Delft University of Technology

National Renewable Policies in an International Electricity Market

A Socio-Technical Study
Iychettira, Kaveri

DOI
10.4233/uuid:d247e271-3321-4661-b29a-08070ae9090f
Publication date
2018
Document Version
Final published version

Citation (APA)
Iychettira, K. (2018). National Renewable Policies in an International Electricity Market: A Socio-Technical
Study. Netherlands. https://doi.org/10.4233/uuid:d247e271-3321-4661-b29a-08070ae9090f

Important note
To cite this publication, please use the final published version (if applicable).
Please check the document version above.

Copyright
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent
of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Takedown policy
Please contact us and provide details if you believe this document breaches copyrights.
We will remove access to the work immediately and investigate your claim.

This work is downloaded from Delft University of Technology.

For technical reasons the number of authors shown on this cover page is limited to a maximum of 10.
 Kaveri K. Iychettira
Invitation
DoDtoral
Thesis
Delft,
The
Netherlands
2018 Fri April 6
10 am

National Renewable
National Renewable Policies in an Policies in an
International
International Electricity Market:

National Renewable Policies in an International Electricity Market: A Socio-Technical Study

Electricity Market:
A Socio-Technical Study
A Socio-Technical Study
Kaveri Iychettira
Kaveri K. Iychettira

TU Delft
Aula
Conference Center
Mekelweg 5
2628 CC Delft
The Netherlands

“For KTH Royal Institute of Technology:

DOCTORAL THESIS IN ELECTRICAL ENGINEERING
TRITA-EECS-AVL-2018:12
www.kth.se
ISBN: 978-94-6233-904-0
N ATIONAL R ENEWABLE P OLICIES IN AN I NTERNATIONAL
E LECTRICITY M ARKET
A SOCIO - TECHNICAL STUDY

Kaveri Kariappa I YCHETTIRA

N ATIONAL R ENEWABLE P OLICIES IN AN I NTERNATIONAL
E LECTRICITY M ARKET
A SOCIO - TECHNICAL STUDY

Dissertation

for the purpose of obtaining the degree of doctor

at Delft University of Technology,
by the authority of the Rector Magnificus prof. dr. ir. T.H.J.J. van der Hagen,
Chair of the Board for Doctorates,
to be defended publicly on,
Friday 6 April 2018 at 10:00 o’clock

by

Kaveri Kariappa I YCHETTIRA

Master of Science in
Engineering and Policy Analysis
Delft University of Technology, the Netherlands
born in Bangalore, India.
This dissertation has been approved by the promotors.

Composition of the doctoral committee:

Rector Magnificus chairman
Prof. dr. ir. M.P.C. Weijnen Delft University of Technology
Dr. ir. R.A. Hakvoort Delft University of Technology

Independent members:
Prof. dr. ir. R.W. Künneke Delft University of Technology
Prof. dr. ir. M.L. Rivier Abbad Comillas Pontifical University
Prof. dr. K. Blok Delft University of Technology
Dr. ir. M.R. Hesamzadeh KTH Royal Institute of Technology
Prof. dr. J. Watson University of Sussex

ISBN 978-94-6233-904-0

The doctoral research has been carried out in the context of an agreement on joint doc-
toral supervision between Comillas Pontifical University, Madrid, Spain, KTH Royal In-
stitute of Technology, Stockholm, Sweden and Delft University of Technology, the Nether-
lands.

Keywords: RES-E, policy design, support schemes, renewable electricity, agent-

based modelling, investment, electricity, cross-border effects, IAD
framework

Copyright © 2018 by K.K. Iychettira. All rights reserved. No part of the material protected
by this copyright notice may be reproduced or utilized in any form or by any means, elec-
tronic or mechanical, including photocopying, recording or by any information storage
and retrieval system, without written permission from the author.
Printed in the Netherlands
Thesis Supervisors:
promotor: Prof. dr. ir. M.P.C. Weijnen
copromotor: Dr. ir. R.A. Hakvoort

Members of the Examination Committee:

Prof. dr. ir. R.W. Künneke Delft University of Technology
Prof. dr. ir. M.L. Rivier Abbad Comillas Pontifical University
Prof. dr. K. Blok Delft University of Technology
Dr. ir. M.R. Hesamzadeh KTH Royal Institute of Technology
Prof. dr. J. Watson University of Sussex

TRITA-EECS-AVL-2018:12
ISBN 978-94-6233-904-0

Copyright © K.K. Iychettira, 2018

Printed by: Publisher
Keywords: RES-E, policy design, support schemes, renewable electricity, agent-
based modelling, investment, electricity, cross-border effects, IAD
framework
Front cover: The cover picture is a depiction of independent entities trying to cre-
ate a common electricity grid together. Kaveri has superimposed an
image of the electricity grid by James W. Smith obtained on Flickr, on a
painting by M.C. Escher.

An electronic version of this dissertation is available at

http://repository.tudelft.nl/.
 SETS Joint Doctorate
The Erasmus Mundus Joint Doctorate in Sustainable Energy Technologies and Strate-
gies, SETS Joint Doctorate, is an international programme run by six institutions in co-
operation:

• Comillas Pontifical University, Madrid, Spain

• Delft University of Technology, Delft, the Netherlands

• Florence School of Regulation, Florence, Italy

• Johns Hopkins University, Baltimore, USA

• KTH Royal Institute of Technology, Stockholm, Sweden

• University Paris-Sud 11, Paris, France

The Doctoral Degrees issued upon completion of the programme are issued by Comillas
Pontifical University, Delft University of Technology, and KTH Royal Institute of Technol-
ogy.
The Degree Certificates are giving reference to the joint programme. The doctoral candi-
dates are jointly supervised, and must pass a joint examination procedure set up by the
three institutions issuing the degrees.
This thesis is a part of the examination for the doctoral degree. The invested degrees are
official in Spain, the Netherlands and Sweden respectively.
SETS Joint Doctorate was awarded the Erasmus Mundus excellence label by the Euro-
pean Commission in year 2010, and the European Commission’s Education, Audiovisual
and Culture Executive Agency, EACEA, has supported the funding of this programme.
The EACEA is not to be held responsible for contents of the Thesis.
 A CKNOWLEDGEMENTS
"It always seems impossible until it’s done," said Nelson Mandela. The same feeling per-
vades even as I write this. During the last four years however, there were a lot of people
in my life who thought it was possible. Some of whom, through their own challenges
showed me it was possible gracefully, and with perseverance. Others showed me the
way, and others still helped by making sure I knew that there was much more to life than
a PhD. I take this opportunity to acknowledge their contribution and care; they are the
village that it took to create this thesis.
I thank my pragmatic, witty, and insightful co-promotor, dr. ir. Rudi Hakvoort, for
giving me the opportunity to undertake this journey. Rudi was a mentor, advisor, and
friend, therefore so much more than the title ’supervisor’ implies. Through it all, he al-
lowed me the freedom to explore my mad ideas, helped keep my perfectionist tenden-
cies in check after first letting them run wild; his light-hearted approach, creative ideas,
and supportiveness ensured that I stayed positive, and kept trying when results didn’t
come through. I thank Prof. dr. ir., Margot Weijnen, my promotor, for her guidance,
especially during the later half of my PhD. I thank her for ensuring that I completed the
dissertation on time, by keeping my ambitiousness in check at the final stages. During
the many meetings with her, she was insightful and friendly, and I never left a meeting
without feeling encouraged. I especially thank both Rudi and Margot for their warmth
and understanding when my wrists gave in due to repetitive stress injury, as well as for
their interest and effort in helping me get to the next step in my career. I also thank
Prof. Pedro Linares at Comillas for the academic year I spent in Madrid, Spain. Weekly
meetings with him were instrumental in shaping my ideas during the early stages of my
dissertation.
I thank the several colleagues I collaborated and co-authored papers with, particu-
larly the EMLab community: Dr. ir. Laurens de Vries, dr. ir. Emile Chappin, Joern, Rob
de Jeu, Salman, and Prad. Laurens was also a mentor to me. I am grateful for the op-
portunities to collaborate with him on several projects, and for all the master students
he sent my way. I thank him for the memorable dinners he hosted, for the fun and lively
conversations, and for his advice on the process of PhD-ing. I thank Joern for teaching
me the ropes of EMLab, and for inspiring me to be a good supervisor and friend to my
master students.
I thank Wiebke, whose friendship was the most valuable outcome of the PhD Start-
Up course! I am incredibly grateful for her friendship, inspiration and counsel. I thank
Yesh for being the big brother he has been: when I had lost hope of ever getting sensible
results from EMLab, he managed to convince me that there was light at the end of the
tunnel. I thank Siddhartha Mukherjee for his friendship, the fantastic biryanis with Udai,
the fun conversations, and for introducing me to Vikram Seth!
I thank the wonderfully ’gezellige’ group at the E&I section. The atmosphere was
one of camaraderie, and a feeling that we all had each others’ backs. The Power Rangers

ix
x

meetings were especially useful, usually indirectly, for my own research. I’m also grate-
ful for conversations with Amineh and Igor, which have influenced ideas in this disser-
tation. I thank Elta for convincing me to take up the PhD programme, and for being the
supportive figure she was through the PhD! I thank Binod, Jorge, and Jose who made
my first year a very happy one. I thank Joao, Ben, Esther, Amineh, Arthur, Amit, Shan-
tanu, Ozge, Kasper, Tristan, Frank, Reinier, Cherrelle, Juliana, Chris, Dierdre, and my
officemates Victoria, Anna, and Graciela. Arthur, Joao, and Shantanu kept spirits in the
department high, and made sure we didn’t stay in the office too long after work hours! I
also thank Rishab and Ad for all the fun times in my final year. I’m glad I got to know you
before I left!
When I went to Madrid for the SETS mobility in Aug 2014, I was immediately made to
feel like a part of the Ricci-Spanish family. I thank Adela, Alessandro, Pablo, Checa, Luis,
Paco, Elena, Eva, and Zarrar. I also thank the SETS Spanish cohort in that year, Katia,
Peyman, Quentin. I thank Sara for her diligence and dedication in helping us get through
the various administrative SETS requirements during the 4 years. In KTH, Stockholm too,
I was very lucky to have met wonderful people like Francisco, Gudrun, Tin, Viktor, Ravi,
Fabian, and Aravind. Their friendship made even Swedish winters seem warm and fun.
I also met a very nice bunch of people from all over the world, in the SETS pro-
gramme, who made the journey easier together: Christian, Marcin, Nenad, German,
Paolo, Anna.
Towards the final stages, a lot of people helped proof-read my dissertation. Some of
them were acquaintances, and others family! I can’t thank them enough for offering to
help under tight time constraints, and doing diligent jobs too. My mum, Raunak Kalra,
Layla Abdulkarim, Vivek anna, Ranjani, Yeshambel, and Vikram Srinivas. I thank Wiebke
and Ozge for standing by my side as paranymphs. I also thank Prof. James W. Smith from
the university of Mc. Master, for letting me use his beautiful photograph for my thesis
cover.
I thank my friends from Bangalore - Ranjani Iyer, Chitwan Mahipal, Kaushik Ran-
garajan, Vidya Shree, Nisha Suresh, Roshan Rao, Santosh Harish, and Anil Kumar who
have all been by my side despite the distance. I thank Daddy for being an inspiration
in his pragmatic approach to life. He made sure that I didn’t let the PhD consume me,
and reminded me that there were more important things in life! I thank Mum for her
constant care, love, and support. She never let me feel like I lived too far away from fam-
ily. And finally I thank Vikram, my fiance, whose presence makes every moment more
joyful. He is a source of calm strength, and unconditional affection. I look forward to all
that life has in store for us!

Kaveri Iychettira
February, 2018
 C ONTENTS

Acknowledgements ix
Summary xvii
Samenvatting xxi
Sammanfattning xxv
1 Introduction 1
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.1 A brief history . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.2 When Renewables Arrived . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Problem Definition and Research Overview. . . . . . . . . . . . . . . . . 4
1.2.1 Research questions . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2.2 Scientific contribution . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2.3 Research framework and thesis structure . . . . . . . . . . . . . . 6
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2 Theoretical Foundations: Identifying the design space 13
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2 Theoretical Foundations and Methodology . . . . . . . . . . . . . . . . . 15
2.2.1 Theoretical Foundations . . . . . . . . . . . . . . . . . . . . . . . 15
2.2.2 A Policy Design Framework . . . . . . . . . . . . . . . . . . . . . 18
2.3 Applying the policy design framework to RES-E support design. . . . . . . 21
2.3.1 Identification of participants, action situations, and exogenous vari-
ables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.3.2 Design Space: Design Elements of RES-E Support Schemes . . . . . 23
2.4 Empirical representativeness of design elements . . . . . . . . . . . . . . 26
2.5 Merits and limitations of the approach . . . . . . . . . . . . . . . . . . . 28
2.6 Conclusions and policy implications . . . . . . . . . . . . . . . . . . . . 29
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3 Modelling: EMLab Generation and Renewable Support 35
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.2 System decomposition of Base Model in EMLab . . . . . . . . . . . . . . 36
3.2.1 Model conceptualisation. . . . . . . . . . . . . . . . . . . . . . . 37
3.2.2 Model implementation . . . . . . . . . . . . . . . . . . . . . . . 37
3.3 Electricity spot market and Investment . . . . . . . . . . . . . . . . . . . 38
3.3.1 Market clearing . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.3.2 Dynamic determination of load duration curve . . . . . . . . . . . 39
3.3.3 Investment in generation capacity . . . . . . . . . . . . . . . . . . 40

xi
xii C ONTENTS

3.4 Modelling renewable electricity support . . . . . . . . . . . . . . . . . . 42

3.4.1 Modelling of Design Elements . . . . . . . . . . . . . . . . . . . . 43
3.5 Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.5.1 Load duration curves . . . . . . . . . . . . . . . . . . . . . . . . 46
3.5.2 Demand growth trends . . . . . . . . . . . . . . . . . . . . . . . 46
3.5.3 Technology characteristics . . . . . . . . . . . . . . . . . . . . . . 48
3.6 Model Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.6.1 Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.6.2 Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

4 Renewable support interacts with the electricity market 55

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.1.1 Motivation and Research Objective . . . . . . . . . . . . . . . . . 56
4.1.2 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.1.3 From Scenario Analysis of Policies to Design Elements . . . . . . . 58
4.1.4 Choice of Modelling Approach . . . . . . . . . . . . . . . . . . . . 58
4.2 Experiment Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.2.1 Input Data: Case of the Netherlands . . . . . . . . . . . . . . . . . 61
4.2.2 Critical Review of Modelling Assumptions . . . . . . . . . . . . . . 61
4.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.3.1 Results and Discussion. . . . . . . . . . . . . . . . . . . . . . . . 62
4.3.2 Discussion and Interpretation . . . . . . . . . . . . . . . . . . . . 67
4.3.3 Applicability of the Design Element Approach . . . . . . . . . . . . 68
4.4 Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

5 Congruency between national support schemes and an international elec-

tricity market 75
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.1.1 Literature review . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
5.2 Experiment design and data . . . . . . . . . . . . . . . . . . . . . . . . 78
5.2.1 Experiment design . . . . . . . . . . . . . . . . . . . . . . . . . . 78
5.2.2 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
5.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
5.3.1 Impacts of interconnection on electricity prices . . . . . . . . . . . 80
5.3.2 Effects on costs to consumer . . . . . . . . . . . . . . . . . . . . . 81
5.3.3 Costs and Revenues to Producer . . . . . . . . . . . . . . . . . . . 85
5.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
5.4.1 Merits and limitations of the model . . . . . . . . . . . . . . . . . 87
5.4.2 Validation: Comparison to observed data . . . . . . . . . . . . . . 88
5.4.3 Distortion of the internal energy market . . . . . . . . . . . . . . . 89
5.5 Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
C ONTENTS xiii

6 RES-E policy design in perspective of the energy transition 95

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
6.2 Subsidy costs and electricity prices . . . . . . . . . . . . . . . . . . . . . 97
6.2.1 Drivers of RES-E subsidy . . . . . . . . . . . . . . . . . . . . . . . 97
6.2.2 Model results on the relationship between subsidy costs and elec-
tricity prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
6.2.3 Insights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
6.3 EU-ETS in relation to Subsidy Costs. . . . . . . . . . . . . . . . . . . . . 103
6.3.1 Efficiency and cost-effectiveness of the EU-ETS and RES-E? . . . . . 104
6.3.2 Interacting price effects . . . . . . . . . . . . . . . . . . . . . . . 105
6.3.3 Implications for model results . . . . . . . . . . . . . . . . . . . . 106
6.4 Impact of flexibility on the economic value of VRE . . . . . . . . . . . . . 107
6.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
6.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
7 Conclusions and reflections 117
7.1 Answers to Research Questions . . . . . . . . . . . . . . . . . . . . . . . 117
7.2 Reflections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
7.2.1 Role of simulation in science. . . . . . . . . . . . . . . . . . . . . 120
7.2.2 How does the model relate to the real world?. . . . . . . . . . . . . 121
7.2.3 Reflection on approach, model . . . . . . . . . . . . . . . . . . . 122
7.3 Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
A Data and Assumptions 127
A.1 Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
A.1.1 Demand and Fuel Price Trends . . . . . . . . . . . . . . . . . . . 127
A.1.2 Renewable Technology Cost Curve Assumptions . . . . . . . . . . 128
A.1.3 Target and Potential Curves . . . . . . . . . . . . . . . . . . . . . 129
A.1.4 Assumptions: Technology Characteristics . . . . . . . . . . . . . . 131
A.2 Current technology cost data trends . . . . . . . . . . . . . . . . . . . . 132
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
B Results 133
B.1 Chapter 4: Results and Sensitivity Analysis . . . . . . . . . . . . . . . . . 133
B.1.1 Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
B.1.2 Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
B.2 Chapter 5: Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
List of Publications 139
Curriculum Vitæ 141
 N OMENCLATURE
N PV Net Present Value

W ACC Weighted Average Cost of Capital

W ACC r ev WACC adjusted for risk aversion

∗ Asterisk in the exponent denotes equilibrium values.

g Power plant index

r ep Repetition index per scenario

s Segment index

t Time step in years

a g ,s Available capacity of power plant g, in segment s [in MW]

C FOp,g Expected cash flow for power plant g, during operation [in Eur]

f c g ,t +n Fixed costs of power plant g, in time t+n [in Eur/MWh]

Ig Investment cost of power plant g in t [in Eur]

Kg Nominal capacity of power plant g [in MW]

n Number of years ahead of current tick, for which value is being computed

n r ep Number of repetitions per scenario [40]

n t i ck Number of ticks per repetition [40]

p s,t +n Electricity spot market price for segment s, estimated at time t, for a period n
years ahead

pa yment g ,t Payment of subsidy to RE producer for plant g at time t [in Eur]

rD Rate of debt

r Eb Basic rate of equity

rE p Price risk component of rate of equity

rE Rate of equity

r g ,t +n Running hours of power plant g, in segment s, at time t+n [in hours]

xv
xvi C ONTENTS

r Gen t Total renewable energy generation at time t [in MWh]

tb Power plant construction time

tD Power plant depreciation time

t ar g et t Total target for renewable energy generation at time t [in MWh]

vc g ,t +n Variable costs of power plant g, in time t+n [in Eur/MWh]

X ant e g Total subsidy per MWh of generation for plant g, discounted to present value
[in Eur/MWh]

X post g Total cost per MWh of plant g, discounted to present value [in Eur/MWh]
 E XECUTIVE S UMMARY
The current regulatory framework under which the support schemes for Renewable en-
ergy sources specifically for electricity (RES-E) operate, is provided for by the Directive
2009/28/EC. It sets a 20% target for energy consumption, while relying on legally bind-
ing, national targets until 2020. The goal to promote RES-E, in the European context,
coexists with goals of ensuring a single internal market for electricity, and security of
supply in the European Union, and these simultaneous goals are not always congruent
with each other.
Today, significant amounts of intermittent RES-E in the energy mix have led to un-
intended effects. An important consequence is the so called ‘merit-order effect’, where
the spot market electricity price reduces to the extent by which the renewable electric-
ity generation displaces demand along the merit order. There is concern that part of
the merit order effect spreads across national borders. Vitally, implications of the merit
order effect on the effectiveness of RES-E support schemes are unclear. Another impor-
tant effect of the price reduction is that, the lower the average electricity market price,
the greater the costs of subsidies, making the phasing out of subsidies for renewable,
intermittent sources more difficult.
With respect to electricity from renewable sources, this achievement of the three ob-
jectives took the shape of "making renewable support schemes more market-based",
"ensuring renewables are driven by market signals". However, it is not often clear what is
meant by such statements in policy documents by the EC. What features of the support
scheme are being referred to? What would it mean for renewables solely to be driven
by market signals? How would features of support schemes impact for instance, the
merit order effect, and vice-versa? These issues are encapsulated in the first problem
addressed in the thesis: to unravel the interactions between renewable support scheme
design and a single isolated electricity spot market, with a long term perspective.
Since countries are now increasingly interconnected, the second major issue tackled
in this thesis concerns cross border effects due to different renewable support schemes be-
tween neighbouring countries in a common electricity market. This issue addresses con-
cerns about the merit-order effect spreading across national borders, and the ensuing
distributional implications.
The final issue addressed in this dissertation relates to the long term economic viabil-
ity of electricity from renewable sources given the current institutional and physical setting
they operate in. Costs of renewable technologies have dropped dramatically and yet ef-
fects such as their reducing market value lead to questions about whether it is possible
for them to attain economic viability in a decarbonised power sector. Accordingly, the
main research question in this dissertation is:

How do national renewable electricity support schemes interact with the electricity

xvii
xviii S UMMARY

market over the long term (20-30 years) as the European Union transitions to a decar-
bonized energy system?

