industrial) 1.603*** 1.053 1.672 1.526* (0.615) (0.721) (1.313) (0.916) Ln EPO patents 0.621*** 0.638*** 0.251 0.425** (0.206) (0.213) (0.188) (0.183) ln public R&D (industrial) 0.0718* 0.0758* 0.00823 0.0304 (0.0424) (0.0425) (0.0454) (0.0545) ln public R&D (residential) 0.0404 0.0355 0.0210 -0.00508 (0.0438) (0.0440) (0.0476) (0.0460) EU Directive 2002/91 ann. 0.284*** 0.319*** 0.234** 0.251*** (0.100) (0.103) (0.0947) (0.0909) Lag tax share (industrial) -1.088 -1.032 (1.735) (1.009) Lag ln price electricity (industrial) -0.224 -0.327 (0.220) (0.235)

Lag ln patents lighting

0.314*** (0.0549)

0.124**(0.0571)

Constant 1.800*** 1.554*** -0.0129 (0.471) (0.522) (0.0171) Observations 233 233 224 207 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1


74

Patenting in LED Technology Light emitting diodes constitute a major technical breakthrough and have established a promising lighting technology in their own right. We test here if patenting in this technolo-gical segment is determined by the same forces as lighting technology in general. The vari-ables correspond to those used for lighting technologies.


75

Determinants of Patenting in LED Technology (1) (2) (3) (4) Ln price electricity (residential) -0.248 -0.528 0.493 0.499 (0.461) (0.591) (0.940) (0.489) Ln price electricity (industrial) 0.0357 (0.455) Tax share (residential) -0.509 -2.797 -0.174 0.372 (1.225) (1.763) (1.681) (1.928) Tax share (industrial) 1.592 (1.476) ln public R&D (industrial) 0.00997 -0.00563 0.0475 0.0541 (0.0797) (0.0786) (0.0964) (0.111) ln public R&D (residential) 0.161** 0.156* 0.0626 0.0439 (0.0802) (0.0792) (0.0733) (0.0780) EU Directive 2002/91 ann. 0.282 0.325* -0.0407 -0.0556 (0.180) (0.184) (0.171) (0.166) EU Directive 2002/91 intr. -0.128 -0.401 -0.0527 -0.0116 (0.220) (0.256) (0.167) (0.193) Lag tax share (residential) 0.357 -2.099 (2.144) (1.611) Lag ln price electricity (residential) -0.773 -1.061** (0.998) (0.540) Lag ln patents LED 0.560*** 0.453*** (0.0601) (0.120) Constant -0.907 -1.012 (0.801) (0.930) Observations 199 194 193 181 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1


76

Patenting in Boiler Technology The relevant price variable for driving incentives in the area of boiler technology could either be the price for natural gas (or other gas fuels) as well as the price for light fuel oil. However, the results for light fuel oil have slightly better economic interpretation and we choose to present these here. Light fuel oil costs simply appear to have a bigger effect than natural gas. The public R&D variable is not significant in these regressions.


77

Determinants of Patenting in Boiler Technology (1) (2) (3) (4) ln price light fuel oil (residential) 0.0999 0.0999 -0.140 -0.0788 (0.0948) (0.0948) (0.104) (0.125) tax share (residential) 0.951* 1.421*** 1.303*** 1.303*** (0.522) (0.274) (0.340) (0.340) ln EPO patents -0.235 -0.235 -0.348 -0.392** (0.198) (0.198) (0.212) (0.165) ln public R&D (residential) -0.0130 -0.0130 0.000391 -0.0338 (0.0362) (0.0362) (0.0291) (0.0325) lag ln price light fuel oil (residential) 0.223 0.0907 (0.140) (0.135) lag tax share (residential) -0.0997 0.0582 (0.634) (0.542) lag ln patents boilers 0.635*** 0.246*** (0.0557) (0.0643) lag ln price light fuel oil (residential) 0.223 0.0907 (0.140) (0.135) Constant 1.606** 1.606** (0.637) (0.637) Observations 276 276 266 251 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Patenting in Biomass Technology in Buildings We take again the prices for light fuel oil and for natural gas as the potentially relevant en-ergy costs which determine incentives for investments and patenting in biomass technology. Again, the price for light fuel oil yields the most intuitive results, so only these are included.


78

Determinants of Patenting in Biomass (1) (2) (3) (4) ln price light fuel oil (residential) 0.318** 0.318** 0.709*** 0.719*** (0.135) (0.135) (0.237) (0.246) tax share (residential) -1.957*** -1.957*** -1.172 -1.168 (0.569) (0.569) (0.783) (0.800) ln EPO patents 0.114 0.114 0.554** 0.555 ln public R&D (residential) -0.0351 -0.0351 -0.0208 -0.0202 (0.0580) (0.0580) (0.0597) (0.0658) lag ln patents biomass 0.358*** 0.313*** (0.0680) (0.0757) lag ln price light fuel oil (residential) -0.498** -0.491*** (0.201) (0.146) lag tax share (residential) -0.422 -0.514 (0.925) (0.354) Constant 1.121 1.121 0 (0.916) (0.916) (0) Observations 188 188 182 175 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Patenting in Ventilation Technology Relevant prices for patenting in ventilation technologies are again for electricity (residential and industrial). For this technology group, the variables capturing the timing of the directive 2002/91 are slightly significant together with one of the country-specific R&D variables.


79

Determinants of Patenting in Ventilation Technology (1) (2) (3) (4) ln price electricity (industrial) -0.387* -0.431 -0.314 -0.291 (0.197) (0.272) (0.230) (0.394) tax share (industrial) 1.777** 1.849** 0.718 0.699 (0.798) (0.889) (1.024) (0.699) ln EPO trend -0.296 -0.305 -0.336 -0.356 (0.285) (0.289) (0.371) (0.270) ln public R&D (industrial) -0.0159 -0.0134 -0.0171 -0.0179 (0.0497) (0.0507) (0.0627) (0.0331) ln public R&D (residential) 0.142*** 0.144*** 0.130*** 0.132* (0.0479) (0.0490) (0.0476) (0.0701) EU Directive 2002/91 ann. 0.134 0.126 0.111 0.113* (0.121) (0.125) (0.104) (0.0685) EU Directive 2002/91 impl. -0.274 -0.273 -0.285 -0.279 (0.179) (0.181) (0.214) (0.213) lag ln patents ventilation 0.0927 0.0503 (0.0753) (0.0636) lag tax share (industrial) 1.431 1.550 (1.210) (1.608) lag ln price electricity (industrial) -0.168 -0.215 (0.223) (0.533) Constant 0.245 0.217 0.0246 (0.641) (0.716) (0.0227) Observations 190 190 184 175 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1


80

Patenting in Fuel Efficiency for Motor Vehicles Since a large number of patents in this field come from car and truck manufacturers in Germany, France and Italy – with considerable investments in diesel technology for passen-ger cars and trucks – the price for diesel fuel is the first relevant price variable. Also this price variable is the only with consistent data coverage over time. The selection of the appropriate gasoline fuel prices is complicated by the fact that leaded gasoline became replace by unleaded variants in the course of the 1980s and 1990s. Thus, we focus exclusively on automotive diesel prices, remembering that most car fuel prices move very much in tandem over longer horizons. Determinants of Patenting in Motor Fuel Efficiency (1) (2) (3) (4) ln retail price diesel fuel 0.223 0.223 0.392 0.484 (0.206) (0.206) (0.308) (0.329) tax share 2.132*** 2.132*** 1.110 1.826 (0.352) (0.352) (1.162) (1.193) ln public R&D in transportation 0.125*** 0.125*** 0.0463 0.110* (0.0447) (0.0447) (0.0502) (0.0579) lag ln patents motor fuel efficiency 0.667*** 0.399*** (0.0736) (0.137) lag tax share -0.380 -1.798 (1.137) (1.186) lag ln retail price diesel fuel -0.365 -0.467** (0.251) (0.207) Constant 1.936*** 1.936*** 0.0238*** (0.203) (0.203) (0.00907) Observations 193 193 186 176 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Patenting in Pulp and Paper Manufacturing In 2008, the European Commission announced the Best Available Techniques Reference Documents (BREFs) - IPCC Directive which targeted explicitly the pulp and paper manu-facturing sector. Since the data series used here end with year 2008, the impact of this direc-tive cannot be ascertained from data presently available.