Method and Key Findings

Identifying Design Elements of RES-E Schemes:
The first step towards understanding the design of RES-E support schemes in this
work has been to identify the design elements that they comprise of. Design elements
are defined as a closed set of components that are common to all renewable electricity
support schemes; they form the smallest level of analysis. The underlying idea is not that
there exists a choice between policy A or B, but how either policy instrument should be
designed. Such a perspective allows the policy maker such as the European Commission
to decide which design features are essential in an RES-E scheme, rather than propose an
entire scheme itself. To this end, in Chapter 2 a formal approach to RES-E policy design
based on Design Theory, the Institutional Analysis and Development (IAD) Framework,
and Agent Based Modelling and Simulation is presented. Using this approach, and as-
sisted by a literature review, ten design elements are identified. They are corroborated
with empirical data by studying six RES-E support schemes in Europe.
RES-E support schemes were implemented in terms of their design feautures, on an
existing, agent-based model of the electricity market, called the Energy Market Laboratory,
or "EMLab". EMLab is an initiative of the Energy and Industry section of the faculty of
Technology, Policy and Management in TU Delft, built to study various aspects of the
energy transition in Europe. At its core, EMLab comprises an electricity market clearing
algorithm and an endogenous investment algorithm, where agents who take investment
decisions are boundedly rational in their knowledge about the future, much like reality.
Modelling Design Elements of RES-E Schemes:
In order to understand precisely how each design element impacts the overall goals
of sustainability, affordability, and security of supply, three design elements are imple-
mented in a model of an electricity market, with RES-E policies. The three design ele-
ments are quantity vs. price based policy, technology neutrality vs specificity, and price
setting ex-post vs. ex-ante. The objective of Chapter 3 is to assess the impact of design el-
ements of RES-E support schemes on a single (isolated, uncongested) region, modelled
approximately similar to the power sector in the Netherlands, using a long-term agent-
based model of the electricity market. An important uncertainty in the real world is that
of long-term electricity price development. The model demonstrates that accounting for
future electricity prices ex-ante in the subsidy calculation may reduce the overall cost of
subsidy by about 15%, since actors are likely to overestimate future electricity price. The
combination of design elements that provides the highest increase in social welfare is
the quantity based policy, with electricity market price accounted for ex-ante, and with
technology specificity.
Cross border impacts of design elements:
Using EMLab Generation, regions based on Germany and the Netherlands were sim-
ulated, to test the effects of different support scheme designs in each country, and the
common electricity market, on distributional implications under different scenarios of
interconnector capacities. As economic theory would predict, it is found that subsidy
costs increase in the smaller country due to spreading of the merit order effect from the
S UMMARY xix

larger neighbouring country. However, the increase in subsidy costs remains lower than
the reduction in costs of electricity, corroborating earlier, single country analyses on the
topic. Therefore, total costs in both countries reduce as interconnection increases, and
the invisible hand works its magic. As the share of RES-E increases, interconnection has
limited impact on reducing spillage of RES-E while storage becomes increasingly im-
portant. The results show that setting targets for RES-E, should not only take into to
account the targets of the interconnected neighbouring countries, but also capacity of
storage that exists in the system. They raise interesting questions about how subsidy
costs should be allocated between countries over the long term when electricity markets
they operate in are so intricately connected.
RES-E design in perspective of the energy transition:
Results from previous chapters are, in this penultimate chapter, regarded from the
standpoint of the EU energy transition. A literature review examines three major aspects
of the energy transition in Europe: the cost and revenue drivers of RES-E technologies,
the EU ETS, and finally, the role of flexibility. We find that at relatively high average elec-
tricity prices, their impact on subsidy costs is direct and clear; as prices increase, subsidy
costs decrease. However, as electricity prices on the spot market reduce on average, the
change in subsidy cost trends depend far more on cost curves of each RES-E technology.
With respect to the EU-ETS, it is clear that, even if EUA prices play a significant role on
the electricity prices today, their influence on electricity prices and RES-E subsidy costs
will reduce as the share of RES-E generation increases with time. From results in Chap-
ter 5, it is evident that at high shares of RES-E, interconnection will play a limited role in
providing flexibility while storage (long-term, seasonal) will play a pivotal role. Conse-
quently, costs of subsidy for RES-E at high shares substantially depend on the presence
of storage in the system.

Conclusion The central question addressed in this thesis is How do national renewable
electricity support schemes interact with the electricity market over the long term (20-30
years) as the European Union transitions to a decarbonized energy system?
Using an approach based on theoretical foundations of institutional analysis, design
theory, RES-E support schemes are broken down into their design elements from a wel-
fare economics perspective. A modelling framework is then introduced, using the mod-
eling paradigm of agent-based modelling, by which RES-E schemes are modelled in EM-
Lab Generation. EMLab Generation mainly comprises an electricity market clearing and
an endogenous investment module, enabling the researcher to study long term dynam-
ics of policy designs in the electricity sector. The model shows that the way electricity
prices are taken into account while designing subsidies makes a substantial difference
to welfare distributions. When multiple countries are considered, spreading of the merit
order effect from a larger country to a smaller one would cause the smaller country’s
subsidy costs to increase. This indicates that even if RES-E support can be designed to
be "market-based" at an operational level, designing them to be "market based" at an
investment level is far more complicated. It indicates that RES-E targets of countries
should take into account the targets of neighbouring countries, as well as the amount of
flexibility in terms of interconnection and storage in the system.
In terms of its scientific contribution, the thesis takes a step towards integrating in-
xx S UMMARY

stitutional analysis and design theory in order to complement the neo-classical school
of thought which arguably dominates energy policy design and analysis in Europe. The
framework introduced in Chapter 2 brings together two separate strands of literature:
the institutional analysis framework and design theory. The combination is shown to
also be coherent with and complementary to the modelling paradigm of agent based
modelling and simulation. While the framework does not provide causal relationships
which provide insight on why a certain actor behaves in a certain manner under cer-
tain conditions, it does make for a strong fundamental basis to simulate a test-bench on
which the real world can be represented, and theories or "speculations" tested. Although
this framework was applied and demonstrated using renewable support schemes in Eu-
rope, its fundamental nature makes it employable to other domains.
A major advantage of the model is that it is open source. Most models that are used
to inform the European Commission’s policy documents are black box simulations. The
assumptions, data, systems used to set policy are not publicly available. Making mod-
els open source allows for replicability, transparency, and debating their validity, which
are all basic tenets of the scientific method. The results from this dissertation also pave
the way towards new questions that need to be resolved as the shares of RES-E increase
in the elecricity system. In terms of methodology for long-term energy studies, it is
paramount that models employ high temporal resolution at high shares of intermittent
RES-E generation, while also being able to model endogenous investment. For instance,
they raise pressing questions about the need to further develop a methodology for shar-
ing the subsidy costs between countries. These questions need to be explored in greater
detail, and possible mechanisms to cope with the complexities should be evaluated.
 S AMENVATTING
De Europese Richtlijn 2009/28/EC biedt het reguleringskader voor de groei van het aan-
deel duurzaam opgewekte energie, specifiek voor elektriciteit. De doelstelling is dat 20
% van het energiegebruik in 2020 uit duurzame bronnen wordt opgewekt, een en ander
te realiseren door middel van bindende nationale doelstellingen. Het bevorderen van
duurzame energie in de Europese context staat naast de andere doelstellingen van het
realiseren van een interne markt voor energie en het waarborgen van de leveringszeker-
heid in de Europese Unie. Deze doelstellingen zijn niet altijd met elkaar in lijn.
De grote hoeveelheden energie uit intermitterende bronnen in de energiemix heb-
ben geleid tot onbedoelde neveneffecten. Een van de belangrijkste is het zogenaamde
‘merit order effect’: de elektriciteitsprijs op de spotmarkt daalt naarmate energie uit
duurzame bronnen conventionele elektriciteitsproductie verdringt. Dit merit order ef-
fect kan zich over landsgrenzen heen verspreiden. De gevolgen hiervan op de effectiviteit
van beleid om duurzame energie te bevorderen zijn onduidelijk (en worden in dit proef-
schrift onderzocht). Een gevolg van de prijsdaling is verder dat hoe lager de gemiddelde
elektriciteitsmarktprijs, hoe hoger de kosten van de toegezegde subsidiëring worden. Dit
maakt het uitfaseren van subsidies voor duurzame energiebronnen moeilijker.
In de praktijk is geprobeerd om de drie hierboven genoemde doelstellingen met el-
kaar in lijn te brengen door “het meer markt-gebaseerd maken van de ondersteuning van
duurzame energie” en “het verzekeren dat duurzame energiebronnen worden aange-
dreven door marktsignalen”. Het is meestal niet duidelijk wat met dergelijke uitspraken
bedoeld wordt. Aan welke eigenschappen van het ondersteuningsmechanisme wordt er
gerefereerd? Wat betekent het dat duurzame energiebronnen alleen door marktsignalen
gestuurd worden? Welke eigenschappen van ondersteuningsregelingen hebben impact
op bijvoorbeeld het merit order effect, en omgekeerd? Deze kwesties worden in het eer-
ste vraagstuk geadresseerd dat in dit proefschrift aan bod komt: het ontrafelen van de
interacties tussen ondersteuningsregelingen voor duurzame energie in een enkele (geï-
soleerde) elektriciteitsmarkt, bekeken vanuit een langetermijnperspectief.
Aangezien de elektriciteitssystemen in de Europese landen onderling verbonden zijn,
gaat de tweede kwestie die in dit proefschrift wordt behandeld, over de grensoverschrij-
dende beïnvloeding van verschillende ondersteuningsregelingen voor duurzame ener-
gie tussen buurlanden in een gemeenschappelijke elektriciteitsmarkt. Deze kwestie adres-
seert het verspreiden van het merit order effect over nationale grenzen heen, en de daar-
uit volgende verdelingseffecten. Het laatste vraagstuk dat in dit proefschrift wordt be-
handeld, heeft te maken met de economische levensvatbaarheid op lange termijn van
elektriciteitsproductie vanuit duurzame bronnen gegeven de huidige institutionele en
fysieke omgeving. Hoewel de kosten van duurzame elektriciteitsproductie zijn afgeno-
men, leidt hun snel dalende marktwaarde tot vragen in hoeverre deze technologieën
economisch levensvatbaar blijven in een koolstofarme stroomsector. Bijgevolg is de
hoofdonderzoeksvraag in dit proefschrift:

xxi
xxii S AMENVATTING

Hoe interacteren nationale ondersteuningsregelingen met de elektriciteitsmarkt op de

lange termijn (20-30 jaar) nu de Europese Unie overgaat naar een koolstofarm energiesys-
teem?

Methode en belangrijkste bevindingen

Identificeren van de ontwerpelementen van ondersteuningsregelingen voor duurzame
energie: De eerste stap om ondersteuningsregelingen voor duurzame energie te kunnen
ontwerpen is het identificeren van de onderliggende ontwerpelementen. Ontwerpele-
menten worden gedefinieerd als een gesloten set van componenten die gemeenschap-
pelijk zijn aan alle ondersteuningsregelingen voor duurzame elektriciteit. Zij vormen
daarmee het laagste analyseniveau. Het idee hierbij is niet dat er de keus moet wor-
den gemaakt tussen beleid A en beleid B, maar dat allereerst de componenten die in
elke beleidsoptie aanwezig zouden moeten zijn, inzichtelijk worden. Een dergelijk per-
spectief maakt het mogelijk dat een beleidsmaker (zoals de Europese Commissie) eerst
beslist welke componenten essentieel zijn om nader uit te werken in plaats van direct
een hele regeling te ontwikkelen. Hiertoe presenteert hoofdstuk 2 een formele aanpak
voor dit soort ondersteuningsregelingen gebaseerd op design theory, het Institutional
Analysis and Development (IAD) Framework, agent-based modellering and simulatie.
Gebruik makende van deze aanpak, en ondersteund door literatuuronderzoek, zijn tien
ontwerpelementen geïdentificeerd. Deze zijn bevestigd met empirische data door het
bestuderen van zes ondersteuningsregelingen voor duurzame energie in Europa. Ver-
schillende soorten ondersteuningsregelingen (ontwikkeld op basis van hun ontwerpe-
lementen) zijn vervolgens geïmplementeerd in EMLab, een agent-based model van de
elektriciteitsmarkt. EMLab is ontwikkeld aan de faculteit Technologie, Beleid en Mana-
gement van de TU Delft en gebouwd om verschillende aspecten van de energietransitie
in Europa te bestuderen (EMLab is een afkorting voor ‘Energie Markt Laboratorium’).
In de kern bevat het een market clearing algoritme voor de elektriciteitsmarkt in combi-
natie met een endogeen investeringsalgoritme, waarin agenten investeringsbeslissingen
maken die rationeel gebonden zijn aan hun kennis over de toekomst.
Modelleren van ontwerpelementen van ondersteuningsregelingen voor duurzame ener-
gie: Om de invloed van elk ontwerpelement te begrijpen op de algemene doelen van
duurzaamheid, betaalbaarheid en voorzieningszekerheid zijn drie ontwerpelementen
geïmplementeerd in EMLab: de keus tussen een gegarandeerde omvang versus een ge-
garandeerde prijs voor duurzaam opgewekte elektriciteit, technologische neutraliteit ver-
sus technologische specificiteit, en een ex-post versus een ex-ante prijsbepaling (dus
ook subsidievaststelling). In hoofdstuk 3 worden de gevolgen van de ontwerpelemen-
ten op de ondersteuningsregelingen onderzocht, waarbij de analyse zich richt op een
enkele (geïsoleerde, niet-verzadigde) regio met kenmerken die soortgelijk zijn aan aan
de elektriciteitssector in Nederland. Een belangrijk onzekerheid in de echte wereld is de
ontwikkeling van de elektriciteitsprijs op lange termijn. Het model laat zien dat reke-
ning houden met toekomstige elektriciteitsprijzen ex-ante in de subsidieberekening de
totale subsidiekosten met ongeveer 15% kan laten reduceren, aangezien de betrokken
partijen de toekomstige elektriciteitsprijs waarschijnlijk zullen overschatten. De combi-
natie van ontwerpelementen die de grootste verhoging van de welvaart realiseert, is een
S AMENVATTING xxiii

volumegarantie waarbij de verkoopprijs ex-ante wordt vastgesteld, en met technologie-

specificiteit.
Grensoverschrijdende invloeden van ontwerpelementen: Gebruik makende van EM-
Lab zijn twee regio’s gesimuleerd (gebaseerd op de elektriciteitssystemen in Duitsland en
Nederland) om de effecten van verschillende ondersteuningsregelingen in buurlanden
te analyseren en de gemeenschappelijke elektriciteitsmarkt te testen op de verdelings-
effecten onder verschillende scenario’s voor interconnectiecapaciteit. Zoals de econo-
mische theorie voorspelt, blijkt dat de subsidiekosten in het kleinere land toenemen als
gevolg van van het merit order effect in het grotere buurland. Echter, de verhoging van de
subsidiekosten blijft kleiner dan de verlaging van de elektriciteitskosten als gevolg van de
interconnectie. De totale kosten in beide landen worden daarmee lager naarmate de in-
terconnectiecapaciteit toeneemt (en de onzichtbare hand zijn magische werk doet). De
resultaten tonen verder aan dat de doelstellingen voor duurzame energie in het ene land
niet alleen rekening moeten houden met de doelstellingen van (elektrisch verbonden)
buurlanden, maar ook met de beschikbare opslagcapaciteit in het systeem. Dit roept
interessante vragen op over hoe de subsidiekosten op lange termijn gealloceerd zouden
moeten worden tussen landen die onlosmakelijk met elkaar verbonden zijn.
Ontwerp voor ondersteuningsregelingen in het perspectief van de energietransitie: De
resultaten van de vorige hoofstukken worden in het laatste hoofdstuk bekeken vanuit
het standpunt van de energietransitie. Op basis van een literatuurstudie zijn drie be-
langrijke componenten geïdentificeerd: de drivers voor de kosten en opbrengsten van
duurzame energietechnologieën, het Europese systeem voor emissierechten (EU-ETS)
en de rol van flexibiliteit. Relatief hoge gemiddelde elektriciteitsprijzen hebben een di-
recte en duidelijke impact op de subsidiekosten: als de prijzen stijgen, dalen de subsi-
diekosten. Echter, als de elektriciteitsprijzen op de spotmarkt gemiddeld dalen, dan zijn
de wijzigingen van de subsidiekosten vooral afhankelijk van de kostencurves van elke
afzonderlijke technologie. Met betrekking tot het Europese systeem voor emissierech-
ten wordt het duidelijk dat hun invloed op de elektriciteitsprijzen en de subsidiekosten
lager wordt naar gelang het aandeel elektriciteitsproductie op basis van duurzame ener-
gie in de tijd toeneemt. Uit de resultaten in hoofdstuk 5 blijkt dat bij een hoog aandeel
duurzame energie de interconnectiecapaciteit een meer beperkte rol speelt in het reali-
seren van flexibiliteit terwijl (langetermijns, seizoensgebonden) opslag een sleutelrol zal
gaan spelen. Als gevolg hiervan zullen de subsidiekosten bij een hoog aandeel duurzame
energie substantieel afhankelijk zijn van de aanwezigheid van opslag in het systeem.