81

Determinants of Patenting in Pulp and Paper Manufacturing (1) (2) (3) (4) ln price electricity (industrial) 0.352*** 0.352*** 0.188 0.184 (0.133) (0.133) (0.239) (0.285) tax share (industrial) -0.547 -0.547 -0.851 -1.181** (0.485) (0.485) (0.846) (0.590) ln EPO patents 0.320** 0.320** 0.165 0.220* (0.158) (0.158) (0.164) (0.118) lag ln patents paper and pulp 0.475*** 0.0842 (0.0466) (0.0856) lag tax share 0.700 0.306 (0.911) (0.612) lag ln price electricity (industrial) 0.0573 0.155 (0.197) (0.114) Constant 4.264*** 4.264*** 0.0143 (0.342) (0.342) (0.0103) Observations 298 298 285 273 Standard errors in parentheses *** p<0.01, ** p<0.05, *p<0.1


82

The model is closely linked to the model developed in Fisher and Newell (2008). In fact, the CERIM is a simplification of their model in order to focus exclusively on relevant policy op-tions. A more thorough model description can be found in the original paper. The CERIM includes two subsectors, one emitting (fossil fuels) and one non-emitting (re-newables), and both are assumed to be perfectly competitive and supplying an identical product, electricity. Fossil-fuelled production includes a CO2-intensive technology (coal) that operates primarily as a base load and a lower-emitting technology (gas turbines) that dominates at the margin. To the extent that renewable energy is competitive, it displaces marginal fossil-fuelled generation. We therefore treat nuclear- and hydro-based generation as fixed in response to the range of policies we model, a reasonable assumption based on other detailed models (see the Numerical application). The model has two stages, each represent-ing a specific number of years. Electricity generation, consumption, and emissions occur in both stages, while investment in knowledge takes place in the first stage and, through tech-nological change, lowers the cost of renewables generation in the second. An important as-sumption is that firms take not only current prices as given, but also take prices in the sec-ond stage as given, having perfect foresight about those prices. To allow for consideration of the length of time it takes for innovation to occur, and for the lifetime of the new technologies, let the first and second stages be made up of n1 and n2 years, respectively. For simplicity, we assume that no discounting occurs within the first stage; this assures that behaviour within that stage is constant. However, let δ represent the discount factor between stages. It is possible to allow for discounting within the second, longer stage by altering n2 to reflect such a discounting; in that case n2 can be thought of as ‘‘effective’’ years. The emitting sector of the generation industry, denoted with superscript F, relies on two fos-sil fuels for production: coal, x, and natural gas, y. Total output from the emitting sector is

in year t. Total emissions from this sector equal

as each fuel has a fixed CO2 intensity. Marginal production costs coal-fired generation is given by

/2 and similarly for natural gas. The opportunities for CO2 abatement in electricity rely largely on fuel switching; although coal gasification or generation efficiency improvements are op-tions, they tend to explain little of the predicted reductions in climate policy models. One

APPENDIX B: DESCRIPTION OF CERIM


83

policy affects the fossil-fuelled sector directly: an emissions price/tax denoted by τt . This gives rise to an intertemporal profit function for the emitting firms:

We assume profit maximisation. Another sector of the industry generates without emissions by using renewable resources (wind, for example). Annual output from the renewables sector is qt. The costs of production, G Kt,qt , are assumed to be increasing and convex in output, and declining and convex in its own knowledge stock, Kt, in the following way:

, 2

Note that we have simplified considerably by assuming there is technological change in the relatively immature renewable energy technologies, but none in the relatively mature fossil-fueled technologies. While it is not strictly true that fossil-fueled technologies will experience no further advance, incorporation of positive but relatively slower innovation in fossil fuels would complicate the analysis without adding much additional insight. The knowledge stock is a function of cumulative knowledge from R&D, Ht, and of cumulative experience through learning by doing, Qt. Cumulative R&D-based knowledge increases in proportion to annual R&D knowledge generated in each stage, ht, so

Cumulative experience increases with total output during the first stage in a completely simi-lar manner

Research expenditures, R ht , are increasing and convex in the amount of new R&D knowl-edge generated in any one year, with

The strictly positive marginal costs imply that real resources—specialized scarce inputs, em-ployees, and equipment—must be expended to gain any new knowledge. As a partial equi-librium model, we do not explicitly explore issues of crowding out in the general economy, but those opportunity costs may be reflected in the R&D cost function.


84

The second policy instrument in the model is a subsidy targeted at R&D in renewable. The subsidy, σ, is implemented as the government offsetting a share research expenditures. Now, we can write up the intertemporal profit function for the renewables sector:

1 , 1 2 , The model is closed by the standard demand equals supply condition:

Where demand is assumed iso-elastic. All parameter values can be found in Fisher and New-ell (2008).


85

Microsoft Office Excel 97-2003 Worksh

APPENDIX C: DESCRIPTION OF IPC CODES

1139:03 RECS International

PPPPROMOTING COROMOTING COROMOTING COROMOTING CO----OPERATION OPERATION OPERATION OPERATION

MECHANISMS TO REACH MECHANISMS TO REACH MECHANISMS TO REACH MECHANISMS TO REACH RESRESRESRES TARGETSTARGETSTARGETSTARGETS RECS POLICY FORUM | 31 MARCH


2

COLOPHON

Author: Project manager and partner H.S. Næss-Schmidt, Analyst Hanna Hedlund

Client: RECS International

Date: 25thMarch 2011

Contact: SANKT ANNÆ PLADS 13, 2nd FLOOR | DK-1250 COPENHAGEN

PHONE: +45 2333 1810 | FAX: +45 7027 0741

WWW.COPENHAGENECONOMICS.COM


3

Executive summary....................................................................................................... 4

Chapter 1 The big picture requirements and problems ............................................ 5

1.1. EU strategy is failing in two major dimensions ...................................................... 6

1.2. How to set the strategy on the right track .............................................................. 9

Chapter 2 Use of co-operation mechanisms: the NO-SE example .......................... 10

2.1. Key requirements for operating a joint TGC scheme ........................................... 10

2.2. Benefits of building on Norway-Sweden .............................................................. 12

Chapter 3 Offshore wind ....................................................................................... 15

Chapter 4 Innovation, investment climate and external competiveness .................. 20

4.1. Boosting innovation ............................................................................................ 20

4.2. Investment climate .............................................................................................. 22

4.3. The industrial perspective: International competiveness ....................................... 24

References .................................................................................................................. 25

TABLE OF CONTENTS


4

The EU and its member states have adopted ambitious long term climate change and energy

security policies including the goal to reduce GHG emissions by 2020 by (at least) 20 per

cent relative to its 1990 level and together with all developed countries a reduction of 85 to

90 per cent by 2050. This requires a massive roll-out of investments and deployment of low

carbon technologies. Relying on more energy efficiency alone would entail a drastic reduc-

tion of living standards. The EU’s directive on renewable energy, setting a target for renew-

able energy of 20 per cent in 2020, should be seen in this long term perspective. The big

question is how? This study has two main and highly linked recommendations:

Strengthening the internal market. Strengthening the internal market. Strengthening the internal market. Strengthening the internal market. While the directive on the renewable energy allows

member states to comply jointly with their national renewable energy targets, very few coun-

tries have so far shown any inclination to move in that direction. We find that this is a very

costly strategy for EU for mainly three reasons. It adds further to the already substantial na-

tional compliance costs, it seriously increase investment risks for firms that operate under 27

constantly changing national support regimes, and finally it deprives EU producers of a

strong and large domestic market that can provide a spring board to compete against main

competitors in China, US, and other major regions. In practical terms we propose the fol-

lowing main actions:

• Norway and Sweden Norway and Sweden Norway and Sweden Norway and Sweden is about to establish a joint support schemes based on so-

called Green Certificates to comply jointly with their targets under the renewable

energy targets. The two countries have expressed an explicit interest in extending

their co-operation to other countries and have the potential to expand relatively

low cost production. This initiative could spearhead a more internal market based

approach to target compliance, a potential that should be pursued.