Conclusie De centrale vraag die in dit proefschrift is geadresseerd is: Hoe interacteren
nationale ondersteuningsregelingen met de elektriciteitsmarkt op de lange termijn (20-
30 jaar) nu de Europese Unie overgaat naar een koolstofarm energiesysteem? Wanneer
we gebruik maken van een aanpak die gebaseerd is op theoretische grondslagen van
de institutionele analyse en design theory, worden de ondersteuningsregelingen voor
duurzame elektriciteitsproductie teruggebracht tot hun ontwerpelementen vanuit een
economisch welvaartsperspectief. Modelberekeningen in EMLab tonen vervolgens aan
dat de manier waarop elektriciteitsprijzen in het ondersteuningsmechanisme worden
ingepast, een substantieel verschil maakt voor de welvaartsverdeling. Wanneer verschil-
lende ondersteuningssystemen in buurlanden worden geïmplementeerd, zorgt het me-
xxiv S AMENVATTING

rit order effect in het grote land dat de subsidiekosten in het kleinere land kunnen stij-
gen. Dit leidt tot de conclusie dat zelfs indien een ‘marktgebaseerde’ ondersteunings-
regeling kan worden ontworpen op het operationele niveau, dit geen garanties biedt
dat ook het ‘marktgebaseerde’ ontwerp op een investeringsniveau is geborgd. Gecon-
cludeerd wordt dat bij het vaststellen van duurzame energiedoelstellingen in het ene
land, de doelstellingen in buurlanden in ogenschouw moeten worden genomen, net als
de beschikbare hoeveelheid flexibiliteit in termen van interconnectie en opslag in het
systeem. Wat betreft de wetenschappelijke bijdrage zet dit proefschrift een stap naar
het integreren van institutionele analyse en design theory om de neoklassieke denkrich-
ting (die het energiebeleid in Europa aantoonbaar domineert) te complementeren. Het
raamwerk uit hoofdstuk 2 brengt twee lijnen in de literatuur tezamen: het institutionele
analyse raamwerk en design theory. Er is aangetoond dat de combinatie coherent is met
en complementair is aan het modelleringsparadigma van agent-based modelleren en
simulatie. Hoewel het raamwerk geen causale relaties biedt die inzichten geven in de
redenen waarom een speler zich op een bepaalde manier gedraagt onder bepaalde om-
standigheden, biedt het wel een fundament voor het ontwikkelen van modelsimulaties
die theorieën en ‘speculaties’ kunnen analyseren. Hoewel dit raamwerk is ontwikkeld
en toegepast op het gebied van ondersteuningsregelingen voor duurzame energie in Eu-
ropa, maakt de fundamentele aard ervan het ook inzetbaar in andere domeinen. Een
voordeel van het gehanteerde EMLab model is dat het open source is. De meeste model-
len die gebruikt worden om de Europese beleidsmaatregelen te analyseren zijn black box
simulaties. De veronderstellingen, data en mechanismen om het beleid te parametrise-
ren zijn niet publiekelijk beschikbaar. Open source modellen zorgen voor een betere
reproduceerbaarheid, meer transparantie en voor meer inzicht in hun geldigheid van
de uitkomsten, alle fundamentele uitgangspunten van de wetenschappelijke methode.
Deze dissertatie leidt ook tot nieuwe vragen die geadresseerd moeten worden naarmate
het aandeel van duurzame energie in het elektriciteitssysteem toeneemt. Wat betreft
de methodologie voor langetermijns energiestudies is het van het belang dat de gehan-
teerde modellen een voldoende (tijds)resolutie aanbieden om de effecten van intermit-
terende duurzame elektriciteitsproducten voldoende nauwkeurig te kunnen weergeven.
Daarnaast moeten ook endogene investeringen kunnen worden gemodelleerd. Dit leidt
vervolgens tot prangende vragen over de noodzaak om (een methodologie te ontwikke-
len om) subsidiekosten te verdelen tussen buurlanden. Deze vragen blijven in dit proef-
schrift liggen om later door anderen te worden opgepakt.
 S AMMANFATTNING
Målen med förnybar el, säkerställande av en inre elmarknad och elförsörjningstrygg-
het i EU är inte alltid förenliga med varandra. Nuförtiden har stora mängder av förny-
bar el i energimixen lett till oavsiktliga konsekvenser på elmarknaden. Detta är ett pro-
blem eftersom mål och stödsystem är utformade på nationell nivå, medan elmarknaden
i allt högre grad innebär en sammankoppling över nationsgränserna. Den centrala frå-
gan som tas upp i denna avhandling är hur interna nationella stödsystem för förnybar
el kommer att växelverka med elmarknaden på lång sikt (20-30 år) då EU övergår till ett
fossilfritt energisystem.
Baserat på teoretiska grunder för institutionell analys, designteori och agentbase-
rad modellering, är förnybar elproduktions stödsystem uppdelade, ur ett välfärdsekono-
miskt perspektiv, i sina designelementöch implementeras i en modell av en energimark-
nad. Energimarknadsmodellen heter Energy Market Laboratory, eller EMLab Generation
vilken huvudsakligen består av en elmarknadsklarering och en endogen investeringsmo-
dul, som gör det möjligt för forskare att studera policydesignens långsiktiga dynamik i
elsektorn.
Modellen visar att sättet att ta hänsyn till elpriserna vid utformningen av subven-
tioner skiljer sig betydligt från välfärdsfördelningen. När flera länder betraktas, skulle
spridningen av prioriteringsordningen från ett större land till ett mindre, leda till att det
mindre landets subventionskostnader ökar. Detta indikerar att även om stödet för för-
nybar el kan utformas för att vara marknadsbaseratpå operativ nivå, är utformningen av
marknadsbaseradestödsystem på investeringsnivå mycket komplicerad. Det indikerar
att ländernas mål för förnybar el bör ta hänsyn till målen i grannländerna, liksom flexi-
biliteten i elsystemet. Med tanke på det vetenskapliga bidraget tar avhandlingen ett steg
mot att integrera institutionell analys och designteori för att komplettera den nyklassiska
tankeskolan som förmodligen dominerar energipolitisk design och analys i Europa.

xxv
 1
I NTRODUCTION

1.1. B ACKGROUND
Energy systems of the world are undergoing a major transformation. Over the last few
years, the sheer magnitude of declining costs of renewable energy technologies has taken
even the most optimistic of forecasters and analysts by surprise (Carrington, 2017). This
decline in costs could be attributed to leaps in technological advancements: increase
in capacity factors, decrease in manufacturing costs, and economies of scale. However,
such a view would miss the bigger picture, since political will and far-sighted policy-
making have been key driving forces. The European Union (EU) and many countries
within Europe have been amongst the first to adopt strong support mechanisms to pro-
mote CO2 abatement technologies. The demand created by Germany for instance, set
up a global production line for manufacturing of PV panels and consequently led to a
decline in costs the world over. Whilst a determined policy-driven effort towards decar-
bonisation was taking place, parallel efforts were being made to liberalize the electricity
sector across Europe. This liberalization meant ensuring competition and affordability
through a single market for electricity across the EU.
This dissertation comprises an analysis of RES-E policies and their design in the con-
text of a liberalized, multi-national electricity market in Europe. The following sections
in this chapter present a brief historical account of renewable electricity governance in
Europe, issues that unfolded as shares of RES-E began to become significant, and the
most recent questions on the topic.

1.1.1. A BRIEF HISTORY

Energy policy in Europe has traditionally been a matter of national concern. The first
indication of a common European energy policy appeared as a consequence of the oil
crisis in the 1970s. However, the outcome was only a loose form of intergovernmen-
tal cooperation around a set of symbolic objectives related to energy security (De Jong,
2008). In the 1986 Single European Act, energy was acknowledged as a critical economic
issue within the process of integration of Europe. It is in this Act that renewable energy

1
 2 1. I NTRODUCTION

Figure 1.1: Development in EU Renewable Energy policy from 1974-2009. Source: (Hildingsson et al., 2012)

was addressed as a policy priority for the first time. The progression of EU energy pol-
icy and the development of renewable energy policy through the decades is presented
succinctly in Figure 1.1.
The early 1990s saw the first steps towards the creation of a single internal energy
market. In 1995, the European Parliament called for an action plan to further the Eu-
ropean Union’s engagement in increasing the EU-wide RES share. Around this time,
renewable energy also came to be viewed as a solution to environmental and climate
change issues, as well as a means to help increase security of supply, and economic
competition. Taking a broad political economy perspective, Hildingsson argues that
although concern for the environment motivated the promotion of renewable energy,
"advancing a policy framework has to be seen against the backdrop of the EU’s longer-
term ambition to promote liberalised and integrated energy markets" (Hildingsson et al.,
2012). The policy entrepreneurship of the Energy Commissioner and the EC during the
early 1990s is credited with developing a comprehensive and ambitious proposal for a
single internal market in gas and electricity (Nylander, 2001). In the process, from the
mid 1990s onwards, the European Commission asserted that the triad of policy objec-
tives of energy security, efficiency, and climate change could be addressed through di-
versification of energy supply, the liberalisation of energy markets, and deployment of
renewable energy (Hildingsson et al., 2012).
The 1997 white paper titled "Energy for the Future: Renewable Sources of Energy"
(COM(97)599) (Commission, 1997) outlined a goal of increasing the renewable energy
share twofold by 2010. This could be considered the beginning of EU RES policy. As
the policy strategy began to transform into a regulatory framework, it met with substan-
tial Member State opposition, especially to the binding nature of the targets. The 2001
RES-E directive therefore only established indicative targets (21% RES-E by 2010) instead
of binding ones. In the course of preparing the 2001 RES directive, although proposi-
tions were made to harmonise national support schemes, they were strongly contested
by some member states (ex: Germany and Spain).
Finally, in January 2007, the European Commission proposed the establishment of
the current framework of regulations, which aimed at increasing the share of renew-
1.1. B ACKGROUND 3

able electricity to 20% by 2020. The current regulatory framework under which support
schemes for RES-E operate, is provided by the 2009 RES-E Directive (2009/28/EC). The 1
Directive sets a 20% target for energy consumption, while relying on legally binding, na-
tional targets until 2020. The goal to promote RES-E coexists with other goals: ensuring a
single internal market for electricity, affordability of supply, and security of supply in the
European Union. These simultaneous goals are not always congruent with each other.
During this process of proposing the current framework of regulations, the principle
of nationally differentiated and binding renewable electricity targets was broadly undis-
puted (Hildingsson et al., 2012). However, harmonization of policies, as with the intro-
duction of a "Guarantees of Origin" trading scheme, remained contentious throughout
the process of its making. This was resolved after much debate by introducing interstate
statistical transfers, joint support schemes, and other mechanisms for cooperation.
The above paragraphs presented a brief history of the governance of renewable en-
ergy sources in the European Union up to the late 2000s, in the context of liberalisation
of electricity markets. In the following section, a more recent history is presented using
both academic literature, and official consultations and reports, to highlight the issues
that emerged as renewable electricity began to form a significant share of the electricity
mix.

1.1.2. W HEN R ENEWABLES A RRIVED

By the late 2000s, renewable electricity technologies had begun to contribute significant
percentages to electricity consumption in certain countries, largely due to generous sub-
sidies. In this section some of the major issues that have appeared since renewables
arrived in the late 2000s are highlighted. The issues described here primarily relate to
concerns raised both by academics and policy makers regarding interactions between
the electricity market and renewable electricity or their support schemes. What are the
impacts of renewables on wholesale prices? How can renewables be increasingly driven
by market signals? Are current institutions suitable for a decarbonised power system?
With increasing shares of renewable electricity production, more and more studies
were published on the impact of zero marginal price bids on the electricity prices of the
spot market. Sensfuss et al. (2008) were amongst the first authors who used empirical
data to demonstrate the reduction in wholesale electricity prices due to the presence of
renewables. They demonstrated that in 2004, renewable electricity decreased wholesale
prices by 2.5 Eur/MWh, and in 2005 by 4.5 Eur/MWh. Several other studies then fol-
lowed with similar objectives but with different scopes of locations, time periods, and
technologies (Cludius et al., 2014; Ederer, 2015; Gelabert et al., 2011; O’Mahoney and
Denny, 2013; Traber and Kemfert, 2009, 2011; Weigt, 2009; Würzburg et al., 2013). The
phenomenon is now well-established and popularly referred to as the merit-order effect.
Another landmark study evaluated the impact of variable renewable electricity on its
own market value (Hirth, 2013). If the market value of renewable electricity were com-
puted as the ratio of its relative price compared to the base price1 , then the authors found
that the value of wind power fell from 110% of the average power price to 50-80% as wind
penetration increased from 0-30% of total electricity consumption. This finding has im-
1 Relative value is measured as the ratio of the hourly wind-weighted average wholesale electricity price and its
time-weighted average (base price)
 4 1. I NTRODUCTION

portant implications on the economic viability of such variable renewable technologies

1 as their market share increases.
Soon after, at the European Commission, questions began to be raised about the
cost effectiveness of the support scheme designs at an EU-wide level. The 2015 doc-
ument "Launching the public consultation process on a new energy market design"
(COM(2015) 340) includes a discussion on adapting support scheme to markets (Com-
mission, 2015). It suggests that the national scope of support schemes hinders cost ef-
ficiency and that a more coordinated approach could deliver substantially higher gains
by promoting investment into renewable electricity in the most optimal geographical
locations. This argument came as no surprise: Germany had invested the most in solar
energy by 2015, while it would have arguably been more cost-efficient to have had that
investment in countries with greater sunshine. Consequently, amongst the questions
posed in the consultation document were those such as, "Should there be a more coordi-
nated approach across member states for renewable support schemes?" and "What needs
to be done to allow investment in renewables to be increasingly driven by market signals?"
In a companion report by the European Commission titled "Investment perspectives
in electricity markets", the role of the current market framework to ensure investment in
a decarbonised power system was questioned (European Commission and Directorate-
General for Economic and Financial Affairs, 2015). The report argued that the cost struc-
ture of a technology mix in a decarbonised power system exhibits decreasing average
costs and positive fixed costs. As long as this cost structure remained, and assuming
perfect competition, the report argued that it was uncertain that marginal pricing would
produce sufficient revenue to cover the fixed costs of the technologies.
In the preceding paragraphs, some fundamental issues that have emerged in relation
to the governance of electricity and the promotion of renewable energy were highlighted.
In the following sections, problems tackled in this dissertation are delineated, following
which the research objective and research questions are presented.

1.2. P ROBLEM D EFINITION AND R ESEARCH O VERVIEW

A defining characteristic of energy policy is its long-term nature. Lifetimes of invest-
ments in the sector are lumpy and commonly extend over several decades. Policies or
regulatory decisions also often have very long-term ramifications. Therefore in this the-
sis, a long-term view spanning several decades into the future is taken.
Under the current framework, renewable electricity support policies are designed
and implemented nationally. They are implemented alongside a common, international
electricity market. Under EU state aid guidelines, support should be ‘market-based’.
Therefore, the cost of RES-E subsidies are intricately related to the electricity markets
they are situated in. As outlined in the above sections, the design of policies to sup-
port renewable electricity in Europe therefore has to contend with multiple objectives of
competition, sustainability, and energy security. This reconciliation of multiple objec-
tives has been a recurring theme throughout the history of EU energy policy.
Such a reconciliation of multiple objectives took the shape of "making renewable
support schemes more market-based", "ensuring renewables are driven by market sig-
nals". However, it is not often clear what is meant by such statements in EU policy docu-
ments. What features of the support scheme are being referred to? What would it mean
1.2. P ROBLEM D EFINITION AND R ESEARCH O VERVIEW 5

for renewables solely to be driven by market signals? How would features of support
schemes impact for instance, the merit order effect and vice-versa? These issues are en- 1
capsulated in the first problem: to unravel the interactions between renewable support
scheme design and a single isolated electricity spot market, with a long- term perspective.
Since countries are now increasingly interconnected, the second major issue tackled
in this thesis concerns cross- border effects due to different renewable support schemes be-
tween neighbouring countries in a common electricity market. This addresses concerns
about the merit-order effect spreading across national boundaries, and its ensuing dis-
tributional implications.
The final issue addressed in this dissertation relates to the long term economic vi-
abiliy of electricity from renewable sources given the current institutional and physical
settings they operate in. While costs of renewable technologies have dropped dramati-
cally, effects such as their reducing market value question whether it is possible for them
to attain economic viability in a decarbonised power sector. Accordingly, this research
tackles the research questions presented below.

1.2.1. R ESEARCH QUESTIONS

The main research question addressed in this thesis is

How do national renewable electricity support schemes interact with the electricity
market over the long term (20-30 years) as the European Union transitions to a decar-
bonized energy system?

The following sub-questions together help address the aforementioned main research
question.

• How can policy design options for RES-E support in Europe be systematically and
comprehensively explored and modelled?

• How can the impact of various RES-E support design elements on the electricity
market be modelled and analysed?

• How do RES-E support policy design elements interact with a single isolated elec-
tricity market and what social welfare implications do they actualise?

• How do RES-E support policy design elements interact with an interconnected,

congested electricity market?

• How could major developments in the energy transition such as RES-E technology
cost trends, the EU ETS, and flexibility influence RES-E support?

1.2.2. S CIENTIFIC CONTRIBUTION

Since the early 2000s, when the first renewable electricity directive appeared, there has
been a vibrant debate in literature as to the most effective ways of supporting renewable
electricity generation (Huber et al., 2004; Voogt et al., 2001; Most and Fichtner, 2010; Fais
 6 1. I NTRODUCTION

et al., 2014). The methods used to perform such analysis were largely based on assump-
1 tions of perfect competition and long-term equilibrium (Capros et al., 2014). Further-
more, most literature used an approach where comparisons were made between exist-
ing policies (Fais et al., 2014; Newbery, 2016; Dressler, 2016; Winkler et al., 2016; Reuter
et al., 2012). More recently, as authors began addressing interactions between electric-
ity markets and support schemes, they proposed the idea that the key to understanding
these interactions were design features of policies, rather than policies as a whole (Held
et al., 2014; Batlle et al., 2012; del Rio and Linares, 2014). However, they have been em-
pirical observations and classifications, rather than a formal approach to policy design.
The scientific contributions emerging from this thesis can be characterised as being of
two kinds: one method-oriented, and the other, application-oriented.
From a methodological perspective, this dissertation contributes to the science of
policy design, by introducing a new modelling framework based on institutional anal-
ysis, design theory, and agent-based modelling. The modelling framework helps struc-
ture, simulate, and analyze socio-technical systems, and consequently design policies
for such systems. Based on this framework, RES-E support schemes were implemented
in terms of their design features, on an existing, agent-based model of the electricity
market, called "EMLab", short for Energy Market Laboratory. EMLab is an initiative of
the Energy and Industry section of the faculty of Technology, Policy and Management in
TU Delft. By creating the RES-E policy analysis module in EMLab, the existing literature
on methods used to analyse RES-E policies using the agent-based modelling paradigm
was also expanded.
The application-oriented scientific contributions emerged when the aforementioned
method was employed to address the research questions mentioned above. The new
modelling approach enabled arriving at insights on the impact of design features of re-
newable electricity schemes on the various actors in the electricity market: consumers,
producers, and the government, thus adding to the exisiting literature in the field. Specif-
ically, cross-border impacts of disparate national renewable support schemes operating
in interconnected electricity markets have received little attention in literature. Chapters
4 and 5 describe the experiments conducted and the insights that emerged, in detail.

1.2.3. R ESEARCH FRAMEWORK AND THESIS STRUCTURE

In order to answer the aforementioned research questions, the following structure has
been adopted. In chapter 2, the theoretical foundations for the modelling and analysis
of renewable electricity support schemes and electricity markets are established. Using a
combination of institutional analysis, the agent-based modelling paradigm, and existing
literature, a modelling framework is presented. The theoretical foundations established
are applied to empirical knowledge of current renewable electricity support schemes, to
identify a set of design elements which provide a sufficient and complete description of
a renewable support scheme from a welfare economics perspective.
In Chapter 3, the model itself is presented. This includes a conceptual representation
of the physical systems in terms of modelling entities, relationships between the enti-
ties, features, agents, their behaviours, and algorithms that represent behaviours and
policies.
Chapter 4 presents an evaluation of the impact of design elements of renewable elec-
1.2. P ROBLEM D EFINITION AND R ESEARCH O VERVIEW 7

Figure 1.2: Reading guide to the thesis

 8 1. I NTRODUCTION

tricity support schemes on a single (isolated, uncongested) region modelled similar to

1 the power sector in the Netherlands, using the long-term agent-based model of the elec-
tricity market, and RES-E support developed earlier. Effectiveness of policy is evaluated
in terms of target achievement, social welfare, and distributional implications to pro-
ducer, consumer, and the government.
In Chapter 5, the model developed earlier is applied to a two-country region, based
on Germany and The Netherlands, in order to evaluate long-term cross-border welfare
impacts of different renewable support schemes in the two countries as they operate in
an interconnected electricity market.
The aim of Chapter 6 is to evaluate factors that impact the phasing out of renewable
electricity subsidies in the long term. This is done by specifically studying three aspects
of the energy transition - drivers of RES-E subsidy, the EU ETS, and options for flexible
capacity. In doing so, this chapter puts the results of previous chapters in a larger per-
spective of the energy transition.
Finally, in Chapter 7, main conclusions of the dissertation are presented. The results,
model, and approach employed are reflected upon, and avenues for future research are
discussed.
R EFERENCES 9

R EFERENCES 1
D. Carrington, ’Spectacular’ drop in renewable energy costs leads to record global boost,
The Guardian (2017).

J. De Jong, THE 2007 ENERGY PACKAGE: THE START OF A NEW ERA? in European energy
law report V, Energy & law series No. 7, edited by U. Hammer and M. Roggenkamp
(Intersentia ; Distribution for the USA and Canada [by] International Specialized Book
Services, Antwerp ; Portland, OR : Portland, OR, 2008) pp. 96–97.

R. Hildingsson, J. Stripple, and A. Jordan, Governing Renewable Energy in the EU: Con-
fronting a Governance Dilemma, European Political Science 11, 18 (2012).

J. Nylander, THE CONSTRUCTION OF A MARKET - A frame analysis of the liberalization

of the electricity market in the European Union, European Societies 3, 289 (2001).

E. Commission, Energy for the future: Renewable sources of energy, Tech. Rep.
COM(97)599 (European Commission, 1997).

F. Sensfuss, M. Ragwitz, and M. Genoese, The merit-order effect: A detailed analysis of

the price effect of renewable electricity generation on spot market prices in Germany,
Energy Policy 36, 3086 (2008).

J. Cludius, H. Hermann, F. C. Matthes, and V. Graichen, The merit order effect of wind
and photovoltaic electricity generation in Germany 2008–2016: Estimation and distri-
butional implications, Energy Economics 44, 302 (2014).

N. Ederer, The market value and impact of offshore wind on the electricity spot market:
Evidence from Germany, Applied Energy 154, 805 (2015).

L. Gelabert, X. Labandeira, and P. Linares, An ex-post analysis of the effect of renewables

and cogeneration on Spanish electricity prices, Energy Economics 33, S59 (2011).

A. O’Mahoney and E. Denny, Electricity prices and generator behaviour in gross pool elec-
tricity markets, Energy Policy 63, 628 (2013).

T. Traber and C. Kemfert, Impacts of the German Support for Renewable Energy on Elec-
tricity Prices, Emissions, and Firms, The Energy Journal 30, 155 (2009).

T. Traber and C. Kemfert, Gone with the wind? — Electricity market prices and incen-
tives to invest in thermal power plants under increasing wind energy supply, Energy
Economics 33, 249 (2011).

H. Weigt, Germany’s wind energy: The potential for fossil capacity replacement and cost
saving, Applied Energy 86, 1857 (2009).

K. Würzburg, X. Labandeira, and P. Linares, Renewable generation and electricity prices:

Taking stock and new evidence for Germany and Austria, Energy Economics Supple-
ment Issue: Fifth Atlantic Workshop in Energy and Environmental Economics, 40,
Supplement 1, S159 (2013).
 10 R EFERENCES

L. Hirth, The market value of variable renewables: The effect of solar wind power variabil-
1 ity on their relative price, Energy Economics 38, 218 (2013).

E. Commission, Launching the public consultation process on a new energy market de-
sign, Tech. Rep. COM(2015) 340 final (Brussels, 2015).

European Commission and Directorate-General for Economic and Financial Affairs, En-
ergy economic developments: investment perspectives in electricity markets. (Publica-
tions Office, Luxembourg, 2015) oCLC: 948760359.

C. Huber, T. Faber, R. Haas, G. Resch, J. Green, S. Olz, and S. White, Deriving optimal
promotion strategies for increasing the share of RES-E in a dynamic European electricity
market, Tech. Rep. (2004).

M. H. Voogt, M. A. Uyterlinde, K. Skytte, M. Leonardi, and M. H. Whiteley, Renewable en-

ergy burden sharing. REBUS. Requirements and expectations of utilities and consumer
organisations in the European energy sector, Tech. Rep. (Energy research Centre of the
Netherlands ECN, Petten (Netherlands), 2001).

D. Most and W. Fichtner, Renewable energy sources in European energy supply and in-
teractions with emission trading, The Role of Trust in Managing Uncertainties in the
Transition to a Sustainable Energy Economy, Special Section with Regular Papers 38,
2898 (2010).

B. Fais, M. Blesl, U. Fahl, and A. Voß, Comparing different support schemes for renewable
electricity in the scope of an energy systems analysis, Applied Energy 131, 479 (2014).

P. Capros, L. Paroussos, P. Fragkos, S. Tsani, B. Boitier, F. Wagner, S. Busch, G. Resch,

M. Blesl, and J. Bollen, Description of models and scenarios used to assess European
decarbonisation pathways, Sustainable Energy System Changes 2, 220 (2014).

D. M. Newbery, Towards a green energy economy? The EU Energy Union’s transition to

a low-carbon zero subsidy electricity system – Lessons from the UK’s Electricity Market
Reform, Applied Energy 179, 1321 (2016).

L. Dressler, Support Schemes for Renewable Electricity in the European Union: Producer
Strategies and Competition, Energy Economics , (2016).

J. Winkler, A. Gaio, B. Pfluger, and M. Ragwitz, Impact of renewables on electricity markets

– Do support schemes matter? Energy Policy 93, 157 (2016).

W. H. Reuter, J. Szolgayová, S. Fuss, and M. Obersteiner, Renewable energy investment:

Policy and market impacts, Applied Energy 97, 249 (2012), energy Solutions for a Sus-
tainable World - Proceedings of the Third International Conference on Applied Energy,
May 16-18, 2011 - Perugia, Italy.

A. Held, M. Ragwitz, M. Gephart, F. ISI, M. Gephart, E. de Visser, C. Klessmann, and

Ecofys, Design features of support schemes for renewable electricity, Tech. Rep. (2014).
R EFERENCES 11

C. Batlle, P. Linares, M. Klobasa, J. Winkler, F. Isi, A. Ortner, and T. U. W. Eeg, Interac-

tions between RES-E support instruments and electricity markets, Tech. Rep. October 1
(Beyond 2020, 2012).

P. del Rio and P. Linares, Back to the future? Rethinking auctions for renewable electricity
support, Renewable and Sustainable Energy Reviews 35, 42 (2014).
 2
T HEORETICAL F OUNDATIONS :
I DENTIFYING THE DESIGN SPACE

Parts of this chapter have been published in Energy Policy 187, 228 (2017) Iychettira et al. (2017)

13
 14 2. T HEORETICAL F OUNDATIONS : I DENTIFYING THE DESIGN SPACE

2.1. I NTRODUCTION
The outcomes of a certain policy depend on far more than variables such as price and
quantity. They depend on explicit or implicit institutions which may be part of the policy
or part of the environment surrounding the policy, that shape the socio-technical sys-
2 tem. As Polski and Ostrom (1999) point out, "Institutions delimit the capacity for social
change. They are important because they are intentional constructions that structure
information and create incentives ...thereby imposing constraints on the range of possi-
ble behaviour and feasible reforms." This makes institutional analysis paramount in the
study of policy design. In addition, such analyses lend to the policy maker, a structured
set of policy design characteristics which to operate on the socio-technical system. The
challenge then lies in identifying the most essential design characteristics of a policy or
set of policies, which are sufficiently informed by their institutional setting, and evaluat-
ing their impacts on the socio-technical system.
Some studies have tried to incorporate a more comprehensive approach to RES-E
policy design, see for instance work by Bergmann et al. (2008), and Batlle et al. (2012a).
Most literature uses a "policy analysis approach" where comparisons and categoriza-
tions are made between and across different existing policies; for examples refer to Batlle
et al. (2012b), Kitzing et al. (2012),Kitzing (2014), and Fagiani et al. (2013). It is proposed
here however, that the basic unit of analysis is not the policy itself, but a set of "design
elements". Design elements refer to the detailed components that make up a certain
policy; for instance, technology specificity, location specificity, duration of support etc.
Two seemingly different RES-E support policies can be designed such that they have an
equivalent effect on the market. This idea has been upheld by several authors such as
Batlle et al. (2012a), del Rio and Linares (2014), del Rio and Mir-Artigues (2014), and
Haas et al. (2011). However, they have been empirical observations, rather than a formal
approach to policy design.
The primary objective of this chapter is to introduce a formal, structured approach
to the design of policies for stimulation of RES-E in Europe. To achieve this, we decom-
pose the objective into the following sub-objectives: (1) to identify a set of necessary and
sufficient policy design elements to incentivise RES-E in Europe, and (2) to introduce
a modelling framework to analyze the impact of policy design elements on the socio-
technical system.
In order to accomplish the above sub-objectives we introduce a formal method based
on design theory and institutional analysis to identify a policy design space, i.e., a set of
necessary and sufficient design variables that we term ’design elements’. These design
elements are identified for a certain level of analysis1 , and for a selected set of partici-
pants in the socio-technical system. Following this, a modelling framework to facilitate
the analysis of the design elements, and identify the impact of each individual design
variable on the socio-technical system is presented. The modelling framework is imple-
mented using agent-based modelling and simulation. Such a formal approach would
not only help analyse existing policies and their impact on the socio-technical system,
but also help explore the full policy design space in a structured fashion, by incorporat-
1 In Chapter 2 of Ostrom (2005) ’levels of analysis’ are described thus: “All rules are nested in another set of
rules that define how the first set of rules can be changed. . . It is useful to distinguish levels of rules that
cumulatively affect actions taken and outcomes obtained in any setting.”
2.2. T HEORETICAL F OUNDATIONS AND M ETHODOLOGY 15

ing the institutional context into the analysis.

2.2. T HEORETICAL F OUNDATIONS AND M ETHODOLOGY

The objective of this section is to introduce the methodology to achieve the objectives
outlined in Section 2.1. The section consists of an introduction to, and a description of 2
different schools of thought on which the methodology rests. It comprises three main
components: the application of design theory to policy design, the application of the In-
stitutional Analysis and Development (IAD) framework for identification of design ele-
ments, and finally, the theoretical foundation to create a modelling framework to analyse
policies in terms of their design elements.

2.2.1. T HEORETICAL F OUNDATIONS

D ESIGN T HEORY A PPLIED TO P OLICY
"Ubiquitous, necessary, and difficult" is how Bobrow (2006) qualifies the act of policy
design. Governments, irrespective of issue type, are interested in effective realization of
their goals, by applying knowledge and empirical data to assess appropriateness of alter-
natives to achieve those goals, and thus engage in ’design,’ (Howlett, 2011). The applica-
tion of (generic) design theory to policy design and policy analysis is not new. Linder and
Peters (1984) are among the earliest, while Howlett and del Rio (2013), Considine (2012),
and Taeihagh et al. (2009) are among the more recent authors who have contributed to
this topic. Read Howlett (2011) for a comprehensive review of policy design literature.

Figure 2.1: Generic Conceptual Design Framework from Herder and Stikkelman (2004)

In Taeihagh et al. (2009), an analogy has been drawn between process design and pol-
icy design, to inform transport policy. Their work is based on the theoretical frameworks
 16 2. T HEORETICAL F OUNDATIONS : I DENTIFYING THE DESIGN SPACE

of Process System Engineering. The framework used in this work, the Generic Concep-
tual Design Framework (GCDF), also has its roots in Process System Engineering.
The Generic Conceptual Design Framework has been developed collaboratively at
the Carnegie Mellon University and Delft University of Technology. It is illustrated in
Figure 2.1. This work is based on the design framework (specifically the problem def-
2 inition and conceptual design aspects) initially developed by Westerberg et al. (1997),
which draws heavily from process system engineering, and is described in detail and ap-
plied in Herder and Stikkelman (2004) and Subrahmanian et al. (2003). The framework
comprises of the following main concepts, which together, structure the content of any
level in a design process: 1. design goals; 2. design objectives (selection of goals to be op-
timized); 3. design constraints (goals that need not be optimised); 4. tests for the goals;
and 5. design space.
One may contend, as Rittel and Webber (1973) did, that for most social planning
problems or ‘wicked problems’, the concept of design is a technocratic activity and is
not applicable to policy making, as policy-making is a value-laden activity, and therefore
its appraisal is highly dependent on each participant’s personal value-set. In response,
Howlett (2011) writes that there must be a distinction drawn between ‘design’ as a verb,
and that as a noun - instead of treating design as an outcome, he urges the reader to
view it as a process of "channelling the energies of disparate actors towards agreement
in working towards similar goals in specific contexts." And that is the viewpoint that we
wish to subscribe to.

I NSTITUTIONAL A NALYSIS TO I DENTIFY G OALS AND P OLICY D ESIGN S PACE

Institutional analysis is a commonly used approach to study socio-technical systems,
and especially so in the field of institutional economics; see for instance North (1991),
Williamson (1998), and Ostrom (2005). There are several frameworks for institutional
studies to describe socio-technical systems. For a concise, yet informative overview of
the different frameworks, refer to Chapter 2 of Ghorbani (2013).
As argued in Section 2.1, institutional analysis is paramount in the study of policy
design. For the purpose of this research, we choose to employ the Institutional Analysis
and Development (IAD) framework developed and applied for several years by Ostrom
(2005). Conceptually, this framework dissects the socio-technical system into compos-
ite holons, defined as ‘a stable sub-whole in an organismic or social hierarchy which
displays Gestalt constancy’ Ostrom (2005). This conceptual foundation, of sub-wholes
and hierarchies, also corroborates with that of process design theory. Ostrom describes
the application of the IAD framework to policy design and analysis, and presents a step-
wise process for it in Polski and Ostrom (1999). It also lends itself easily to analysis by
computational social sciences such as ABMS, which help construct testable models of
socio-technical systems, as Ghorbani et al. (2010) illustrate; this is explained in greater
detail in Section 2.2.1.
Ghorbani (2013) describes the IAD thus: "This framework is an institutionally driven
tool for (1) understanding the underlying structures of a social system, (2) capturing the
operational environment, and (3) observing the patterns of interaction and outcomes,
given a set of evaluation criteria. The result of this social system analysis is used to give
feedback to the system, and as such support institutional change." The framework is
depicted in Figure 2.2.
2.2. T HEORETICAL F OUNDATIONS AND M ETHODOLOGY 17

Figure 2.2: IAD Framework from Ostrom (2005)

A note on institutional economic theories: In this work the authors put forth a frame-
work while remaining agnostic with regard to the theory that should be used; whether
it should be transaction cost economics, neoclassical economics or a combination. We
emphasize that we present and apply a framework, and not a theory, for policy design
2
In order to demonstrate the framework’s application to a model, in this particular in-
stance we adopt thories based on neoclassical economics, and also incorporate strong
assumptions of imperfect information and bounded rationality. Imperfect information
and bounded rationality are assumptions common in the institutionalist perspectives.

A GENT-B ASED M ODELLING AND S IMULATION

Agent-based Modelling and Simulation (ABMS) has established itself as being naturally
well-suited to represent socio-technical systems (Conte et al., 1998; Arthur, 2006). ABMS
is a form of computational social science that enables one to model individual enti-
ties and their interactions with the environment (Gilbert, 2004). It is then possible to
generate emergent patterns at the macro level, simply by specifying properties and in-
teractions at the micro level. They have been successfully used to implement various
socio-technical systems, including energy and industrial networks, as shown in Dam
et al. (2012). Ghorbani (2013) have shown how agent-based modelling can be used to
incorporate institutions into social simulations.

2 We adopt the definitions of a theory and a framework presented by (Ostrom, 2005, chap .2); they are repro-
duced below.
The development and use of a general framework helps to identify the elements (and the relationships among
these elements) that one needs to consider for institutional analysis. Frameworks organize diagnostic and pre-
scriptive inquiry. They provide the most general set of variables that should be used to analyze all types of
settings relevant for the framework.
The development and use of theories enable the analyst to specify which components of a framework are rele-
vant for certain kinds of questions and to make broad working assumptions about these elements. Thus, theo-
ries focus on parts of a framework and make specific assumptions that are necessary for an analyst to diagnose
a phenomenon, explain its processes, and predict outcomes.
 18 2. T HEORETICAL F OUNDATIONS : I DENTIFYING THE DESIGN SPACE

2.2.2. A P OLICY D ESIGN F RAMEWORK

Figure 2.3: A Generic Policy Design Framework

Drawing from the aforementioned frameworks, a new framework for policy design is
introduced in this section. This policy design framework maintains the basic structure
of the generic conceptual design framework, while allowing different components of the
IAD framework to inform it.
On the one hand, the design framework facilitates specification of goals and con-
straints of the policy maker, the specification of a design space, and provides a frame-
work to evaluate alternatives based on the goals. The IAD framework on the other hand
helps, decompose the socio-technical system, and specifications of interactions between
participants, and interactions between participants and the physical environment. The
latter therefore plays a paramount role in delineating the design space, understanding
and specifying possible behaviours of actors, understanding action-outcome linkages,
which can then be tested, while the former provides a structure to the process of formu-
lating the goals, and evaluating potential alternatives.
The policy design framework, called ‘A Generic Policy Design Framework’ is depicted
in Figure 2.3. The original design framework itself is depicted within dark, bold lines
in the figure, while the grey boxes and dashed lines indicate how the IAD framework
contributes to the design process. The Generic POlicy Design framework is explained
below.

1. Design goals: The intended goals of the policy to be designed are usually set by the
community itself. The concept of ’multiple levels of analysis’ described in Chapter
2 of Ostrom (2005), helps identify which participants at what level, frame these
goals, and/or constraints. According to Ostrom’s definition, the rules-in-use3 at
3 Rules which affect day-to-day behaviour of participants, in the context of the issue being analyzed.
2.2. T HEORETICAL F OUNDATIONS AND M ETHODOLOGY 19

an operational level are set one level deeper, at a ’collective’ level. This is shown
in Figure 2.4. In reality, the policy maker exists at least in two levels: the member-
state level, and at the European level. However, for the sake of illustration in Figure
2.4, it is assumed that values and objectives of the two entities are aligned. The
policy objectives therefore, are derived from the broad objectives of the European
Commission as mentioned in European Commission (2014). These objectives are 2
mentioned below. To improve the ease of associating between policy attributes
and overall objectives, we operationalize the objectives into specific ones.

• Affordability - low cost per unit production or investment

• Sustainability - effective investment in low carbon technologies and RES-E
production
• Security of Supply - also known as ‘energy adequacy’ refers to whether suffi-
cient operational capacity exists to meet demand, at any given point in time.
• Competition - preventing distortions, when multiple countries are consid-
ered, in cross border trade and investments.