• Expansion of offExpansion of offExpansion of offExpansion of off----shore windshore windshore windshore wind is projected to deliver a substantial part of the in-

crease in renewable electricity production. We find that a voluntary co-operation

among the key interested parties – primarily in the North and Baltic Sea – could

deliver substantial benefits to energy companies, consumers and tax payers. The key

aim of such a co-operation should be that: (1) the location should be based on un-

derlying potential not the generosity of subsides; (2) an improved framework for

managing and financing the required expansion of the grid system.

Shifting over time from subsidies to deployment to support for Shifting over time from subsidies to deployment to support for Shifting over time from subsidies to deployment to support for Shifting over time from subsidies to deployment to support for innovationinnovationinnovationinnovation. A massive

reduction in green house gases requires a substantial increase in public support for innova-

tion. EU needs major leaps of technological progress, not just marginal improvements of ex-

isting technologies. Hence higher carbon prices need to be supplemented with targeted

technology policies. In our view there is though a risk that public support over the coming

decade seem to be directed too much towards deployment in large quantities of immature

and hence very costly technologies while more targeted public support for basic and applied

research, demonstration projects, is but a fraction of deployment subsidies. This is the wrong

order of priorities.

The two recommendations are strongly linked in practice. If member states co-operate on

joint compliance, billions of Euros can be saved and used to boost focused innovation poli-

cies. The present fiscal crisis in EU in particular suggests the need for such a switch.

EXECUTIVE SUMMARY


5

EU has committed itself to massive long term reductions in greenhouse gases to address cli-

mate change requiring a combination of energy savings and expansion of low/zero carbon

energy sources. As recently as February 2011 at the European Council, EU confirmed its

target of having GHG (Greenhouse Gas) emissions reduced in 20501 by 80 to 95 per cent

relative to 1990. By 2020 the target is to reduce GHG emissions by 20 per cent (or 30 per

cent if part of a global effort) while also increasing the share of RES (Renewable Energy

Sources) in final energy consumption to 20 per cent by 20202.

To reach the targets, the EU has pledged to cut energy related greenhouse gas emissions

(CO2) by 20 per cent in 2020, relative to 1990 emissions, and by 82 per cent in 2050, cf.

Figure 1.1. Long term cuts in that order will make massive deployment of RES viable as a

means to reduce GHG in general and energy related emissions in particular. Such a reduc-

tion in energy consumption from fossil fuels will also require a massive increase in carbon

prices: from an estimated € 17-30 per ton in 2020 to € 117 per ton in 2050 cf. Figure 1.1.

Figure 1.1 Energy related CO2 emissions (GHG emissions) and prices in EU

Note: Pledged target 2030 and 2050 is represented by the level needed to reach the 2 degree target, stated but not pledged. Business as usual 2050 is calculated assuming that energy related CO2-emissions will increase as much as all greenhouse gases.

Source: EU Commission (2010), SEC (2010) 650, p 39-45. Capros et al. (2008) p 2. OECD (2009).

However, the GHG reduction target is far too low to drive up carbon prices consistent with

prices needed to reach the 2020 renewable energy target. The power price from coal plants,

the most likely alternative to renewable energy, in 2020 is expected to be just below € 40 per

1 The 2050 objective is a reduction of 80 to 95 per cent by developed countries as a group. The 2020 and 2030 tar-get is for EU specifically. 2 Europen Council, Conclusions on Energy, 4 Februray 2011, European Council, Transport, Telecommunications and Energy, Press release, 28 Feb 2011, 6950/11

0

20

40

60

80

100

120

140

0

1000

2000

3000

4000

5000

6000

7000

2005 2020 2030 2050

CO2 price(€ per ton)

GHG emissions (Gt) Cuts

additional cut to reach 30% target, 2020

Pledge

CO2 price (OECD)

CO2 price (EU)

CO2 price ( 30% target 2020)

Chapter 1 THE BIG PICTURE REQUIREMENTS AND PROBLEMS


6

MWh 3 (with GHG emission reduction target of 20%). Looking at renewable combined

heat and power generation which has the least cost options, at that price only 25% of the re-

newable supply needed to reach the target4 is economically viable. A 30 per cent reduction

target for the ETS sector will lift power market price to just over € 60, and roughly 40 per

cent of the RE target can be met, cf. Figure 1.2.

Figure 1.2 Stylised supply curve for output from power generation in EU by 2020

Source: Eurelectric (2008) and Copenhagen Economics calculations

Assuming the most cost-efficient EU roll out of renewable energy supply in the power gen-

eration sector, a price around € 100 is required to make enough RE power production viable

without specific support, cf. Figure 1.2. Hence, support to RE will be required, an issue we

discuss in the following section.

1.1. EU STRATEGY IS FAILING IN TWO MAJOR DIMENSIONS While reaching the 2020 RE target will require very substantial subsidies in any case, the

EUs implementation strategy is failing in two main dimensions.

First, there is too much focus on domestic compliance driving up costs further. The EU

RES directive allows member states to reach target compliance by trading credits by various

means, cf. Box 1.1 below, but very few countries have so far shown any inclination to do so.

As a consequence, some estimations indicate that overall costs of reaching targets may go up

further € 8-15 billion per year by 2020 as a result5.

3 Power price from coal plants at 20 and 30 per cent reduction of GHG is based on a carbon price of € 30 and € 17 per ton, cf Figure 1.1. The carbon price is added to the electricity price from a marginal coal power plant, assuming cost effectiveness. 4 To achieve 20% renewable energy in 2020, it is estimated that 1069 TWh electricity and heat will be needed, Eurelectric(2008), p 13. 5 Eurelectric (2008), EWI (2009). A report focusing on joint implementation of renewable energy targets in the Baltic regions, also find sizeable gains from such trade (BDF(2009)). EU Impact assessment (2008)

0

50

100

150

200

250

300

0 200 400 600 800 1000 1200 1400 1600 1800 2000

€/MWh

TWh

Biomass heat

Geothermal

Hydro

Heat pumps

Onshore wind

Biowaste

Biomass

Offshore wind

Biogas

Solar

Tidal stream

Wave

Power price, 30% target

Power price, 20% target

RES TargetRES TargetRES TargetRES Target (electricity/heat)(electricity/heat)(electricity/heat)(electricity/heat)


7

Box 1.1 Co-operation mechanisms The EU directive has set out four co-operation mechanisms:

1. Statistical transfer 2. Joint projects among EU Member States 3. Joint support systems 4. Joint projects among EU Member States and third countries

Statistical transfersStatistical transfersStatistical transfersStatistical transfers Member states that exceed their target can transfer (sell) target accounting units to countries that miss their target. Selling accounting units is only allowed if it does not affect the target of the selling country. The transfer must be reported to the commission no later than 3 months after the end of the year for which they are valid. Joint projects among EU Member StatesJoint projects among EU Member StatesJoint projects among EU Member StatesJoint projects among EU Member States Two or more member states (and private operators) can share the financing of a project and share the real-ised target units. They have to specify the proportion or amount of electricity, heating or cooling from the project that will count towards the target of each member state and report this to the commission no later than 3 months after the end of the year for which the energy units are valid. Joint support systemsJoint support systemsJoint support systemsJoint support systems Joint support systems require a high degree of cooperation between countries. Examples are feed-in tariffs or trade in green certificates. A certain amount of energy produced in one country may count towards the target for another country (statistical transfer or a distribution rule). A distribution rule must be reported to the commission no later than 3 months after the end of the first year it is in effect Joint projects among EU Member StatesJoint projects among EU Member StatesJoint projects among EU Member StatesJoint projects among EU Member States and third countriesand third countriesand third countriesand third countries Basically the same as “Joint projects among EU Member States” but including non-EU Member States.