Figure 2.4: Example Specification of Policy Design Framework

2. Design space: Discerning the design space requires the policy analyst to make
decisions regarding which design variables are relevant. The action situation in
the IAD framework is defined thus, "whenever two or more individuals are faced
with a set of potential actions that jointly produces outcomes, these individuals
can be said to be "in" an action situation" according to Ostrom (2005). Therefore,
the action situation outlines potential actions that participants can take, and out-
comes an action could lead to, based on their perceived notions of which actions
lead to which outcomes, called ‘action-outcome linkages’. For instance, the energy
producer must make decisions regarding which technology to invest in, where to
place the power plant, and its size. The policy maker would, at the very least, have
 20 2. T HEORETICAL F OUNDATIONS : I DENTIFYING THE DESIGN SPACE

to make decisions regarding the manner in which the remuneration is provided

- whether a price or quantity warranty 4 , whether the cost burden is distributed
among the tax payers or only the consumers, whether there is a penalty to non-
compliance. Combined across different levels of analysis (i.e., the producer at the
operational level and the policy maker at the collective level), this would present
2 a complete set of potential actions that a policy may be designed with, i.e., the
design space of the policy. This is depicted in Figure 2.4.

3. Tests: The next step in the design framework calls for developing and executing
a test. This test would simulate patterns of interactions, based on design objec-
tives, constraints and design variables discerned in the previous steps. In order to
simulate patterns of interactions that lead to outcomes, tests must include specifi-
cations of action-outcome linkages. This would mean that simulations must spec-
ify behaviour that is theoretically or empirically supported, and that participants
are expected to show given certain setting of exogenous variables. The test should
help understand and explain which outcomes are created under what different
conditions of the design space,. Agent-based modelling is a suitable approach for
creating such simulations, as described in Section 2.2.1. A detailed description of
the modelling framework created in order to simulate the impact of RES-E design
elements is presented in section 3.4.

4. Outcomes and Selection: The outcomes from the simulation help identify which
configurations of design variables (design elements) lead to desired outcomes.
With the help of the goals identified in step one, it is possible to select a config-
uration of design elements that meet the objectives and constraints of the policy
issue to be resolved.

So far, a rather general overview of the Policy Design Framework has been provided.
Central to the objectives of this paper is the identification of a closed set of design ele-
ments. Therefore, it is befitting that this aspect of the policy design framework is paid
further attention.

D ESIGN S PACE : O N I DENTIFICATION OF D ESIGN E LEMENTS

Under step 2 above, the Design Space, i.e., the set of ‘design elements’ were established
as a combination of decision variables across the relevant levels-of-analysis (collective
and operational). In order to elucidate the process of arriving at the decision variables,
and consequently the design elements, it is necessary to apply the IAD framework to the
policy issue; specifically, this would include description of the action arenas relevant to
the energy producer and to the policy maker. In Section 2.3.1, the framework has been
applied to RES-E policy making in Europe: the physical and community attributes are
outlined, followed by the action arenas themselves.
The design space is a set of design elements defined at the community-level, ie., at
the level of the policy maker, as a combination of
4 The word warranty here is used to define a promise or committment or guarantee, that a policy-maker makes
to an energy producer. For instance, a mandated quantity of electricity supply or consumption from RES
technologies would mean a quantity warranty, while a mandated price per unit of electricity generated from
RES would mean a price warranty.
2.3. A PPLYING THE POLICY DESIGN FRAMEWORK TO RES-E SUPPORT DESIGN 21

1. decision variables of the policy-maker at the community level (Decision Variables

3 and 4, in Figure 2.4), such as type of warranty, cost-burden, scheme duration etc.,
and

2. variables which indicate whether the purview of one or more of the above decision
variables further apply to each decision variable at the operational level (Decision 2
Variables 1 and 2, in Figure 2.4); for instance whether the warranty could be tech-
nology neutral or specific, location neutral or specific, and size neutral or specific.

Depending on the objectives of, and assumptions underlying the analysis, not all
design elements may be considered for evaluation. The choice of design elements for
evaluation may be strongly influenced by the frame of analysis. In an exhaustive work,
Bobrow and Dryzek (1987) highlight the different frames of analysis within policy analy-
sis; two such frames are outlined here. A welfare economics perspective would assume
a benevolent policy-maker whose only interest is to increase social welfare. A public
choice perspective posits that politicians and bureaucrats are more interested in serving
their own interests, rather than that of the public. For instance, in the particular case
of designing RES-E support, the decision regarding whether the subsidy costs are borne
by the state budget versus electricity consumers, would be much more relevant in the
public choice perspective, rather than one of welfare economics. This idea is revisited
and clarified in Section 2.3.1 and critically examined in Section 2.5.

2.3. A PPLYING THE POLICY DESIGN FRAMEWORK TO RES-E SUP-

PORT DESIGN
In this section, the Policy Design Framework introduced above is applied to the specific
case of RES-E support scheme design. In doing this, three steps are followed: the IAD
framework is first applied to the case. This forces the analyst to delineate the problem
and participants, thus creating boundary conditions, which forms the first and crucial
step towards identifying design elements. This also leads to specification of the policy
design framework introduced in Figure 2.4, to RES-E support. Subsequently, the design-
elements of RES-E support schemes are identified from literature in Section 2.3.2. Fi-
nally, a conceptualization of the agent based model for the testing and evaluation of the
different designs is presented in Section 3.4.

2.3.1. I DENTIFICATION OF PARTICIPANTS , ACTION SITUATIONS , AND EX -

OGENOUS VARIABLES
The IAD framework shown in 2.2 requires the identification of physical attributes, com-
munity attributes, and "action situations" related to investment in RES-E. Electricity,
from renewable energy or otherwise, is an excludable and subtractable commodity. RES-
E in particular is produced by installing renewable energy generating capacity, such as
for instance, solar PV panels or offshore wind farms.
While there are no specific European level targets beyond 2030 for consumption of
electricity from renewable sources, it is recognized in the 2050 roadmap that ‘power gen-
eration system would have to undergo structural change and achieve a significant level
of decarbonisation already in 2030 (57-65% in 2030 and 96-99% in 2050)’(Commission,
 22 2. T HEORETICAL F OUNDATIONS : I DENTIFYING THE DESIGN SPACE

2011). In order to realize this level of decarbonisation, policy makers design schemes to
incentivize investment in RES-E sources.
The policy maker implementing a support scheme therefore forms one action-situation.
The second action situation involves producers of renewable electricity; they are as-
sumed to be boundedly rational actors attempting to maximize their profits via actions
2 such as investments in electricity producing technologies. Their strategy for investment
in RES-E generation capacity is to make a cost benefit analysis, i.e., a net present value
calculation for each investment decision, under uncertainty. This is a stylized represen-
tation of actors, to make the analysis tractable. It must be noted however that represen-
tation is only one instantiation of the modelling framework presented.
The disadvantage of this abstraction is that it does not consider actors who might
have other motives (to be autarkic, to self-consume, to create an energy community, to
produce green energy for its own sake). On the other hand, the largest share of current
energy production comes from utility companies whose primary motive is profitability,
irrespective of differences in ownership or corporate structure. Also, even if other own-
ership structures were in place, it is hard to imagine a scenario where an actor would not
be concerned with the profitability of the project. Therefore, in the current modelling
framework, we assume that the Energy Producer agent makes an investment only when
the cost-benefit analyses indicate profitability.
Ostrom (2005) defines an action situation in terms of the following elements: par-
ticipants, positions, actions, outcomes, action-outcome linkages, information about the
action situation, and payoffs or the costs and benefits. These elements have been de-
fined, and their corresponding values have been identified for the two action situations
presented above.
However, information about future prices, and therefore revenue, and therefore whether
the decision is a viable one, is notoriously difficult to predict. Uncertainty comes in the
form of other actors’ decisions on investment, regulatory uncertainty, fuel price uncer-
tainty which impact the electricity market price and therefore the revenue, and uncer-
tainty regarding future electricity prices. It is assumed that producers’ strategy is profit
maximising, while the government’s goal is to reach the renewable target (constraint) in
an affordable (cost-effective) manner. It is also assumed that regulators have as much
information as the energy producers. And that each energy producer is aware of others’
past decisions on investment.

Table 2.1: Specification of Action Arenas

Elements of an Action Operational Action Arena Collective Action Arena

Situation, defined
Participants: Decision- Energy Producers Government(s)
making entities.
2.3. A PPLYING THE POLICY DESIGN FRAMEWORK TO RES-E SUPPORT DESIGN 23

Positions: Anonymous Energy producers are sellers A policy-maker is a partici-

slots into and out of of electricity and investors in pant with the authority to de-
which participants move. power plants. They are as- cide on which type of elec-
sumed to be boundedly ratio- tricity production is preferred,
nal, and profit maximizing. and how the remuneration is
organized. She is assumed to
be benevolent and is primar-
2
ily interested in increasing so-
cial welfare.
Action: A selection of a Act of deciding whether to Governments make decisions
setting or a value on a make an investment in a on how to incentivize RES-E.
control variable which a power plant. This would en- Based on literature on RES-
participant hopes will af- tail decisions on technology, E support in Europe, and on
fect the outcome variable. location, and size empirical evidence, the deci-
sion variables include price
warranty, quantity warranty,
contract type (risk allocation),
distribution of cost burden,
budget limits. These are de-
scribed in greater detail in
section 2.3.2.
Action-outcome linkage: A decision to invest is taken Each setting of the policy-
A setting on a control if the net present value of the maker’s decision modifies the
variable is considered investment is positive. An NPV calculation of the pro-
“linked” to a state vari- investment occurs, causing ducer in a certain way; differ-
able when it is possible a structural change in the ent combinations of settings
to use that setting to physical system, and there- lead to different repeated pat-
cause the state variable fore changing the (expected terns of interaction, which
(1) to come into being, electricity price) for the next further lead to different out-
(2) to disappear, or (3) to investment. comes.
change in degree.
Information: Access to Information about how many It is assumed that the govern-
information regarding plants have been invested into ment have the same informa-
other participants, their is available to each partici- tion as the energy producers.
positions, their action pant at any point in time.
sets, and payoffs However, information about
future electricity prices, fu-
ture fuel prices and future de-
mands, much like in reality,
are not available. The produc-
ers use forecasting techniques
for the same.
Costs and benefits Given by the financial returns Share in RES-E electricity, at
of the energy producer equa- low costs
tion.

The IAD framework, thus set-up for impacts of RES-E policies on energy investment,
is illustrated in Figure 2.5. The figure also illustrates how information from the IAD
framework feeds into the policy design framework, and populates the design space. In
the next section explanations of the design elements are provided.

2.3.2. D ESIGN S PACE : D ESIGN E LEMENTS OF RES-E S UPPORT S CHEMES

The two action arenas and potential actions at each arena were outlined in the previous
section. The design space is a collection of decision variables or potential actions that
 24 2. T HEORETICAL F OUNDATIONS : I DENTIFYING THE DESIGN SPACE

Figure 2.5: Specification of policy design framework for RES-E support

participants at two levels can make: the operational (energy producer), and the collec-
tive (policy-maker) levels. As mentioned earlier, the assumption is that the actions are
of a benevolent policy maker concerned with welfare economics. This is shown in Fig-
ure 2.5. The design elements are based on insights from literature on renewable support
schemes. Looking at the problem from the lens of a welfare economics perspective nar-
rows down literature to that extent. The formulation of design elements is also an iter-
ative process, where comparisons are made with a set of existing policies implemented.
The comparisons with empirical experiences is presented in Section 2.4.
One could then wonder how the design elements below are different from those de-
termined based on neo-classical economics. While the design elements quantity or price
warranty do indeed originate directly from neoclassical economics, many others do not.
In a neoclassical firm, the only function of management is to select profit-maximizing
quantities of outputs and inputs, which means, determining the quantity and the conse-
quent price that is established; see North (1991). Other core assumptions of neoclassical
economics include rationality (utility maximization), and a focus on the existence of an
equilibrium; read Himmelweit et al. (2011).
Other design elements mentioned in the current paper however, are not typically
considered in neoclassical economics. In order to make analyses tractable, equilibrium
models frequently make abstractions regarding perfect knowledge of costs, revenues,
and competitive positions. Therefore, design elements such as contract duration and
frequency of change in warranty do not typically feature in partial or general equilibrium
models. Another important abstraction in neoclassical economics is that firms exist to
produce an already fully defined product. The idea of diversification or specification of
products as interchangeable depending on changing institutions (such as policies on as-
2.3. A PPLYING THE POLICY DESIGN FRAMEWORK TO RES-E SUPPORT DESIGN 25

set specificity) is difficult to incorporate into the analysis as the product has already been
defined. Therefore, design elements such as asset (technology, size, location) specificity
also do not typically feature in partial or general equilibrium models. Some argue that
the institutionalist perspective itself is characterised in terms of how it differs from the
neoclassical perspective Himmelweit et al. (2011). We provide a method to incorporate
these differences and operationalize their analysis. 2
Another aspect of the way the design elements are chosen, is that they are mutually
exclusive from another; if two actions are substitutable, then they become alternative
states for one design element. For instance, either a price warranty or a quantity war-
ranty must be chosen by the welfare maximizing policy-maker, he does not set both si-
multaneously. However, he can choose technology specificity in addition to say, quantity
warranty.
Table 2.2 below lists the complete set of design elements, and their possible impacts
on the socio-technical system, as hypothesized in literature. Their impacts on the socio-
technical system is referred to with the term "action-outcome linkages" as per the IAD
framework.
Table 2.2: Design Elements

Design Element Definition Action-Outcome Linkages References

Quantity War- A mandated quantity Under a quantity warranty with no long- Batlle
ranty or Price of electricity supply term contract, investors face a substantial et al.
Warranty or consumption from price risk. With a price warranty, quan- (2012b)
RES technologies, tity of investment highly sensitive to the set
or a mandated price price.
per unit of electricity
generated from RES.
Contract w.r.t. Revenue from the When revenue is calculated ex-ante, the
electricity mar- electricity market uncertainty in future electricity price and
ket price can be accounted for consequently the revenue risk lies with
ex ante, or ex post. agent (either regulator or energy producer)
which calculates the subsidy. When cal-
culated ex-post, the electricity market rev-
enue risk lies with the subsidising agent
(government).
Contract Length The length of time Support schemes that last longer are sub- Batlle
or Project Dura- for which the support ject to lower regulatory uncertainty, which et al.
tion scheme is valid could mean lower risk for an investor. (2012a)
Technology The design element It encourages immature technologies. It Fais
Specificity which specifies may lead to more expensive technologies et al.
which technolo- being incentivized early on, but the overall (2014);
gies are eligible for cost of RES generation could be lower than Huber
a certain support a technology neutral scenario, due to the et al.
scheme. lack of windfall profits for the energy pro- (2004)
ducers.
Location Speci- This element would If higher support is given to locations with Batlle
ficity allow the differentiat- less resource endowment, RE power plants et al.
ing of support levels would be more evenly distributed in the re- (2012a)
by location. gion, which might lead to a reduced need
for grid infrastructure. However, the incen-
tive to use the best locations might be lost.
 26 2. T HEORETICAL F OUNDATIONS : I DENTIFYING THE DESIGN SPACE

Size specificity This element would Larger installations allow for economies of Batlle
allow the differentiat- scale, while incentivizing smaller sized ap- et al.
ing of support levels plications lead to more decentralized gen- (2012a)
by size. eration, and could reduce market power.
With greater smaller sized technologies,

2 distribution grid would need to be rein-

forced, and the impact on the transmission
grid is unclear.
Cost burden The cost of the RES- When financed by consumers, the scheme Batlle
E support could be is perceived as less risky as compared to et al.
borne either by the when financed by the state budget. This is (2012b)
consumers or by the because taxpayers finance is usually nego-
tax payers (state bud- tiated annually, while laws involving con-
get). sumers typically last longer. Financing
by tax-payers sets up an implicit cross-
subsidy between the tax-payers and elec-
tricity consumers. However, since RES-E
support is justified by the public good of
driving down costs so as to benefit all fu-
ture users of RES-E, an argument is that the
funds should come from general taxation .
Cost contain- Adaptation of sup- They are necessary because controlling the del Rio
ment mecha- port levels to tech- costs of RE support is argued to be abso- and
nisms nology costs and lutely vital for its political feasibility and so- Mir-
state budget related cial acceptability. Artigues
political feasibility (2014)
concerns. Ex: capac-
ity caps generation
caps, cost caps.
Penalty for non Penalties are a means Support schemes are ineffective if develop- del Rio
compliance to deter non – com- ers have the possibility to back-off with- and
pliance of the regula- out consequences. However, penalties may Linares
tion. just increase the cost as bidders could in- (2014)
clude the cost of the penalty into the bid, if
the risk of not complying is high enough.
Furthermore, penalties may deter partici-
pation of small actors.
Frequency of The number of times If the frequency of change is high, like
Change in War- the price or quantity with the Scandinavian tradable green cer-
ranty signal changes over tificate system, where the signal changes
the duration of the each year, the risk to investment increases.
support scheme. Long term contracts lead to lower prices
and they can be used to compensate low
support levels.

2.4. E MPIRICAL REPRESENTATIVENESS OF DESIGN ELEMENTS

RES-E support schemes take various forms across Western Europe - tenders, feed-in-
tariffs, tradable green certificates, and their variations and combinations. Here, six RES-
E support schemes across five countries in Western Europe have been studied and rep-
resented in terms of the design elements that were formulated in Section 2.3.2. Table 2.3
demonstrates that the design elements can indeed be used to represent, and differenti-
ate between, a variety of policies. The table below also shows that this approach lends
itself to a broader and better empirically-founded representation of policies than pure
neo-classical economics allows for.
2.4. E MPIRICAL REPRESENTATIVENESS OF DESIGN ELEMENTS 27

Table 2.3: Existing RES-E Support Schemes in terms of Design Elements. Source: Commission (2012)

Design element DE EEG DE Pre- FR Ten- NL SDE+ UK Con- Sweden

FiT mium der EOLE tract for Quota
Tariff Differ- System
ences
2
Quantity War- Price war- Price war- Quantity Quantity Quantity Quantity
ranty / Price ranty ranty warranty Warranty Warranty warranty
Warranty (Auction) (Base cost (Strike (TGC)
deter- price de-
mined termined
based on based on
auction) auction)
Contract w.r.t Ex-post Ex-ante Ex-ante Ex-post, Ex-post Remuneration
Electricity Mar- yearly solely de-
ket Price pends on
certificate
market
price
Contract Length 20 20 15 8,12,15 <15 15
(project duration years,
in years) based on
technol-
ogy
Technology technology technology technology technology technology technology
Specificity specific specific neutral neutral specific 5 neutral
Location Speci- location location location location location location
ficity neutral neutral neutral neutral neutral neutral
Size Specificity <= 10 kW <= 10 kW >12MW differs per none
technol-
ogy
Cost Burden Consumers Consumers Consumers State Consumers Consumers
Budget
Cost Contain- Quantity Max Ca- Quantity Capped Capped Capped
ment Mecha- cap of pacity cap of by bud- by quan- by quan-
nisms 52GW 400MW 52GW get (€ 4 tity tity
billion
in sring
2016)
Penalty for Non- None None None None None Exists
compliance
Frequency of Warranty Warranty Warranty Warranty Warranty TGC
Change in War- remains remains remains remains remains market
ranty per Project same same same same same cleared
during during during during during once a
project project project project project year for
duration duration duration duration duration all plants;
warranty
varies
annually

5 Source: BEIS (2016)

 28 2. T HEORETICAL F OUNDATIONS : I DENTIFYING THE DESIGN SPACE

2.5. M ERITS AND LIMITATIONS OF THE APPROACH

M ERITS
The approach presented enables us identify what actions can be taken by whom under
the current framework of rules and regulations, and therefore identifies the ‘levers and
2 knobs’ so to say, of energy policy design. Systematically identifying these levers that we
call ‘design elements’, at the level of the producers and then at the level of the national
regulations, provides the policy analysts and policy makers at the level of the European
Commission a much, much wider range of variables to use in their policy recommenda-
tions.
Understanding these ‘levers and knobs’ is especially important because over the past
several decades, the European Commission has been implementing directives towards
one common internal electricity market. At the same time, national policies and regula-
tions in related topics (such as renewable support or security of supply policies) some-
times seem to work exactly in the opposite direction; see Glachant and Ruester (2013).
The approach presented identifies more levers or variables than just quantity and price,
with the help of an institutional analysis approach and empirical evidence and thereby
assists us in resolving this dichotomy.
For instance, having technology neutral policies in one country would lead produc-
ers of non-marginal technologies to establish themselves in the same country, even when
purely in terms of wind-resources, a different country would be a better choice. In a sim-
ilar fashion, policies which shield producers from the risk of electricity price volatility in
a certain country might make that country far more attractive than another country with
no such policy, but with far better natural resources. Therefore, a design element such
as technology specificity or a method of risk allocation (ex-ante vs ex-post electricity
price setting) could severely undermine the idea of efficient resource allocation, which
the single internal electricity market promotes. The modelling framework, as demon-
strated, thus provides a method to identify which of these design elements impact effi-
cient resource allocation among different member states, and to what extent. The model
itself indicates that technology specificity vs neutrality would have a much larger impact
(60%) on subsidy costs, than the impact of price setting being ex-ante or ex-post (15%).