Source: EU Directive 2009/28/EC, Article 6 (1)

Secondly, there is an imbalance between costs associated with deployment of RES technolo-

gies over the coming 10 years running up to around € 100 billion per year by 2020 and the

relatively speaking miserly low level of public funding for research, development and innova-

tion which by 2009 at EU level reached roughly € 3 billion per year cf. Table 1.1.

Table 1.1 Current RDI funding and estimated costs and required RES investments 2020

Type of cost

Current cost or

investment,

per year (RDI)

Estimated deployment cost

or investment, 2020.

WITH internal market

Estimated deployment cost or

investment, 2020.

WITHOUT internal market

Cost to reach RES target (ECOFYS)

€ 97-105bn

Cost to reach GHG target (EU)

€ 48bn *

Cost to reach GHG and RE target (Eurelectric)

€ 80 bn (incl € 4bn to reach RE target)

€ 97 bn (incl € 21bn to reach RE target)

EU investment in and support to RDI

€3.26bn*

Required capital invest-ment in RE (per year up until 2020)

€ 35bn € 62bn € 70bn

Note: *Cost to reach target represents an additional energy cost, not a reduction of GDP. Including additional investments and energy savings

** Per year 2007-2009. Source: Cost to reach target: European Parliament COM(2010) 265 final, May 2010 p 3. Eurelectric/ Pöyry p 19.

Investment and support to RDI: EU Commission SEC(2011) 131final, p 4. Capital Investments: EU Commission COM(2011) 31 final p7-8 and SEC(2011) 131 final p 4.

Indeed, the scale of the challenge of meeting the global challenges of climate change, and the

need to bring new technologies to the market, requires a massive increase in RDI. Known

technologies such as biomass and wind power will become increasingly viable over the com-


8

ing years as carbon prices increase and on-going progress at the same time bring down de-

ployment. However, there are limits on how far such sources can be expanded: biomass due

to physical supply constraints, wind power due to its intermittence that require other con-

trollable energy sources to provide balance in power due to errors in day-ahead forecasting

and back-up due to volatility in production6. As a result, we need research that can bring

new technologies to the forefront, a job that private investors cannot provide alone even with

much higher incentives from carbon prices given the huge uncertainties involved and the

substantial public benefit nature of such R&D7. Based upon this reasoning, OECD has pro-

vided model calculations that suggest that public R&D investments should increase from

their present 0.04 per cent of GDP in 2009 to perhaps 0.12 per cent in 2020 cf. Figure 1.3.

Figure 1.3 OECDs call for substantial increases in public R&D support

Source: IEA data services, Bosetti et al (2009, EIA home page, Copenhagen Economics calculations

This number of 0.12 per cent is however not a huge amount compared to historical spend-

ing before late 80’ties.

6 Producing (growing) biomass is constrained by limited natural resources (land), ecological conditions and com-petes with other land use. EWI (2009) p. 66, and European Biomass Industry Association. 7 A new innovation may create positive spillovers to other firms and the rest of the economy since innovations can be improved, standardized and create the basis for new technology classes. But the positive effects are not fully ap-propriated by the company financing the research. A tax that that lowers emissions might not increase the return on R&D investments enough to make investments high enough to achieve the economically optimal level (where spill-over effects are taken into account). An EU study concludes that the combination of rising carbon prices and up-front R&D support for green tech-nologies yields the most favourable outcome. It also concludes, similar to an OECD study, that R&D support in the long run should not favour green technologies specifically, compared to all sectors of the economy. Copenhagen Economics (2010b)

0,00%

0,10%

0,20%

0,30%

1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 2018

OECD, RD investment in energy, per cent of GDP, 1974-2020

RD investment in energy, per cent of GDP, historical

RD investment in energy, per cent of GDP, to reach the 2 degree scenario


9

1.2. HOW TO SET THE STRATEGY ON THE RIGHT TRACK We identified two main failings in the present EU strategy namely too little focus on the in-

ternal market as a driver for cost-efficient deployment and development of low carbon tech-

nologies, and too little focus on fostering long term innovation and investment in such tech-

nologies.

As the core of this project is a discussion about choice of support mechanisms for deploy-

ment of RES resources within the context of the existing RES directive, as described in Box

1.1. The task is to put this choice in the perspective of the larger challenges. In doing so we

will favour a pragmatic, non-dogmatic approach that essentially accepts that progress will

depend on a voluntary cluster of countries seeing the benefit of co-operating on RES de-

ployment rather than relying exclusively on own resources for target compliance. The very

fact that so few countries have so far been inclined to go down that route suggests that it

needs to be demonstrated that this is a doable strategy and a way can be found to deal with

the objections earlier raised by member states and evaluation studies against the use of inter-

nal market based mechanisms for target compliance such as Tradable Green Certificates

(TGC).

Hence we have divided the rest of the report into three chapters:

• Chapter two on Norway’s and Sweden’s TGC coChapter two on Norway’s and Sweden’s TGC coChapter two on Norway’s and Sweden’s TGC coChapter two on Norway’s and Sweden’s TGC co----operation:operation:operation:operation: the two countries

have signed a principal agreement for joint implementation of their RES targets.

What are the main ingredients in this agreement and what could be the benefits of

extending it to more countries as a practical application of a “joint support scheme”

under the RES directive?

• Chapter three on Chapter three on Chapter three on Chapter three on offshoreoffshoreoffshoreoffshore windwindwindwind: Massive increases in production are projected

but location of production is at substantial risk of being driven more by differences

in generosity of support schemes than underlying cost advantage. What is the main

scope and benefits of co-operation on deployment of offshore wind be, not the

least as a practical application of the “joint project” mechanism within the RES di-

rective?

• Chapter four on TGC and the innovation and investment climate: Chapter four on TGC and the innovation and investment climate: Chapter four on TGC and the innovation and investment climate: Chapter four on TGC and the innovation and investment climate: TGCs are

in evaluation studies often held to be inferior to feed-in support schemes because

they are seen as failing to provide support to innovation by being focused on ma-

ture RES technologies. It is argued that TGCs are less friendly to investors given

perceived less certain returns on investments and providing excess returns to low

cost RES producers. We suggest that these concerns are largely misplaced given

proper overall policy design.


10

We listed in chapter one that there were three different approaches to using the co-operation

instruments under the present RES directive: (1) statistical trading (2) joint support schemes

and (3) joint projects. We will discuss joint projects largely in the context of offshore wind

in chapter 4.

It is our conclusion that pure statistical trading without underlying market organisation and

price formation lacks credibility; a buyer country is faced with risk of too low seller surplus.

Harmonised feed-in tariffs could deliver within technology neutrality and hence allow com-

parative advantage in location to dominate, but may de facto be more difficult to agree on

than TGC since it requires explicit agreement on subsidy rates for a wide range of technolo-

gies.

Hence we focus our attention on the Norwegian-Swedish TGC market as a means of using

co-operation instruments for two reasons. First, the principal reasons listed above, which es-

sentially are that TGC is the instrument designed to exploit low cost deployment options in

an internal market context. Second, the Norway-Sweden agreement is so far the only pro-

spective “game in town”, no other countries are anywhere as far in doing joint compliance

among EU and EEA countries.

2.1. KEY REQUIREMENTS FOR OPERATING A JOINT TGC SCHEME The preparatory work for setting in motion a joint scheme for TGC between Norway and

Sweden has provided a good testing ground for determining if key principles of such co-

operation can be implemented in practice. In turn, these lessons are partly based upon the

almost decade long experience Sweden has had in operating its own TGC certificate schemes

since its inception in 2003.

The three prime objectives when setting up a joint TGC-market (trade in green certificates)

is to (1) avoid distortions (2) have a transparent and smooth investment climate, and (3) ob-

tain stabile prices for producers of renewable energy (relative to other energy producers)8.