L IMITATIONS
Important assumptions have been made regarding characteristics of the participants
and the action situations that they might find themselves in. For instance, a producer
makes decisions mainly regarding economic and physical aspects of the technology. In
reality however, there are other action situations through which policy makers could be
eventually influenced. For instance, if a severe penalty or taxes were to be introduced,
workers could organize a protest. Or if a technology were to be completely banned, as
nuclear energy has been in Germany, producers could file a lawsuit against the govern-
ment like Vattenfall did. Therefore, the analysis is limited in that the design space does
not include say, the ’political man’ or the ’emotional man’, but mainly focuses on the
’rational man’, although boundedly so. In that sense, the analysis so far could be charac-
terised as being technocratic. Indeed, if the energy producer agent were to assume mul-
tiple identities, such as being politically active and strongly pushing for local autarky, the
design elements would be different. Another limitation of the approach as presented, is
2.6. C ONCLUSIONS AND POLICY IMPLICATIONS 29

its computationally intensive nature; further attention could be paid to the process of
reducing the number of design elements to suit computationally constrained situations.
Despite the limitations, within the action situation and roles outlined in this analysis,
and from a welfare economics paradigm, the approach presented provides a methodol-
ogy for creating, simulating, and testing a complete policy design space.
2
2.6. C ONCLUSIONS AND POLICY IMPLICATIONS
Energy policy design in Europe is a complex issue: not least because of the co-existence
of a common European policy, along with very disparate policies at the member state
levels. The policy maker is faced with the daunting challenge of analysing multiple ac-
tors, multiple decision criteria, at multiple levels of operation and/or governance. Us-
ing a combination of design theory, institutional analysis, and agent-based modelling
(ABM), we provide a method to systematically explore policy design options for RES-E
support in Europe. This is done firstly by identification of decision variables, which then
lead to the design elements of a policy, and secondly by evaluating the impact of each
design element on the socio-technical system using an agent-based model.
Given a certain frame of analysis, we propose that it is theoretically possible to iden-
tify the complete policy design space. Crucially, this aspect potentially opens up to the
policy analyst new avenues for intervention, and allows her to explore, given a range of
uncertainties, which element(s) of intervention is(are) the most vital to achieve goals of
the community. The applicability of the approach is demonstrated by representing and
differentiating between six renewable electricity support schemes from Western Europe
in terms of the design elements. The applicability of the modelling framework using
ABM, and consequently the Design Element Approach, is demonstrated by evaluating
the long-term, dynamic impact of three design elements: price warranty versus quan-
tity warranty, technology neutrality versus specificity, and accounting for the electricity
market price ex-ante versus ex-post on the Dutch electricity sector. A vital result, demon-
strated in Chapter 4 and described in section 4.3.1, is that technology specificity leads to
making the scheme 60% more cost effective than technology neutrality.
It is important to note here that claims of completeness of the design space come
with limitations. For instance, if the energy producer agent were to assume multiple
identities, such as being politically active and strongly pushing for local autarky, the de-
sign elements would be different. The design framework published here therefore per-
tains mainly to an analysis which lies within the scope of welfare economics, although
founded firmly within an institutional framework and empirical experience. Other lim-
itations of the approach include its computationally intensive nature, and the need to
prudently select the most important design elements necessary for the analysis.
Avenues for future work are many. The foremost of these involve demonstrating
with modelling, the application of this framework to understand impacts of different re-
newable electricity policy designs in neighbouring countries sharing one common elec-
tricity market, on cross-border welfare effects. Such work would pave the way towards
quantitatively understanding whether and how renewable support scheme designs in
neighbouring member states should be harmonised, in view of the common electricity
market. This analysis is being performed for a forthcoming paper. Other possibilities
for future work include designing an endogenous policy-maker agent, who dynamically
 30 2. T HEORETICAL F OUNDATIONS : I DENTIFYING THE DESIGN SPACE

changes values of design variables based on indicators in the model. The most challeng-
ing avenue for further exploration would be to identify a design space involving agents
with more than just economical considerations and identities, but are more complex
involving perhaps political and cultural considerations as well.
The implications of this work are, from the perspective of the authors, most useful
2 for policy makers of RES-E support schemes, at both the member-state and at European
levels. Given that governance of renewable energy support beyond 2020 at the European
level is still undefined, while a European target for renewables has been set, this work
paves the way for a more comprehensive, formal, empirically founded analysis of RES-E
policy design than what currently prevails.
R EFERENCES 31

R EFERENCES
K. K. Iychettira, R. A. Hakvoort, P. Linares, and R. d. Jeu, Towards a comprehensive policy
for electricity from renewable energy: Designing for social welfare, Applied Energy 187,
228 (2017).

M. M. Polski and E. Ostrom, An Institutional Framework for Policy Analysis and Design,
2
(Indiana University, Bloomington, IN, 1999).

J. Bergmann, C. Bitsch, V. Behlau, and S. Jensen, A European harmonised policy to pro-

mote RES-electricity - sharing costs and benefits, Tech. Rep. D17 (Vienna University of
Technology, 2008).

C. Batlle, P. Linares, M. Klobasa, J. Winkler, F. Isi, A. Ortner, and T. U. W. Eeg, Interac-

tions between RES-E support instruments and electricity markets, Tech. Rep. October
(Beyond 2020, 2012).

C. Batlle, I. J. Pérez-Arriaga, and P. Zambrano-Barragán, Regulatory design for RES-E sup-

port mechanisms: Learning curves, market structure, and burden-sharing, Energy Pol-
icy Modeling Transport (Energy) Demand and Policies, 41, 212 (2012b).

L. Kitzing, C. Mitchell, and P. E. Morthorst, Renewable energy policies in Europe: Con-

verging or diverging? Energy Policy 51, 192 (2012).

L. Kitzing, Risk implications of renewable support instruments: Comparative analysis of

feed-in tariffs and premiums using a mean–variance approach, Energy 64, 495 (2014).

R. Fagiani, J. Barquín, and R. Hakvoort, Risk-based assessment of the cost-efficiency and

the effectivity of renewable energy support schemes: Certificate markets versus feed-in
tariffs, Energy Policy Special section: Long Run Transitions to Sustainable Economic
Structures in the European Union and Beyond, 55, 648 (2013).

P. del Rio and P. Linares, Back to the future? Rethinking auctions for renewable electricity
support, Renewable and Sustainable Energy Reviews 35, 42 (2014).

P. del Rio and P. Mir-Artigues, Combinations of support instruments for renewable elec-
tricity in Europe: A review, Renewable and Sustainable Energy Reviews 40, 287 (2014).

R. Haas, C. Panzer, G. Resch, M. Ragwitz, G. Reece, and A. Held, A historical review of

promotion strategies for electricity from renewable energy sources in EU countries, Re-
newable and Sustainable Energy Reviews 15, 1003 (2011).

E. Ostrom, Understanding Institutional Diversity (Princeton University Press, Princeton,

New Jersey, 2005).

D. B. Bobrow, Chapter 5: Policy Design: Ubiquitous, Necessary and Difficult, in Handbook

of Public Policy (SAGE Publications, 2006).

M. Howlett, Revisiting Policy Design: The Rise and Fall (and Rebirth?) of Policy Design
Studies, (Reykjavik, Iceland, 2011).
 32 R EFERENCES

S. H. Linder and B. G. Peters, From Social Theory to Policy Design, Journal of Public Policy
4, 237 (1984).

M. Howlett and P. del Rio, Policy Portfolios and their Design: A Meta Analysis, (Grenoble,
France, 2013).
2
M. Considine, Thinking Outside the Box? Applying Design Theory to Public Policy, Politics
and Policy 40, 704 (2012).

A. Taeihagh, R. Banaras-Alcantara, and M. Claire, Development of a novel framework for

the design of transport policies to achieve environmental targets, Computers & Chemi-
cal Engineering 33 (2009).

P. Herder and R. M. Stikkelman, Methanol-Based Industrial Cluster Design: A Study of

Design Options and the Design Process, Industrial & Engineering Chemistry Research
43, 3879 (2004).

A. W. Westerberg, E. Subrahmainan, Y. Reich, and S. Konda, Designing the process de-

sign process, Computers & Chemical Engineering 21, Supplement, S1 (1997), supple-
ment to Computers and Chemical Engineering6th International Symposium on Pro-
cess Systems Engineering and 30th European Symposium on Computer Aided Process
Engineering.

E. Subrahmanian, A. W. Westerberg, S. Talukdar, J. Garett, A. Jacobson, C. Paredis,

A. Christina, P. Herder, and T. Adam, Integrating Social Aspects and Group Work Aspects
in Engineering Design Education, International Journal of Engineering Education 19
(2003).

H. W. J. Rittel and M. M. Webber, Dilemmas in a general theory of planning, Policy Sci-

ences 4, 155 (1973).

D. C. North, Institutions, Institutional Change and Economic Performance (1991).

O. E. Williamson, Transaction Cost Economics: How It Works; Where It is Headed, De

Economist 146, 23 (1998).

A. Ghorbani, Structuring Socio- technical Complexity, Ph.D. thesis, Delft University of

Technology (2013).

A. Ghorbani, A. Ligtvoet, I. Nikolic, and G. Dijkema, Using institutional frameworks to

conceptualize agent-based models of socio-technical systems, Proceeding of the 2010
workshop on complex system modeling and simulation 3, 33 (2010).

R. Conte, N. Gilbert, and J. S. Sichman, Multi-Agent Systems and Agent-Based Simula-

tion: First International Workshop, MABS ’98, Paris, France, July 4-6, 1998. Proceedings,
(Springer Berlin Heidelberg, Berlin, Heidelberg, 1998) pp. 1–9.

W. B. Arthur, Chapter 32 Out-of-Equilibrium Economics and Agent-Based Modeling, (El-

sevier, 2006) pp. 1551 – 1564.
R EFERENCES 33

N. Gilbert, Agent-based social simulation: dealing with complexity, The Complex Systems
Network of Excellence 9, 1 (2004).

K. H. V. Dam, I. Nikolic, and Z. Lukszo, Agent-based modelling of socio-technical systems,

May (Springer, 2012).
2
European Commission, A policy framework for climate and energy in the period from
2020 to 2030, (2014).

D. B. Bobrow and J. S. Dryzek, Policy Analysis by Design (Pittsburgh University Press,

1987).

E. Commission, Energy Roadmap 2050, Tech. Rep. COM/2011/0885 (Brussels, 2011).

S. Himmelweit, R. Simonetti, and A. Trigg, Microeconomics: Neoclassical and Institution-

alist Perspectives on Economic Behaviour, 1st ed. (Cengage Learning, 2011).

B. Fais, M. Blesl, U. Fahl, and A. Voß, Comparing different support schemes for renewable
electricity in the scope of an energy systems analysis, Applied Energy 131, 479 (2014).

C. Huber, T. Faber, R. Haas, G. Resch, J. Green, S. Olz, and S. White, Deriving optimal pro-
motion strategied for increasing hte share of RES-E in a dynamic European electricity
market, Tech. Rep. (2004).

E. Commission, RES Legal Europe Database for Regulation on Renewable Energy Gener-
ation, (2012), rES LEGAL Europe is the website on regulations on renewable energy
generation. In the database you will find information on the important legislation on
the support schemes, grid issues and policies for energy from renewable sources cov-
ering all three energy sectors: electricity, heating & cooling and transport. The scope
of this database covers all the EU 27 Member States, the EFTA Countries and some EU
Accession Countries. The website offers links to all relevant original legislation, is free
of charge and is updated on a regular basis.

BEIS, Contracts for Difference: An explanation of the methodology used to set adminis-
trative CFD strike prices for the next CFD allocation round, Tech. Rep. (Department for
Business, Energy and Industrial Strategy, 2016).

J.-M. Glachant and S. Ruester, The EU internal electricity market : done forever?, Working
Paper (2013).
 3
M ODELLING : EML AB G ENERATION
AND R ENEWABLE S UPPORT

Parts of this chapter have been published in Applied Energy 106, 169 (2017) Iychettira et al. (2017)

35
 36 3. M ODELLING : EML AB G ENERATION AND R ENEWABLE S UPPORT

3.1. I NTRODUCTION
In the previous chapters the problem was described, and theoretical, conceptual foun-
dations to perceiving the problem and evaluating alternatives were established. The ob-
jective of this chapter is to describe the "testing enviroment", or the model that is used
to evaluate various policy options over the course of this dissertation.
The Energy Modelling Laboratory, known as EMLab, is an open source energy mod-
elling project initiated at the Energy and Industry section of the Technology Policy Man-
agement faculty at the Delft University of Technology. Multiple doctoral and master’s
3 students have learnt from and contributed to this project over the past several years. This
dissertation builds on the shoulders of those proverbial giants who have contributed to
EMLab before me, and have helped me render my own contributions to it.
The initial conceptualisation and implementation of EMLab-Generation was carried
out by de Vries et al. (2012). The AgentSpring agent-based modelling framework created
by Chmieliauskas et al. (2012) provided the software architecture platform for the model.
Work on power market transition models by Chappin (2011) also contributed to the con-
ceptualisation and implementation of the core modules of EMLab. Richstein (2015) fin-
ished its implementation and verified and validated the model. Several conceptual and
practical improvements were made: the investment algorithm was improved, banking
was incorporated into the CO2 market. This chapter seeks to describe in detail the parts
of existing EMLab that have been used for this work, and the additional parts that were
especially built to address the research questions of this dissertation.
To help the reader peruse relevant sections, a short reading guide is presented here.
Firstly, based on the analysis from the previous chapter, the system to be modelled is de-
lineated, and the main agents whose behaviours are modelled are presented in Section
3.2. Secondly, the parts of EMLab that existed prior to the start of this work, and that
are relevant to this work are descibed in Section 3.3. Finally, the parts of EMLab, that
were created by the author specially for the purposes of her doctoral programme, i.e.,
the models of the different RES-E support schemes are described in 3.4.

3.2. S YSTEM DECOMPOSITION OF B ASE M ODEL IN EML AB

This section follows directly from Section 2.3.1 in the previous chapter, where partici-
pants and action situations were identified. In this section, a testing environment de-
signed to evaluate the impact of design elements that were identified in the previous
chapter, is described. The testing environment simulates an electricity market, producer
agents taking investment decisions, and a ’government’ agent implementing RES-E sup-
port policies in terms of their design elements. A brief description of the conceptualiza-
tion of the model is presented here, to demontrate one instantiation2 of the framework
presented in the previous chapter. The schematic representation of the testing environ-
ment for the policy design framework is presented in Figure 3.1. The model consists of
a ’base model’ where energy producers’ decisions are simulated, and an ’RES-E support
scheme model’.

2 The word instantiation is used to mean: ‘(of a universal or abstract concept) have an instance; be represented
by an actual example.’
3.2. S YSTEM DECOMPOSITION OF B ASE M ODEL IN EML AB 37

Figure 3.1: Design space is evaluated by simulating patterns of interactions1 .

3.2.1. M ODEL CONCEPTUALISATION

Agents are the basic units of an agent-based model. They are characterised by their
states, behaviours, and ability to interact. Entities that are not capable of decision-
making are objects. Agents can interact with objects.
For instance, the main agents are electricity generation companies; they make short
term decisions such as bidding competitively into the electricity spot market (an object),
purchasing fuel, and long term decisions such as investing in new power plants. They af-
fect the model environment with such decisions, and consequently their own state (e.g.
cash position). The other agents are regulators. They make decisions regarding the de-
sign of renewable support scheme. For instance, in a price warranty scheme, the regula-
tor agent decides the amount of subsidy based on his knowledge of the costs of the tech-
nology. For a quantity warranty scheme, the agent decides the quantity, and organises
an annual market clearing process, in which energy generation companies participate.
The environment is composed of objects such as the electricity market, the annual
load duration curve, the renewable support schemes, power plants of various technolo-
gies, locations which correspond to electricity market clearing zones. Each of these ob-
jects is characterized by its own set of properties. For instance, the electricity market
object has the properties clearing price and clearing volume, instantiated for every year
(tick), and for every segment of the load duration curve. The main agents and objects in
EMLab are schematically represented in Figure 3.2.

3.2.2. M ODEL IMPLEMENTATION

The model has been implemented in Java, which is an object oriented programming
language. The EMLab-Generation model identifies different types of Java classes and
other files, described below. Classes are grouped into packages in the following manner.

• Domain classes are the definitions of things and their properties. Classes such as
Energy Producer and EnergyConsumer under package Agent, and classes such as
Bid and ClearingPoint under package Market are examples.
 38 3. M ODELLING : EML AB G ENERATION AND R ENEWABLE S UPPORT

Figure 3.2: System represented as agents and objects

• Role classes capture behaviour, executed by specific classes from the domain pack-
ages. For instance, the agent EnergyProducer acts the SubmitOffersToElectricity-
MarketRole.

• Repository classes contain functions that deal with interaction of model code and
the database. They also assist in updating current information and storing new
information.

• Scenario xml files consist of all the data required to define and initialize a simula-
tion run. It contains data as well as relations between objects.

While this section described the what and the who, i.e., things, objects and agents,
the following sections are a description of the model narrative. The model narrative is
a story that relates the behaviours of the agents with time, i.e., it explains which agent
does what with whom and when.

3.3. E LECTRICITY SPOT MARKET AND I NVESTMENT

This section briefly describes the ’base model’, on which the design elements are built.
The base model includes two main algorithms: 1) an electricity spot market clearing
algorithm, and 2) the investment algorithm. A brief description of the two algorithms is
presented below. They are complemented by flowcharts.

3.3.1. M ARKET CLEARING

A uniform electricity market clearing has been implemented algorithmically. The load
duration curve for a full year is represented in terms of 20 load segments, where each
load segment is a demand (in MW) and time (in hours) pair. For each load segment, the
bids (price, quantity pairs) from the energy producer are stacked according to their merit
3.3. E LECTRICITY SPOT MARKET AND I NVESTMENT 39

order, and a uniform market clearing price is determined at the intersection of demand
and supply for that load segment. The two flowcharts in this section indicate the main
algorithmic processes in EMLab. Market clearing within one tick (year) is performed
using an annual load duration curve. The time resolution is indeed yearly. However, the
annual load duration curve, comprising 8760 hours of different loads, is approximated
into twenty segments in view of computational resource constraints. Each segment is
represented by a pair of values: a load (in MW), and period (in hours). For instance,
segment 1 is (8160.778 MW, 17h), segment 2 is (8390.36, 77h) and so on. For each load
segment, the electricity spot market is cleared individually according to uniform price
clearing, and price volume pairs are determined for each of the 20 load segments. 3
Start
Invest into new capacity

For each load segment

Current data
for
forthcoming
plants
Determine load, energy producer Gather alternatives
submits offers to electricity market among technologies
Physical
constraints:
resource
Clear load segment by potential
uniform market clearing Forecast expected
demand and fuel prices
Past data on
for future base year by
demand and
simple regression
fuel price
trends

Determine expected
Run RES-E Support Scheme
revenues and costs for
future base year

For each energy producer No attractive

alternatives Select power
exist plants with NPV
Invest into new capacity >0

Attractive
alternatives
exist
Select power
Stop plants with
Increment tick, update system data highest NPV

Choice

Is tick<40?
Create new power plant,
Yes
start construction, pay
No downpayment

End

(a) Main EMLab Algorithm (b) Investment Algorithm

Figure 3.3: Flowcharts showing the overall EMLab algorithm, and the investment algorithm.

3.3.2. DYNAMIC DETERMINATION OF LOAD DURATION CURVE

The intermittent nature of the system is represented for renewable electricity generation
using actual hourly capacity factors, and for demand with hourly load data. The hourly
load data is then aggregated into twenty segments, and called the load duration curve.
The generation from renewable electricity is reduced from the load duration curve, leav-
ing the ‘residual load duration curve’ to be cleared against the remaining supply. Further
investment in generation capacity of intermittent RES-E leads to changes in the residual
load duration curve, which needs to be reflected correspondingly.
 40 3. M ODELLING : EML AB G ENERATION AND R ENEWABLE S UPPORT

Figure 3.4: Dynamic determination of load duration curve

This dynamic determination of the load duration curve is demonstrated in the flowchart
presented in Figure 3.4. The figure shows that the residual load duration curve is deter-
mined for each price zone separately. If spillage3 exists in all zones, there is no exchange
between the zones. If spillage exists in either zone, it is exported to the other country
upto the extent that the interconnector capacity allows it. Hourly values of load and
supply are then aggregated into twenty load segments and the electricity prices are de-
termined in each zone.

3.3.3. I NVESTMENT IN GENERATION CAPACITY

Every agent makes decisions about investments of plants by forecasting demand and
fuel prices based on past data, thereby estimating their own merit order, and future elec-
∗
tricity prices p s,t +n . Producers differ from each other in terms of the initial mix of their
generation portfolios, and the order in which they take investment decisions. Each agent
considers demand and fuel price data of the previous 5 years to create geometric regres-
sion trends for the future. The future time point, n, for which they make investment
decisions is 2 years ahead. They do have perfect knowledge only about investments
made thus far by the other agents, and when they will come online. That the agents

3 Spillage is the RES-E production that is greater than the sum of the demand and interconnection capacity.
3.3. E LECTRICITY SPOT MARKET AND I NVESTMENT 41

have a limited knowledge of the future is an important feature of the model, as it leads to
sub-optimal decisions being made. This corresponds to reality where expectations often
differ from actual outcomes, as explained by Richstein et al. (2014).
Based on the expected electricity market prices, marginal costs vc g ,t +n ,the fixed op-
eration and maintenance cost f c g ,t +n , segment-dependent available capacity of power
plant a g ,s , and the expected running hours r s,g ,t +n ,which is also calculated from the ex-
pected electricity prices and marginal cost per segment, the cash flow for reference year
t + n of operation for the power plant is calculated as follows.

3
C FOp,g = C I n f l ow Op,g −COut f l ow Op,g
∗
= Σs p s,t +n × r s,g ,t +n × a g ,s (3.1)
! "
− Σs vc g ,t +n × r s,g ,t +n × a g ,s + f c g ,t +n

The economic viability of each power plant of capacity K g , is then assessed with ini-
tial capital costs,I g , over the building period 0..t b , and the service period, t b +1...t b +t D .
The Weighted Average Cost of Capital (WACC) is used as the discount rate. The Net
Present Value (NPV), which discounts all future costs and benefits into present value, is
calculated by each energy producer for each technology in order to make an investment
decision:

# −I g
t
N PVg = Σtb=0
(1 + W ACC )t
(3.2)
t b +t D C I n f l ow Op,g COut f l ow Op,g $
+ Σt =t − /K g
b +1 (1 + W ACC r ev )t (1 + W ACC )t

W ACC = r D × (D/V ) + r Eb × (E /V ) (3.3)

W ACC r ev = r D × (D/V ) + (r Eb + r E p ) × (E /V ) (3.4)

Where D is the debt value, E is the equity value, V is the total value. The debt equity
ratio is set at 70:30. In Equation (3.2), risk aversion to price volatility is incorporated in
the inflow or revenue component by an adjusted W ACC , called the W ACC r ev . The rate
of equity component in the W ACC r ev , described in Equation (3.4), r E , is expressed as the
sum of a basic equity rate, r Eb set to 11%, and a price risk equity rate, r E p set to 3%.The
cost of debt, r D is set at 5.5%. This is based on data from DiaCore (2016).
Each agent thus iteratively computes the NPV for every technology, and invests in the
technology with the highest positive NPV. This algorithm is presented in Figure 3.3b. This
description so far forms the base model, on which RES-E support design elements have
been built. The conceptual model of RES-E policies is explained below. The model is
implemented in Java and the source code is openly accessible4 . It is important to note at
this stage, that economic decommissioning of the power plants has not been modelled;
the plants are operational through the whole of their technical lifetimes, even if their
marginal profits are negative.
4 https://github.com/Kaveri3012/emlab-generation/tree/feature/SocialWelfareAnalysis
 42 3. M ODELLING : EML AB G ENERATION AND R ENEWABLE S UPPORT

3.4. M ODELLING RENEWABLE ELECTRICITY SUPPORT

In the previous section, the main algorithms of EMLab have been described. In this
section I present the methodology for modelling of RES-E schemes in the form of their
design elements. The design elements follow from those identified in 2.2 of this the-
sis. By specification and configuration of design elements, the policy maker attempts to
achieve outcomes of affordability, and sufficient share of RES-E in the energy mix. The
modelling framework is represented structurally using a UML class diagram, presented
in Figure 3.5. The structure of the model indicates that RES-E Support Scheme is a class,
3 whose attributes are the design elements identified in the previous step. The algorith-
mic relationship between the RES-E schemes and the investment behaviour is indicated
with the help of a flowchart in Figure 3.6.

Figure 3.5: Specification of Java class structures of Agents and RES-E scheme using design elements.

At the outset, it must be noted that three design elements have been modelled, while
keeping the others constant. A simplification to three design elements allows for clarity
in interpretation, is sufficient for demonstration of the framework, and is a strong first
step towards incorporating more design elements. Due to these reasons, and in order to
keep within time and other resource constraints, we settled with modelling only three
design elements.

Figure 3.6: Relationship between base model and RES-E schemes

As far as modelling is concerned, an RES-E scheme is represented as an entity with

a set of properties, and related methods, much like a ’class’ in object-oriented program-
3.4. M ODELLING RENEWABLE ELECTRICITY SUPPORT 43

ming. The design elements identified in the previous step together make up the proper-
ties of the RES-E class. This is represented in 3.5. The processes or behaviours related to
the different properties are the ’methods’ of the class. The source code is openly acces-
sible 5 .
Figure 3.7 shows four of the eight possible inherited RES-E Schemes for three design
elements. The quantity based schemes include a function or a method to organize auc-
tions based on the other design elements of the scheme, such as technology specificity,
location specificity, contract (ex-post or ex-ante). The price based schemes includes a
function to compute the remuneration, depending on specified design elements. A high-
level flow chart of the process flow in the model is presented in Figure 3.6. 3
Different representations of the RES-E Support Schemes are be inherited6 , and con-
tain processes that are functions of design elements. Other classes in the model repre-
sent the agents Energy Producer and Government, and their decision-making processes.

Figure 3.7: Specification of Class Structure

3.4.1. M ODELLING OF D ESIGN E LEMENTS

The objects characterising the RES-E support schemes were introduced in the previous
section. This section consists of the processes, or the sequences of action that imple-
ment the eight RES-E support schemes, in EMLab-Generation.

P RICE VS QUANTITY WARRANTY

This design element can be defined as a mandated quantity or price for electricity supply
or consumption from RES technologies. It is modelled as two separate algorithms, their
descriptions are as follows.

Quantity Warranty Scheme The quantity warranty scheme, is algorithmically imple-

mented in the form of yearly auctions, as per the following steps.
5 https://github.com/Kaveri3012/emlab-generation/tree/feature/SocialWelfareAnalysis
6 Inheritance is the defining of new classes as extensions of existing classes: the existing class is the parent class
and the new class is the child class.
 44 3. M ODELLING : EML AB G ENERATION AND R ENEWABLE S UPPORT

1. Quantitative targets for renewable energy generation are exogenously, for each
year, set by extrapolating the targets mentioned in the National Renewable Energy
Action Plan of Economic Affairs Agriculture and Innovation (2010). This comprises
the demand side of the auction.
2. The quantity warranty is implemented as a sealed bid uniform price auction, for
contracts that span a pre-decided period of years7 , like a tender 8
3. Depending on the specification of design element 3, technology specificity, annual
auctions are organized for each technology separately or for all technologies si-
multaneously.
3 4. Producer agents submit bids each year for new projects, by computing the ex-
pected cost and benefit of the project either by Equation (3.5) or (3.7), depending
on whether the scheme is designed ex-post or ex-ante.
5. The payments are then made annually for the winning bids for the duration of the
contract period (20 years) according to Equation (3.6) or (3.9).

Price Warranty Scheme

1. The price warranty is computed by matching the exogenously specified inelastic

target on the demand side, with the (cost, quantity) pairs on the supply side9 .
2. The ‘regulator’ agent depending on specification of design element 3, computes a
price warranty for each eligible technology, or a single price for all technologies if
the scheme is technology neutral.
3. The price, with ex-ante considerations of electricity market price, is computed as
per Equation (3.5) and (3.6), and with ex-post considerations of electricity market
price is computed as per Equation (3.7) and (3.9).
4. Investment decisions are made by each energy producer taking into account pub-
lished revenue from the applicable subsidy schemes. Payments are made annually
till the end of the contract duration (20 years) according to Equation (3.6) or (3.9).

T ECHNOLOGY N EUTRALITY VS . S PECIFICITY

In the technology-specific scenarios, a different quantity X is calculated for each technol-
ogy. When technology specificity is applied with quantity warranty of design element 1,
a different auction is cleared for each technology by the regulator agent, resulting in one
X for each technology type, where supply and demand meet. Inelastic RES-E production
targets (demand-side) are set for each technology type at each year exogenously. Pro-
ducer agents compute their offer prices for each available technology type in the model,
either by Equation 3.7 or 3.5, and submit it to the auction. In a price warranty scheme,
the regulator agent is assumed to have the same information on costs, and assumptions
regarding discount rates, as the producer agent. Again, the regulator agent consequently
determines the quantity X for each technology.
In the technology-neutral scenarios, a single quantity X is calculated irrespective of
the technology type. In a quantity warranty scheme, a single auction is conducted for all
7 The duration of contract is 20 years.
8 This step is approximately modelled on the French EOLE auctions (Laali and Benard, 1999).
9 It is assumed that the regulator has full knowledge of power plant costs and realistic technology potentials.
3.4. M ODELLING RENEWABLE ELECTRICITY SUPPORT 45

technologies. In a price warranty scheme, the regulator agent is assumed to have infor-
mation regarding technology costs and technology potentials. With this knowledge and
given the exogenously set RES-E target, the agent constructs a supply-demand curve,
and computes a single quantity X for all technologies.

E X - ANTE VS . EX - POST PRICE SETTING

The contract can be designed in a way that for computing the subsidy i.e., the additional
remuneration for RES-E technologies, revenue from the electricity market is accounted
for either ex-ante (before the actualization of electricity prices) or ex-post (when the
electricity price is known). This process of organizing the remuneration takes place in
3
two steps. A first step is where supply and demand are matched, to arrive at a quan-
tity X , and a second step where payment is made to the energy producer, based on the
amount of generation each year. It is important to note that this quantity X holds differ-
ent meanings in ex-post and ex-ante versions of remuneration.

Ex-ante In this version, the revenue from the electricity market is taken into account
X ant e g
ex-ante for the calculation of the remuneration. In the first step, a quantity Σt =0..d (1+W ACC )t
equivalent to the total subsidy required by a plant is computed. As can be seen in Equa-
tion 3.5, this quantity is computed in the following way: the estimated revenue is sub-
tracted from the sum of the discounted value of investment cost and operating cost. The
annual payment to eligible power plants is organized by Equation 3.6. This way, the risk
of volatility of future electricity prices is relegated to the producer.

X ant e g t Ig
Σt =0..d = Σtb=0
(1 + W ACC )t (1 + W ACC )t
# C I n f l ow Op,g
t b +t D
− Σt =t (3.5)
b +1
(1 + W ACC r ev )t
COut f l ow Op,g $
+
(1 + W ACC )t

pa yment g ,t = Σg Σs (X ant e g × a g ,s ) wher e, t ∈ {t b ...t D } (3.6)

Ex-post In this version the electricity market prices are accounted for after the prices
have been realised in actuality. Since the subsidy is only paid once the electricity price is
known, the only quantity that needs to be published ahead is the ‘total cost per unit’of
the technology, variously known as the ‘base cost’or ‘strike price’in the different sup-
port schemes that implement ex-post remuneration. In the model, this is implemented
in two steps; in the first step, a quantity equivalent to the total discounted cost (fixed
X post g
and variable) of a plant, represented by the term Σt =0..d (1+W ACC )t is calculated in equa-
tion 3.7. In the second step, the annual payment to eligible power plants is organized
by equation 3.9. This shifts the price related uncertainty and risk from the electricity
producer to the government.
 46 3. M ODELLING : EML AB G ENERATION AND R ENEWABLE S UPPORT

X post g t Ig
Σt =0..d = Σtb=0
(1 + W ACC )t (1 + W ACC )t
(3.7)
t b +t D COut f l ow Op,g
+ Σt =t
b +1 (1 + W ACC )t

where,

3 COut f l ow g ,t +n = Σ(vc g ,t +n ) ∗ r s,g ,t +n ∗ a g ,s )

(3.8)
− Σ f c g ,t +n

pa yment g ,t = Σg Σs (X post g − p s∗ ) × a g ,s wher e, t ∈ {t b ...t D } (3.9)

The risk faced by the energy producer is lower in the ex-post scenario, since there is
no price risk in the revenue component of the NPV. This is represented in the following
manner. The rate of equity component, which indicates price risk, r E p in Equation (3.4),
is set to 0%.

3.5. D ATA
This section consists of a description of the data used to model the electricity system.
It consists of assumptions regarding load duration curves and demand growth trend as-
sumptions. Characteristics of the technologies, their costs, availability of intermittent
technologies are also described.

3.5.1. L OAD DURATION CURVES

As mentioned earlier, hourly load is aggregated into 20 segments of individual load, pe-
riod pairs, to form a load duration curve. The intial load duration curve is based on
2014 ENTSO-E hourly data for consumption in the Netherlands and Germany (ENTSO-
E, 2015). The aggregated curves are shown in Figures 3.8 and 3.9. The load curves for
each consecutive year after the first are computed based on a ‘demand growth factor ’,
as explained below.

3.5.2. D EMAND GROWTH TRENDS

The demand growth trends are stochastic trends and determined in the following way.

D t +1 = αt × D t (3.10)

where, D is the demand at year t and α is the “growth rate” calculated from a trian-
gular distribution. A triangular distribution is a is a continuous probability distribution
with lower limit (min) a, upper limit (max) b and mode c, where a < b and a ≤ c ≤ b. The
values for a, b, and c are presented in Table A.1, in the appendix. The values are based
on the EU reference scenario of 2014 (European Commission and Direction générale de
la mobilité et des transports, 2014), which predicts a 1% growth rate. The trends are de-
picted pictorially in Figure A.1.
3.5. D ATA 47

15000

Load [in MW]

10000 3

5000

0 2500 5000 7500

Time [in hours]

Figure 3.8: The initial load duration curve for the Netherlands split into 20 segments

80000

60000
Load [in MW]

40000

20000

0 2500 5000 7500

Time [in hours]

Figure 3.9: The initial load duration curve for Germany split into 20 segments
 48 3. M ODELLING : EML AB G ENERATION AND R ENEWABLE S UPPORT

3.5.3. T ECHNOLOGY CHARACTERISTICS

The problem being tackled in this thesis relates to the long-term impact on investment
in generation capacity and its interaction with the electricity market under different poli-
cies. Therefore the generation technologies have been described or modelled, in terms
of such properties as are relevant to the longer-term: their nominal capacity, construc-
tion time, technical lifetime, depreciation time, fuels used, their availability if intermit-
tent, and efficiency. Characteristics such as ramp rates which are relevant to the short
term operation of the power plant are not considered. The technologies implemented
and their characteristics are mentioned in Table 3.1.
3
Table 3.1: Assumptions about technical characteristics of power generating technologies, where [a] is number
of years

a]

a]
a]
e[

e[
e[
tim

tim
tim
a]
]
MW

on

ion
e[

ife
y

im
cti
log

all
y[

t
cia
tru

it t

nic
cit
no

pre
rm

els
pa

ns
ch

ch

De
Co
Ca

Fu
Pe
Te

Te
Wind Offshore 600 1 0 20 20 -
PV 500 1 0 20 20 -
Wind Onshore 600 1 0 20 20 -
Coal Pulverised SC 758 4 1 50 20 Coal
Lignite 1000 5 1 50 20 Lignite
Nuclear 1000 5 2 40 25 Uranium
CCGT 776 2 1 40 15 Natural Gas

Fixed costs of technologies Costs of investment, costs of fixed operation and mainte-
nance, and efficiencies are specified for each technology for the duration of the simula-
tion. The learning curves for solar PV were created based on data from ISI (2015). The
learning curves for wind, and wind offshore were created based on data from IRENA re-
ports on costs of renewable technologies (IRENA, 2015). The yearly values for the learn-
ing curves, and the R code that was used to create them have been uploaded on a publicly
accessible data platform, cited in Iychettira (2015). The assumptions made regarding
fixed costs of conventional technologies are presented in Table A.3 in the appendix.

Availability of intermittent technologies Data for the hourly availability of intermit-

tent technologies for onshore wind, solar PV, and wind offshore plants were taken from
the supplementary data used in the paper by Lion Hirth on the assessment of market
value of renewable electricity Hirth (2013). The files containing the data are publicly
accessible in CSV format at an online data-hub, here Lion (2013).