We will discuss important issues that need to be agreed upon in a joint TGC market. The is-

sues are split into two levels of “urgency of priorities”: “must” and “should”.

Must have

First of all there must be a well defined framework for determining a quota curve going for-

ward. Such an agreement must define, over a foreseeable period, the level of buyer obliga-

tions to be reached, and specific rules for lowering/raising ambitions going forward. The

point of this is to provide predictable signals to investors both in terms of overall expected

deployment and prices. The experience from Sweden suggests that a well managed process

can deliver very stable prices to the markets, hence prices have moved between 20 and 30 €

per MWh since the opening of the market in 2003, cf. Figure 2.1.

8 Swedish Energy Agency and Copenhagen Economics

Chapter 2 USE OF CO-OPERATION MECHANISMS: THE NO-SEEXAMPLE


11

Figure 2.1 Prices and buyer obligations for Swedish TGC, 2003-2020

Note: i) Exchange rate 9 SEK/€ used. ii) After 2020 the obligations fall, reaching 0 after 2035. This is because the TGC is a temporary subsidy. After 2035 the market is supposed to supply enough green electricity on its own. iii) Obligations planned for the Swedish system. Obligations for the Norwegian-Swedish TGC not yet set, but projected to follow a similar path,

Source: Swedish Energy Agency – Elcertifikat Kvotpliktig, Svenska Kraftnät – Cesar.

As the quota system is based on buyer obligations it needs to ensure full coverage and non-

leakage. All suppliers of electricity have an obligation to deliver a set amount of green certifi-

cates (supply to the electricity intense industry can be exempted)9. If one of the suppliers is

left outside the system, it gets cheaper costs and thus an unfair advantage compared to other

suppliers.

To achieve stable prices a long term perspective and banking options are required. Swedish

obligations are already set up until 2035 as a yearly percentage of electricity use, which pro-

vides stability for investors. Companies that have acquired certificates in excess of their obli-

gations can bank or sell them. This provides sellers and buyers of green certificates a chance

to counteract fluctuations on the market.

A joint control function is also important to make sure that the same rules apply and the cer-

tificate has the same value in all countries. This ensures a transparent and well functioning

market.

Should

Rules such as what technologies to be included, length of support period, rules for updating

certificates for existing installations running out of support period should be harmonized as

much as possible. This also applies for implied network costs, priority access rules, balancing

9 Swedish Energy Agency – Elcertifikat, Kvotpliktig

0

2

4

6

8

10

12

14

16

18

20

0

10

20

30

40

50

60

70

80

90

100

2003 2008 2013 2018

% of supplied electricity€ per MWh

Total price (electricity + certificate) Price certificates

Obligation (right axis) Obligation path (right axis)


12

cost contributions and paying for additional grid investments. Rules do not have to be ex-

actly the same but should be harmonized so that producers do not face different conditions

in different countries and competition distortions are created. Norway and Sweden have

started discussions on this topic, aiming at making conditions as equal as possible10101010.

A further issue to be solved is code of conduct for support outside TGC system: how should

countries deal with pilot and demonstration projects, and R&D support without creating

competition distortions. A good proposal would be to use EU standard state aid rules for

R&D rather than very weak rules applied for support to energy technologies.

2.2. BENEFITS OF BUILDING ON NORWAY-SWEDEN Sweden has expressed a willingness to expand the scheme to more countries. Studies show

that there is substantial net additional capacity in the overall NO-SE system, mainly in hy-

dro and wind power11. Options for countries that want to join are to either fully join, with

reorganisation of national schemes, or to piggy back.

An expansion of the Scandinavian TGC market has its first mover advantages and disadvan-

tages. It depends on the supply curve for land based wind power which is estimated to be the

marginal supply element. Early-in as buyers could give low prices but also facing more teeth-

ing problems.

The estimated potential and cost for onshore wind energy in Norway and Sweden in 2020

(in addition to existing wind power in 2008) are shown in Figure 2.2. The two countries are

expecting to expand onshore wind by 16 TWh by 202012, which gives a marginal cost of

around € 75 per MWh, where the cost of the tradable green certificate is approximately € 20

per MWh13. The combined goal will be reached by only utilising a small part of the potential

wind power. The marginal cost curve will be almost flat with an unused potential of almost

80 TWh, cf. Figure 2.2.

10 Swedish Energy Agency (2010) 11 Swedish Energy Agency (2010) 12 NO+SE must together increase their renewable energy by 26.4 TWh by 2020. Norway is estimated to add 7 TWh water power and 1-2 TWh biofuel, Sweden could add 7-8 TWh biofuel and 0.5 TWh water power to exist-ing renewable energy sources. 13 Swedish Energy Agency (2010) pp. 54-62.


13

Figure 2.2 Supply curve 2020, onshore wind, Norway and Sweden

Note: Exchange rate used: 9 SEK/€. Source: Swedish Energy Agency (2010), p. 29.

A number of countries should find the Norway-Sweden TGC prices attractive. According to

estimations14 in 2020 UK will have a marginal cost for renewable energy of approximately €

120-180 per MWh, for the last two steps to reach their national target. Netherlands is esti-

mating a production of around 20 TWh offshore wind at a cost of € 120-135 per MWh to

reach their target and Germany is projected to produce 15 TWh photovoltaics at €180-200

per MWh15 and 65 TWh offshore wind at €120-135 per MWh, cf. Figure 2.3.

The unused potential in the Norway-Sweden system could cover around 60-70 per cent of

the offshore wind and solar PV in the UK, Germany and Netherlands, as illustrated in Fig-

ure 2.3. Co-operation with Norway and Sweden would allow the three countries to meet

their target compliance at a lower cost than with national actions only.

14 ECN (2010) p. 22 , HM Government (2009) p. 40 15 Estimated future PV price in requires large cost decreases from current levels. Average production cost in Germa-ny in 2009 was €550 per MWh, spanning from € 250 to € 1050 (Ecofys p. 18). The subsidy rate (feed-in-tariff) for PV in 2009 was € 430 per MWh compared to a subsidy of € 92 per MWh to onshore wind (RWI p. 6).

50

60

70

80

90

100

110

0 20 40 60 80 100

TWh

€ per MWh

ProjectedexpansionNO+SE


14

Figure 2.3 Potential excess supply NO+SE and marginal cost for energy to reach RES-E

targets in UK, NL and DE, 2020

Note: Exchange rates used: 9 SEK/€, 1,2 £/€ Source: Swedish Energy Agency, ECN and (UK source), and Copenhagen Economics calculations

As marginal production costs in Sweden and Norway for the excess potential are only mar-

ginal above present remuneration, we would see the risk of expansion in terms of providing

“excess” profits (wind fall profits to Norwegian and Swedish producers) to be minimal and

in reality without relevance for the buyers in other countries.

70

90

110

130

150

170

190

0 20 40 60 80 100 120

TWh

€ per MWh

Series1 UK marginal cost to reach target

NL offshore wind DE offshore wind, solar PV


15

Projections of electricity generation in North-West Europe show a massive increase in off-

shore wind capacity as a response to EU and national ambitions to replace fossil based elec-

tricity generation with renewable energy. Rising from a very modest share of 0.1 per cent of

power generation in 2010, offshore wind is set to rise to 5 per cent in 2020 and 10-13 per

cent in 2030 cf. Figure 3.1, competing with biomass based production as the main source of

expansion.

Figure 3.1 Share of biomass and offshore wind in total and renewable based power gen-

eration, 2008, 2020 and 2030

Note: Shaded area represents difference between estimated minimum and maximum. Average share stated.

Biomass only from EWI. Source: EWI (2010) and EWEA (2009)

Offshore wind production is a potentially attractive candidate for “joint projects”. The point

is that offshore wind farms are by nature placed in the sea between member states, suggesting

that coordinated decisions about grid investments and their financing as well as the financ-

ing of the actual installation could be productive. Moreover, financing from companies or

3%

8% 8%

0%

5%

11 %

0%

2%

4%

6%

8%

10%

12%

14%

2007 2020 2030

Windpower and biomass share of power generation

Biomass Windpower

14%

21%

16%

0%

10%

21 %

0%

5%

10%

15%

20%

25%

30%

2007 2020 2030

Windpower and biomass share of renewable energy

Chapter 3 OFFSHORE WIND


16

countries not adjacent to the physical location of the installation could also be foreseen: the

purpose for them would clearly be to get credits under national RE targets for the wind

power produced. It could for example be a retail operator with a domestic national compli-

ance target which could be off-set with such wind energy production.

However, there is also relatively widespread consensus that such projections and opportuni-

ties come with substantial caveats and raise a number of concerns for policy makers as well as

investors.

While more or less model based, projections typically take relatively little account of the

pressures such an increase in a volatile energy source put on absorption capacity in power

markets (or it is assumed these will be solved by massive expansion of grid investments and

improvement in demand and supply management).

The absence of an agreed EU support framework implies that support schemes differ sub-

stantially in levels of generosity, structure of support, suppliers’ contributions to pay for bal-

ancing costs as well as required grid investments, likely leading to substantial distortions to

competition. And further, the absence of an agreed EU support framework also entails risk

of “national competition” in subsidy schemes, especially for offshore wind, as the supply

chain within offshore activities may show huge bottlenecks in the coming years, due to an

expected steep increase in offshore deployment.

Current subsidy rates for offshore wind are very high relative to other renewable energy

sources, reflecting that it has a long way to go before it can compete against lower costs re-

newable energy sources such as biomass, onshore wind and further expansion of hydro

power, as demonstrated by Figure 1.2.

There is a considerable political risk that future subsidy rates may be challenged once actual

deployment in depth starts affecting tax budgets and/or bills for electricity consumers. Cur-

rent deployment of offshore wind accounts for less than 1 per cent of total power generation:

as most subsidies are linked to production not investments, this implies that the financing

cost has yet to find its way to consumers and tax payers.

The current crisis has already triggered reductions in subsidy levels in Germany and Spain

(for solar cells), consideration of moving from fixed to premium feed in tariffs in Germany

and a total freeze in the level of total subsidies to deployment of renewable energy in Nether-

lands16. Policy risks are particularly important for offshore wind over the next decade since

the bulk17 of total remuneration is derived from subsidies on top of the support from the

ETS system.

16 Duurzame Energie Koepel (2010). Bloomberg (2011), Guardian Environment Network (2010) 17 One example is the Danish Anholt offshore wind park that will receive 40 to 70 per cent of its remuneration from subsidies. According to the tendered agreement, the park will receive 1.051 DKK per kWh. If the whole sale electricity price is 0.3 DKK per kWh 70 per cent of the payment will be subsidies, with a market price of 0.6 DKK subsidies will make up 40 per cent of the payment.


17

Certain features make offshore wind deployment costs higher than necessary. One such fea-

ture in particular is the insistence that crediting of supply from installations requires physical

imports into national grid systems. This has been demonstrated both in the case of the pro-

spective Cobra line between Netherlands and Denmark as well as in the discussion of the

Kriegers Flak project. German offshore wind turbines have a right to be directly connected

to the German soil ultimately to be paid by German electricity consumers. Moreover, the

wind farm can only receive German subsidies if the wind farms deliver the electricity physi-

cally on German soil. Hence, in the case of the Cobra line the producers have every eco-

nomic incentive to refuse an offer of building a shorter cable to interconnect with Cobra

rather than a longer and more expensive line to Germany, despite this being likely the best

economic solution.

Now is probably a very good time to act to try to get some order into this investment cli-

mate, focusing on more transparent, reliable and efficient regulation of the offshore wind in-

stallations and the way production is credited against national RE targets:

� While the present RES directive is likely to stand largely unchanged in its main sub-

stance until the first major review in 2014, the fact that so little actual investments has

taken place in the offshore sector suggest that “lock in” of national support schemes has

not completely taken hold.

� The fact that the physical location of installations is basically taken place in the North

and Baltic Sea, requiring new grid systems to be produced, which in turn often requires

joint financing from several countries, suggests that a common framework that defines

levels and structures of support, as well as rules for crediting production against target

compliance, could be of substantial value for investment certainty and transparency.

There are signs that the EU Commission would be willing support such an approach, as can

be seen in the quote below. There is no urgent need that they step forward quickly with any

legally binding proposals or directive. They could in the first place act to facilitate a dialogue

on developing guidelines or good practice in the area, which could become the linchpin for

later legislation. Such an approach would allow the EU commission to play an actual role

early without triggering a broader discussion about revising the EU directive now.

As the EU Energy Strategy notes, a greater convergence of national support schemes to

facilitate trade and move towards a more pan-European approach to development of re-

newable energy sources must be pursued. This repeated call for convergence is not new,

and is to flag up the need to start adjusting to a European electricity market in particu-

lar, where over a third of power will come from renewable energy. In some situations,

such as offshore wind development, the need for an integrated strategy is imminent. In

this instance, the relevant Member States and the Commission are acting, having just

signed the Memorandum of Understanding of the North Seas Countries Offshore Grid

Initiative. The Commission will closely follow this development and will report on the

offshore wind and grid developments in 2012.


18

European Commission18

There might be overlapping interests for energy companies with ambitions to expand their

offshore wind turbine capacity and wind turbine producers. They must weigh the risks of

questioning the current system without knowing what they gain in return from the advan-

tages that a more transparent and (a more) EU regulated system could bring to the stability

and predictability of the overall investment climate. A more active approach, provided it

succeeds in producing the desired solutions, could at the same increase the pressure on TSOs

to bring forward the required grid investments and improved regulatory changes to deal with

increased wind power production.

There are a number of options that could be looked at as possible features of a more com-

mon framework for support for offshore wind mills:

First there is a variety of support instrument that could be considered:

• Tendering approach, forcing private companies to come up with best bid offer (DK ap-

proach)19.

• Feed-in, in principle fixing support levels at estimated lowest production costs in best

location among collaborating countries.

• Green Certificate Instruments for offshore wind. This is by far the most ambitious

schemes as it would require collaborating countries to agree on a commonly agreed tra-

jectory for commitments to offshore wind to create a stable setting for investments and

pricing.

Second more specific performance criteria could be included in award and support criteria.

Offshore wind will continue to produce wind energy at costs well above competing renew-

able energy sources. Major reductions in costs are required to make offshore wind an eco-

nomically viable contributor to EUs generation of electricity. This may suggest that the sup-

port systems may need to be broken up into several parts:

• Potential direct integration of offshore wind production in international green cer-

tificate schemes of near-land, lower cost “offshore” wind with no limits to level of

deployment.

• Moderate to high support to more “bulk” oriented offshore wind production, fo-

cusing on technologies and types of location that allows production costs close to,

or not far ahead of, onshore wind power, with overall support potentially con-

trolled by tendering procedures.

• Higher support levels to projects with promising, but yet far from proven, offshore

wind technologies. Such projects could be said to represent the next step from

18 Communication from the commission to the European Parliament, Jan 2011, Renewable Energy: Progressing towards the 2020 target, COM(2011) 31 final, p11. 19 Off shore wind farms in Denmark are support by a fixed feed-in tariff. However, contrary to support by fixed feed-in of other RES-E in Denmark and also in many other countries the feed-in level is not set politically. The level is found by running an auction, where the bidder who accept the lowest feed-in to settle the generation will win the tender and the right to conduct to project put up for tendering.


19

purely research testing in laboratory conditions to on site, i.e. in the sea, condi-

tions.

• Pure research funding in the pre-deployment phase.

As a whole, we consider it important to choose a mix of instruments that reflect a number of

important characteristics of offshore wind as well as the need to underpin a competitive EU

energy industry over the longer term:

• Different locations present very different generation costs, so broader, potentially

regulated support schemes should shift production in the direction of location of

high load and low maintenance cost locations.

• As many other energy technologies, offshore wind production is not one technol-

ogy. Different generations exist at the same time, and support instruments need to

be tailored to the maturity of the technology.

• The long term perspective: projections of carbon prices consistent with ambitious

global climate policies will dramatically increase the viability of renewable energy

over the coming decade. Intelligent support schemes explicitly recognise such tim-

ing issues.

• The interests of ambitious EU energy companies might best be served by helping

them to develop the competitive solutions for the future rather than focus on rapid

and massive deployment of high cost generations.


20

The choice of support mechanisms for compliance with EU renewable energy targets over

the coming years needs to be narrowly linked to the policy requirements for the next four

decades. As underlined in the introductory chapter 1, expansion of renewable energy be-

comes steadily more important as lower hanging fruits of energy savings are already har-

vested and cuts become ever more binding. Indeed, projections from IEA and OECD show

the decline in the relative contribution from improved energy efficiency towards reduction

of energy related CO2 emissions towards 2030 and the increasing role of CCS and renewable

energy cf. Figure 4.1. Energy efficiency accounts for 46 per cent of the emission reduction in

2020 and 35 percent in 2030 whereas the reduction contribution from renewable energy

and CCS increases from 18 and 4 per cent respectively in 2020 to 25 and 20 per cent in

2030.

Figure 4.1 Contributions to CO2 emissions reduction from energy efficiency, nuclear

power, CCS and renewable energy

Source: IEA (2009) World Energy Outlook

4.1. BOOSTING INNOVATION There is strong consensus in empirical research that a wide palette of instruments is neces-

sary to drive forward the needed investments in deployment and innovation of new tech-

nologies. Carbon pricing is required to provide economic incentives to deploy energy saving

and low carbon technologies as well as investment in innovation; entirely new research show

that firms patenting activity related to energy saving technologies is strongly related to en-

ergy taxes20. However, tax induced innovation is not sufficient: returns on investment in in-

novation with benefits only to be reaped in one to two decades time are very uncertain and

20Popp (2006), Johnstone et al (2009) and Copenhagen Economics for DG TAXUD (2010) show that higher ener-gy prices provide incentives for investments in R&D, which leads to patents, since higher prices change the relative returns to the benefit of energy and GHG displacing technologies. A one per cent increase in energy prices (through for example taxes) implies approximately a 0.4 per cent increase in energy technology patenting. The literature ac-knowledges a time lag between the price signal and new patents: typically half of the induced innovations have oc-curred 3-5 years after the price increase.

2

2,2

2,4

2,6

2,8

3

3,2

3,4

3,6

3,8

4

2007 2020 2030

Gt CO2

Energy efficiency

Renewable energy

Nuclear energy

CCS

Chapter 4 INNOVATION, INVESTMENT CLIMATE AND EXTERNAL

COMPETIVENESS


21

the benefits from such innovation will often come to society at large, not just the private

firms providing the funding. Thus public sector funding for research, development and in-

novation is required as well.

Hence, our shot at the best policy mix towards 2020 has the following ingredients. Carbon

pricing such as ETS can help the most mature renewable technologies to reach target com-

pliance21. On top of that we recommend a TCG scheme that includes also the most mature

technologies mainly hydro power, biomass and land based wind power with the NO-SE as a

possible forerunner of a wider European based system. Healthy competition among such al-

ready near stand alone technologies with the safety net of a TGC – higher ETS price means

lower TGC price et vice versa22 – could help ensure that investment and innovation is fo-

cused on the most promising areas.

Furthermore, we also need public support for less mature technologies. We have already dis-

cussed offshore wind in the previous chapter, so we will focus the discussion here on other

issues and technologies. The essential point is that the support for non-mature technologies

should be about speeding up progress of technology not deployment per se. The success cri-

terion is to bring down future costs of mitigation, not reducing emissions today and tomor-

row: for that we need carbon taxes and some measured support for mature technologies.

Hence for the least mature technologies, e.g. 2nd generation biofuels, we need demonstration

projects, capital cost incentives, credits etc, cf. Figure 4.2. Other more developed non-

mature technologies need feed-in tariffs, either fixed or tendered (high cost-gap technologies,

e.g. PV) or TGCs (low cost-gap technologies, e.g. onshore wind).

21 it may be worthwhile to follow the EU Commission May 2010 proposal to go with the 30 per cent reduction tar-get for 2020: it could bring the ETS allowance price up to € 30, which is still less than the € 41 needed to reach the 20 per cent goal foreseen by DG TREN “Trends to 2030 – Update 2007”, before the economic downturn in 2008. 22 If the ETS prices go up, the price of carbon dioxide goes up. This makes renewable energy relatively cheaper and more renewable energy is produced, and that lowers the TGC price.


22

Figure 4.2 From research to market deployment: instruments tailored to maturity of

technology

Source: Copenhagen Economics based on IEA (2008)

A part of the economic literature suggests that massive early deployment can reduce future

costs through “learning”; we suggest that this is very doubtful at best. Indeed, there seems to

be a tendency to assign too much of the costs reductions over time for technologies to his-

torical deployment and far too little to independent as well as research driven cost savings

over time. Consider, for instance, the case of solar energy. Despite research efforts that began

during the energy crises of the 1970s, solar energy is still only cost competitive in niche mar-

kets, such as remote off-grid locations. This leaves a potential role for government-sponsored

R&D to fill in the gaps23.

4.2. INVESTMENT CLIMATE One of the most prevailing criticisms towards TGC vis-à-vis the main alternative, namely

feed-in tariffs, is that it provides less investor safety and has proven less capable of driving

expansion of renewable energy. A (fixed) feed-in tariff provides a legal guarantee by a mem-

ber state that production from a given installation for a given time period receives a prede-

fined remuneration level. This is presumably better than being part of a TGC market where

prices can go up and down.

The most significant form of operating support for electricity, heating and transport are

feed in tariffs and obligations. Reviewing the relationship between project risk and in-

strument choice, the empirical evidence suggests that the more reliable revenue stream

provided by feed in tariffs is generally more effective in driving renewable energy

growth, particularly for a broad range of technologies. Obligations and tradable green

23 Copenhagen Economics for DG TAXUD ( 2010)

Development � Niche markets � Mass marketsTime

Prototype Prototype Prototype Prototype and demonstration and demonstration and demonstration and demonstration stage stage stage stage technologiestechnologiestechnologiestechnologies

(e.g. 2nd generation biofuels)

ContinuityContinuityContinuityContinuity, R&D, , R&D, , R&D, , R&D, createcreatecreatecreatemarketmarketmarketmarket attractivenessattractivenessattractivenessattractivenessCapital cost incentives;

investment tax; credits; discounts; loan guarantees,

etc.

HighHighHighHigh costcostcostcost----gapgapgapgaptechnologiestechnologiestechnologiestechnologies(e.g. PV)

LowLowLowLow costcostcostcost----gapgapgapgaptechnologiestechnologiestechnologiestechnologies

(e.g. onshore wind)

MatureMatureMatureMaturetechnologiestechnologiestechnologiestechnologies(e.g. hydro)

TechnologyTechnologyTechnologyTechnology----neutralneutralneutralneutralcompetitioncompetitioncompetitioncompetition

TGC; Carbon trade (EU ETS)

StabilityStabilityStabilityStability, , , , lowlowlowlow----riskriskriskrisk incentivesincentivesincentivesincentivesPrice based: FIT, FTP

Quantitybased: tender

Imposed market risk, guaranteed Imposed market risk, guaranteed Imposed market risk, guaranteed Imposed market risk, guaranteed but declining minimum returnbut declining minimum returnbut declining minimum returnbut declining minimum return

Price based: FIP

Quantity based: TGC with technology association

StimulateStimulateStimulateStimulate marketmarketmarketmarket pullpullpullpullVoluntary (green) demand


23

certificates often suffer from revenue volatility and require payment of a risk premium,

which appears to make them both less effective and efficient.

European Commission24

However, this argument should be weighed against a number of counter arguments.

First, investor confidence cannot be narrowed down to certainty about return of one year’s

deployment. What matters for an investor is the probability that deployment and invest-

ments in innovation in low carbon technologies in the coming decades are being rewarded in

the market place. We would argue that an international binding agreement between a group

of member states committing to buying renewable energy in a predictable and pre-

announced trajectory in the context of EU wide binding legal targets may provide as much

investor confidence as relying on the willingness of individual member states to provide sub-

sidies over the annual budget also in the future.

Second, a well functioning TGC scheme can provide very stable investment signals and re-

wards, which the Swedish version is a clear example of. A clearly announced trajectory, ad-

justed on an ongoing basis to take into account new assessments of supply potential can pro-

vide a flexible instrument to achieve targets while keeping economic incentives stable as dis-

cussed in chapter 2.

Third, an overall support structure that unnecessarily inflates costs is at risks of a policy

backlash. As noted several times in this report, actual subsidy levels to renewable energy in

EU member states are presently far below the level needed to deliver at 2020 targets even

with the most efficient approach. As actual costs rise going forward, support for the most ex-

pensive technologies may be at risk of being discontinued or reduced sharply downwards.

Indeed, the current crisis has led to some countries scaling down their subsidies to renewable

energy, first of all in Germany, Spain and Netherlands.

Fourthly, providing investment guarantees to one specific part of the economy, in this case

investors in renewable energy, has costs for other actors in the energy market, often other di-

visions within the same energy companies. To provide an example: if producers of renewable

energy are “protected” from more general market forces for example a longer period of lower

wholesale power market prices, they may reinforce this market weakness by maintaining

higher production and investments, leading to even more depressed earnings for coal and gas

based producers. Such policy induced transfers of risk between producers illustrate that there

is no “free lunch”: higher risk premium for producers of fossil based production may hold

back needed investments in “clean” coal and gas based production. For decades still, EU will

continue to need such production to provide electricity and district heating to its consumers.

24 Commission staff working document, Review of European and national financing of renewable energy in accor-dance with Article 23(7) of Directive 2009/28/EC, accompanying document to Communication from the com-mission to the council and the European parliament, Renewable Energy: Progressing to 2020. (2009)


24

4.3. THE INDUSTRIAL PERSPECTIVE: INTERNATIONAL COMPETIVENESS Over the coming 10 years, as much as much as € 500 billion25 is to be invested in power

generation to deliver on new energy and replacing old installations. More than half of that

will be investment in renewable energy. This provides a massive home market for EU pro-

ducers to deploy their products, such as new generations of wind power, new technologies

that combines co-firing of biomass with coal and gas, which can be exploited outside EU as

well. European producers are increasingly focusing their marketing efforts also outside EU

as the climate and energy agenda is spreading to large emerging economies such China and

India as well as US and Canada.

The lack of a strong internal market dimension to this massive expansion of renewable en-

ergy within EU may hurt EUs own industry and is strongly at odds with the EU 2020

agenda that emphasises “smart, green growth”. Continuing to operate with 27 fragmented

national markets with 27 different support schemes constantly being revised prevents EU

firms from exploiting the potential economies of scale and scope that in other fields of EU

industrial policy is seen as essential for competiveness. Indeed, denying the benefits of an in-

ternal market to precisely the industry that is marked to a drive for new jobs and growth

over the coming 10 years appears to be a major industrial policy failure.

Indeed there is a risk that a renewable energy policy too much focused on larger scale de-

ployment of non-mature technologies in the context of fragmented markets will fail to pro-

vide EU firms with the required edge in EU and export markets. The very substantial sup-

port to deployment of solar cells in Germany (and Spain) over the last 10 years may provide

the best example of this. It has cost Germany electricity consumers more than € 50 billion26

while failing to bring production costs to anything close to levels that justify major deploy-

ment: the shadow price of each ton of CO2 replaced still exceeds € 700 27, far above the level

of ETS allowances prices of usually around € 20-25. At the same time, the market is now

largely dominated by producers from China and Taiwan that in 2009 accounted for over 50

per cent of the world market (and their market share is projected to grow) while German

producers account a market share under steady decline, now below 10 per cent.

25 EU Commission, COM(2011) 31 final, p 7. 26 Copenhagen Economics for Danish Energy Association (2010) 27 RWI, p 13.


25

ANP Pers Support, Duurzame Energie Koepel, Kabinet kiest kernenergie boven versnelling

duurzaam, 01.10.2010.

http://pressrelease.perssupport.nl/pressrelease/detail.do?pressId=46761&searchKey=1

&languageId=NL

Bloomberg (20 Jan 2011), Germany to Speed Solar-Subsidy Cuts to Undercut Boom,

www.bloomberg.com/news/2011-01-20/german-subsidies-for-solar-power-to-be-

reduced-as-much-as-15-from-july-1.html

Bosetti et al (2009), ”The role of R&D and technology diffusion in climate change mitiga-

tion: New perspectives using the Witch model”, OECD, Working Paper

Communication from the commission to the European Parliament, Jan 2011, Renewable

Energy: Progressing towards the 2020 target, COM(2011) 31 final and SEC(2011)

131 final

Commission staff working document. Review of European and national financing of renew-

able energy en accordance with Article 23 (7) of Directive 2009/28/EC, SEC(2011)

131final, accompanying COM(2011) 31

Communication from the commission to the European Parliament (COM(2010) 265 final)

May 2010

Commission staff working document, Review of European and national financing of renew-

able energy in accordance with Article 23(7) of Directive 2009/28/EC, accompany-

ing document to Communication from the commission to the council and the Eu-

ropean parliament, Renewable Energy: Progressing to 2020. (2009)

Copenhagen Economics for Danish Energy Association (2010a), Grøn Energi Innovation i

Danmark

Copenhagen Economics for European Commission DG Taxation and Customs Union

(2010b) Innovation of energy technologies: the role of taxes

ECN (2010) What is the scope for the Dutch government to use the flexible mechanisms of

the Renewables Directive cost-effectively?

Ecofys et al. (2011), Financing Renewable Energy in the European Energy Market

REFERENCES


26

Eurelectric (2008), Enhancing Economic Efficiency – Benefits of trade

European Biomass Industry Association, Biomass resources and production potential

www.eubia.org/215.0.html

Europen Council, Conclusions on Energy, 4 Februray 2011

www.consilium.europa.eu/uedocs/cms_data/docs/pressdata/en/trans/119253.pdf

European Council, Transport, Telecommunications and Energy, Press release, 28 Feb 2011,

6950/11,

www.consilium.europa.eu/uedocs/cms_data/docs/pressdata/en/trans/119531.pdf

EWEA (2009) Pure Power – Wind energy targets for 2020 and 2030, 2009 update

EWI (2010) European RES-E Policy Analysis – A Model-based analysis of RES-E deploy-

ment and its impact on the conventional power market

Guardian Environment Network (3 Aug 2010), Spain proposes deep cuts to solar PV sup-

port, www.guardian.co.uk/environment/2010/aug/03/spain-cuts-solar-pv

HM Government, Secretary of State for Energy and Climate Change (2009), The UK Re-

newable Energy Strategy

Johnstone et. al., 2009, “Renewable energy Policies and technological innovation: Evidence

based on patent counts”, Springer, Environ Resource Econ

OECD (2009), The Economics of Climate Change Mitigation: How to Build the Necessary

Global Action in a Cost-Effective Manner, Economics Department Working Paper

Popp, 2006, “R6D subsidies and climate policy: Is there a “free lunch”?”, Climatic Change,

77, 311-341

RWI (2009), Economic Impacts from the Promotion of Renewable Energy Technologies –

The German Exerience, Ruhr Economic papers #156

Swedish Energy Agency (2010), Gemensamt elcertifikatsystem med Norge, ER 2010:28

Swedish Energy Agency, Elcertifikat, Kvotpliktig, accessed on 7 March 2011,

www.energimyndigheten.se/sv/Foretag/Elcertifikat/Kvotpliktig/


27

Svenska Kraftnät, Cesar – Svenska Kraftnäts system för kontoföring av elcertifikat m.m.,

accessed on 7 March 2011, https://elcertifikat.svk.se/cmcall.asp