3.6. M ODEL E VALUATION

The step, "model evaluation", also commonly referred to "verification and validation", is
an integral part of building a computer simulation; they are meant to evaluate whether
the simulation model is an adequate representation of the target system, relative to the
goals of the modelling study. Wendy Parker, a scholar of Epistemology of Computer
3.6. M ODEL E VALUATION 49

Simulation, characterises the process of model evaluation in the following way (Parker,
2008).
"...the question of whether the computer simulation model is an adequate represen-
tation of the target system, relative to the goals of the modelling study, is of utmost im-
portance. The activity of model evaluation (also sometimes known as ‘validation’) aims
to collect evidence regarding precisely this question. Depending on the goals of the mod-
elling study, the process of model evaluation might treat the simulation model as a black
box and focus only its output, or it might involve opening the black box to investigate
the accuracy of particular modelling assumptions and/or the adequacy of the process by
which solutions to the continuous model equations are estimated." 3
In her writing, she is also at pains to emphasize that model evaluation should be
perceived as "an investigation of the model’s adequacy for purpose, not an investigation
of its truth or falsity, whatever that might mean. A model that is constructed with the use
of a variety of false assumptions about a target system might nevertheless be an adequate
representation of that target system, relative to the goals of the modelling study. "
Drawing from the above, this section is split into two parts: Verification, where the
mathematical task of confirming whether the simulation code is implemented correctly
with respect to the conceptual model is discussed, and Validation, where the fitness of
purpose is discussed.

3.6.1. V ERIFICATION
In order to assess whether the simulation code of the agent-based model represents the
conceptual model, the following steps can be employed, based on the recommendations
in Dam et al. (2012).

1. Recording and tracking agent-behaviour: This procedure involves recording the

relevant input, the states, and the outputs of the individual agents and each of
their internal processes. This is usually performed by using loggers throughout
the code, or by using a debugger, available in the integrated development envi-
ronment.

2. Unit testing: In this procedure, small parts of the code are tested by predefining
inputs and providing expected outputs, after which they can be automated to eval-
uate individual units through various inputs. Two important parts of unit testing
are: 1) theoretical prediction and sanity checks 2) extreme value testing. The dis-
advantage of this method is that for a large code composed of a multitude of mod-
ules, they only provide insight on individual parts of the code, and not whether the
code works correctly together as a whole.

3. Interaction testing in a minimal model: In this step, an evaluation is performed to

test whether multiple agents interact with each other in a manner that is expected,
using a minimal model. This step is also used to test whether multiple modules
work as expected.

4. Multi-agent testing: In this step, the full model, with the complete number of
agents is tested for coherent behavioural patterns. In addition to sanity checks
described above, variability testing and timeline sanity check is also performed.
 50 3. M ODELLING : EML AB G ENERATION AND R ENEWABLE S UPPORT

The above tests were applied the implemented model, as described below. In tables
3.2 and 3.3, each test, the module or role or agent behaviour to which they were em-
ployed, and outcomes are mentioned.

Table 3.2: List of checks performed to verify the correct implementation of quantity warranty schemes

Unit Function Test Outcome

Calculate Re- An exogenous target, • Check whether the target for renew- Verified
newable Tender specified as a percentage ables is computed as expected for each
Target of generation to come technology in technology specific pol-
3 from renewable sources, icy designs.
is converted into its cor- • Check whether target is computed
responding value in MW, as expected for all renewable technolo-
for every year, depending gies together, in technology neutral
on whether the policy scenarios.
is technology specific or
technology neutral.
Submit Tender • Each energy pro- • Check whether the bids correspond Verified
Bid Role ducer bids for as many to the costs specified exogenously.
plants as the target al- • Check whether expected revenue
lows, using only his costs in the bid calculation corresponds to
for the technology if the agent’s forecasted demand, fuel prices,
support scheme employs and merit order 2 years into the future.
ex-post pricing, and his • Check whether, under technology
expected revenue if the neutrality, the bids correspond to the
support scheme uses ex- cheapest technology first.
ante pricing. • Check whether the number of bids
are limted also by the technology po-
tential, in case targets are higher than
potential.
Clear Renewable Given bids from all pro- • Check whether the accepted quan- Verified
Tender Role ducers, the regulator tity matches the target, as determined
matches supply and de- in step 1.
mand, and determines • Check whether the clearing price cor-
the clearing price. This responds to last accepted bid.
is organized with tech-
nology specificity or
neutrality, as determined
by the support scheme
design.
Create Power To declare and initialize • Check whether a new plant has Verified
Plants of Ac- new power plant ob- been created, corresponding to the ac-
cepted Bids jects, their correspond- cepted bid.
Role ing financial entities,
including loans, down-
payments.
Organize Renew- To process cash flows for • For ex-post policies, check whether Verified.
able Tender Pay- each tender contract with annual subsidy = cost - revenue from
ments Role. a power plant in each electricity market.
year. • For ex-ante policies, check whether
subsidy = tender clearing price * power
generated.
Full tender algo- Are the investments upto the target Verified
rithm set? Are the cash flows as expected?
3.6. M ODEL E VALUATION 51

Table 3.3: List of checks performed to verify the correct implementation of price warranty schemes

Unit Function Test Outcome

Compute pre- The regulator agent com- • Check whether the regulator agent Verified.
mium role putes the premium for computes the same costs in this role as
each technology, based the power plant owner does during an
on the costs and expected investment decision.
revenues to the power
plant owner, assuming he
has the same information

Feed in premium
as the power plant owner.
For a new elegible plant, • For ex-post policies, check whether Verified. 3
role. a contract is made. For annual subsidy = cost - revenue from
each eligible power plant, electricity market.
the support price is calcu- • For ex-ante policies, check whether
lated and payment made subsidy = tender clearing price * power
generated.
Full price war- • The investments should be upto the Verified.
ranty algorithm potential of the technology.

3.6.2. VALIDATION
Wendy Parker provides the following strategies to determine the internal and external
validity of a code10 (Parker, 2008). Although it is too early in the dissertation to discuss
the validity of the results, a framework for testing validity is presented, based on the
insights drawn from strategies in physical experiment validation to code evaluation. The
strategies and their application to the code presented is discussed below.

• Simulation output fits closely with observational data: In the context of the model
presented, this strategy would be similar to historical validation. When input data
such as demand, initial supply, and fuel prices match historical values, the outputs
such as electricity prices should also match their corresponding historical values.
While the analysis presented in Chapter 4 is indicative, historical data has been
used for the analysis presented in Chapter 5, and consequently a discussion is pre-
sented on the similarities of the results with observational data in that Chapter.

• Simulation results change as expected after intervention on model parameters: This

strategy is largely evident in the observations and interpretations of the results.
For instance, when revenue from electricity prices increase, the subsidy costs are
expected to decrease. When the price warranty is sufficient, there is investment up
to the realistic technical potential of the technology.

• Simulation model is constructed using well confirmed theoretical assumptions: For

the simulation of a socio-technical system, assumptions need to be made regard-
ing the behaviour of agents who represent actors in the system. Predictions from
neo-classical economics are well-confirmed under certain conditions, but not all.
For instance, in a spot market, actors largely behave as theory predicts: submit-
ting bids at the marginal costs. However, for investment decisions, behaviour is
10 Internal validity concerns what is true of a particular experimental system and the validity of its associated
result statements, while external validity concerns generalizability
 52 3. M ODELLING : EML AB G ENERATION AND R ENEWABLE S UPPORT

far more complicated than what neo-classical economics allows for. While deci-
sions are taken based on cost-benefit analyses, the outcome is rarely a long-term
equilibrium due to factors such as imperfect information. For such analyses, as-
sumptions that are face validated are used; assumptions that are validated by ex-
perts in the field - through peer review, presenting works at conferences, talking to
practitioners etc.

• Simulation results are reproduced in other simulations or in traditional experi-

ments: For a socio-technical system simulating 3 decades into the future, a "tra-
3 ditional experiment" is impractical to conduct. However, comparisons with other
simulations are made in the discussion sections of Chapters 4 and 5, to the extent
that the scope of other studies are comparable to the ones presented here.

3.7. C ONCLUSION
Having laid down theoretical foundations in the previous chapter, this chapter concep-
tualized and formalized the algorithms employed, and the primary narrative embodied
in the simulation. The instrument used to perform the core of the analysis was described
in detail in this chapter. The process of gathering data was also briefly described: the
sources and means of obtaining the information on generation technologies, demands,
costs, plausible assumptions for trends related to expectations of future growth.
No simulation is complete without checking whether it is verified and validated. The
final section of this chapter therefore described what verification entails for an agent-
based model of investment in the power sector. The various checks performed to ensure
the code was built as intended were described. This was followed by a discussion on
the framework to assess its suitability for its intended purpose, and how that has been
implemented in the thesis. The following chapters present the experiments that were
conducted with the simulation described, the major analysis, and insights that follow.
R EFERENCES 53

R EFERENCES
K. K. Iychettira, R. A. Hakvoort, and P. Linares, Towards a comprehensive policy for elec-
tricity from renewable energy: An approach for policy design, Energy Policy 106, 169
(2017).

L. de Vries, E. J. Chappin, and J. C. Richstein, Decarbonisation of the Power Sector, Tech.

Rep. (Delft University of Technology, Delft, 2012).

A. Chmieliauskas, E. J. L. Chappin, C. B., I. Nikolic, and G. P. J. Dijkem, New Methods for

Analysis of Systems-of-Systems and Policy: The Power of Systems Theory,Crowd Sourc- 3
ing and Data Management, in System of Systems, edited by A. V. Gheorghe (InTech,
2012) dOI: 10.5772/28796.

E. J. L. Chappin, Simulating Energy Transitions. (Next Generation Infrastructures Foun-

dation, Delft, 2011) oCLC: 840444784.

J. Richstein, Interactions between carbon and power markets in transition (Delft Univer-
sity of Technology, Netherlands, 2015).

J. C. Richstein, E. J. L. Chappin, and L. J. de Vries, Cross-border electricity market effects

due to price caps in an emission trading system: An agent-based approach, Energy Pol-
icy 71, 139 (2014).

DiaCore, The impact of risks in renewable energy investments and the role of smart poli-
cies, Work Package 3 Leader Organisation: ECOFYS IEE/12/833/SI2.645735 (Ecofys,
2016).

M. of Economic Affairs Agriculture and Innovation, National Renewable Energy Action

Plan of the Netherlands: Directive 2009/28/EC, Submitted to European Commission
(2010).

A.-R. Laali and M. Benard, French wind power generation programme EOLE 2005 Results
of the first call for tenders, Renewable Energy 16, 805 (1999).

ENTSO-E, Country Data Packages, (2015).

European Commission and Direction générale de la mobilité et des transports, EU en-

ergy, transport and GHG emmissions trends to 2030: reference scenario 2013 (Office
for official publications of the european communities, Luxembourg, 2014) oCLC:
889309007.

F. ISI, Current and Future Cost of Photovoltaics. Long-term Scenarios for Market Develop-
ment, System Prices and LCOE of Utility-Scale PV Systems., Study on behalf of Agora
Energiewende. 059/01-S-2015/EN (Freiburg, 2015).

IRENA, Renewable Power Generation Costs in 2014., Tech. Rep. (2015).

K. Iychettira, Fixed costs of renewable technologies. - the Datahub, (2015).

 54 R EFERENCES

L. Hirth, The market value of variable renewables: The effect of solar wind power variabil-
ity on their relative price, Energy Economics 38, 218 (2013).

H. Lion, Hourly capacity factors for intermittent generation - Datasets - the Datahub,
(2013).

W. S. Parker, Franklin, Holmes, and the Epistemology of Computer Simulation, Interna-

tional Studies in the Philosophy of Science 22, 165 (2008).

K. H. V. Dam, I. Nikolic, and Z. Lukszo, Agent-based modelling of socio-technical systems,

3 May (Springer, 2012).
 4
R ENEWABLE SUPPORT INTERACTS
WITH THE ELECTRICITY MARKET

Parts of this chapter have been published in Applied Energy 187, 228 (2017) Iychettira et al. (2017)

55
 56 4. R ENEWABLE SUPPORT INTERACTS WITH THE ELECTRICITY MARKET

4.1. I NTRODUCTION
4.1.1. M OTIVATION AND R ESEARCH O BJECTIVE
In a recent article on the transition towards a green economy, Newbery (2016) argues
for the merits of a renewable support policy comprising of a Contract for Differences
(CfD) with a standard Feed-in-Tariff (FiT) as opposed to a Premium FiT proposed by
the 2015 EU Energy Union Package (European Commission and Directorate-General for
Economic and Financial Affairs, 2015). It has been more a decade since the first Renew-
able Energy Sources (RES) directive, and the debate on how best to design support for
renewable electricity is still raging. The European Commission only specifies that there
will be no national level targets beyond 2020, and that most Renewable Energy Sources
for Electricity (RES-E) support schemes should take the form of competitive bidding. It
still remains to be seen whether these choices will lead to the triad of competition, sus-
4 tainability, and affordability being achieved in the energy sector.
Since the first RES-E Directive was released in 2001, there have been numerous works
that have evaluated renewable support schemes from theoretical and empirical stand-
points; refer for instance Batlle et al. (2012a); Couture and Gagnon (2010); Schmalensee
(2012); Neuhoff et al. (2013); (IEA) (2008). Such literature so far on renewable support
schemes has mainly focussed on comparing different policies1 or support schemes 2
that have been implemented in various member states of the European Union (EU). The
key here, however, is not a choice between policy A or B, but between how either pol-
icy instrument should be designed. This allows the policy maker such as the European
Commission to decide what design features are essential in an RES-E scheme, rather
than propose an entire scheme itself. This idea has been upheld by several authors such
as del Rio and Linares (2014); del Rio and Mir-Artigues (2014); Haas et al. (2011).
It was proposed in earlier chapters that any RES-E policy can be broken down into a
closed set of components that are common to all renewable electricity support schemes.
These components are referred to as ‘design elements’; the design elements now form
the smallest level of analysis. The objective of this chapter is to assess the impact of de-
sign elements of Renewable Energy Source – Electricity (RES-E) support schemes on a
single (isolated, uncongested) region modelled approximately similar to the power sec-
tor in the Netherlands, using a long-term agent-based model of the electricity market,
with endogenous investment. We introduce the design elements in Section 2.2.2, and
demonstrate that it is possible to model elements individually in Chapter 3. The policies
are then modelled as combinations of design elements. The design elements analysed
are price warranty versus quantity warranty, electricity market revenue accounted for ex-
post or ex-ante, and technology specificity versus technology neutrality. The performance
indicators in this study are effectiveness of policy in terms of cost and target achieve-
ment, and social welfare and distributional implications on producer, consumer, and
the government.
The following subsection comprises of a review of literature in the field, and outlines
how this work contributes to literature. This is followed by Section 4.2, which includes
a detailed description of the methodology used: the design elements considered, the
1 Policy is a general term used to describe a formal decision or a plan of action adopted by an actor, such as the
government, to achieve a particular goal.
2 The word policy is used interchangeably with the word scheme in this work.
4.1. I NTRODUCTION 57

model, the hypotheses and experiment design. The subsequent section includes the
results and their discussion, followed by the conclusion.

4.1.2. L ITERATURE R EVIEW

The current work relates to two strands of literature, one where RES-E schemes have
been analysed, and the other where they have been modelled.
RES-E schemes have been compared analysed at great depth since the first RES-E di-
rective. Recent literature in the field still indicates that policy comparisons dominate the
field (Newbery, 2016; Fais et al., 2014; Dressler, 2016; Winkler et al., 2016; Reuter et al.,
2012). Nevertheless, perceiving RES-E support schemes in terms of design elements has
been done qualitatively before by some authors. For instance, Held et al. (2014) and
the beyond2020 project by Batlle et al. (2012b) present a list of design elements for RES-
E support schemes. del Rio and Linares (2014) provide an in-depth review of auction 4
schemes for renewable electricity around the world; they identify and assess design ele-
ments of such auctions and propose a coherent integration of several design elements to
improve auction designs. The design elements described in the above papers however
are not common across all policies, thus still making them policy-specific; the disadvan-
tage being that it is not possible to objectively analyse the impacts of specific features of
a policy on the system. Also importantly, all the aforementioned works only qualitatively
discuss design elements, but provide no quantitative analysis regarding their long-term
dynamic effects and welfare distributional implications.
There have been several quantitative modelling efforts to evaluate the effectiveness
of RES-E support schemes. Capros et al. (2014) provide a detailed description of seven
large scale EU energy economy models that have been used to model decarbonisation
pathways. Works which use models that have simulated and quantitatively compared
different RES-E support policies in some detail include the Green-X model (Huber et al.,
2004), the REBUS (Renewable Energy Burden Sharing) model (Voogt et al., 2001), the
PERSEUS-RES-E (Programme-package for Emission Reduction Strategies in Energy Use
and Supply-Certificate Trading) model by Most and Fichtner (2010), and an extended
version of the TIMES-D (The Integrated MARKAL-EFOM System) model by Fais et al.
(2014), henceforth referred to as the TIMES-D-Extension Model.
In terms of the research objective and experiment design, the work using TIMES-D-
extension model is the most comparable to the current one. Like the others, it compares
support schemes themselves - the Feed-in-Tariff scheme to a Tradable Green Certifi-
cate mechanism. However, like this work, it comprises of a long-term evaluation of the
support schemes, under design criteria which include technology specificity and tech-
nology neutrality. Hence, further comparisons to literature will primarily be limited to
the TIMES-D-extension model. The TIMES-D-extension model is a partial equilibrium
energy system model, which employs an objective function representing the total dis-
counted system costs across the years 2000-2050,
These models can be classified into one of the three trends in electricity market mod-
elling: optimization models, equilibrium models, and simulation models Ventosa et al.
(2005). Optimization models include both deterministic, and stochastic programming.
Typically, with respect to investment decisions, the aforementioned models assume per-
fect foresight, and perfect competition. Some models use stochastic parameters and/or
 58 4. R ENEWABLE SUPPORT INTERACTS WITH THE ELECTRICITY MARKET

scenario analysis to account for certain types of uncertainties. However, even these sce-
narios or probability distributions need to be estimated by the analyst.
Such methods imply that investment decisions are made under the premise of min-
imisation of system expenditure across time. As Most and Fichtner (2010) and Olsina
et al. (2006) point out however, such assumptions imply that capacity or production de-
cisions can be taken instantaneously, under conditions of free entry and exit. These as-
sumptions can hardly be expected to hold in the real-world, especially in sectors where
investment decisions, which happen with knowledge of past trends, and imperfect fore-
sight, are a major determinant of welfare outcomes.

4.1.3. F ROM S CENARIO A NALYSIS OF P OLICIES TO D ESIGN E LEMENTS

In a scenario analysis, the uncertainty about parameters or components of the system
4 is modelled by a small number of versions of sub-problems derived from an underlying
optimization problem. These correspond to different “scenarios,” suggesting some kind
of limited representation of information on the uncertain elements or how such infor-
mation may evolve.
The critical question then is how to determine which components of the system
comprise each scenario, and why a certain set of scenarios are sufficient. So far, in mod-
elling studies related to renewable electricity support schemes such as those aforemen-
tioned, different scenarios are formed by established current policies in their entirety.
In other papers, variations of designs within one single established policy are analysed.
However, it is critical to note that two seemingly different policies can be designed such
that they have an equivalent effect on the market. For instance, a Tradable Green Certifi-
cate (TGC) scheme with long term contracts resembles a tender. A Feed in Premium (FiP)
scheme with long term contracts resembles a Feed-in-Tariff (FiT) (Batlle et al., 2012b).
The underlying idea therefore is that it is not the policy but the design element which is
the vital component of analysis. In effect, the decision variables are no longer the poli-
cies, but the design elements that they are composed of.
The design element approach allows us to systematically explore the entire RES-E
policy design space, even create new policies that have not been implemented before.
More importantly, it allows us assess the impact of a specific feature of a policy on the
system. This feature could be technology specificity, price vs. quantity warranty, or type
of price setting. With such information, it is possible to advice the EC on what design
features are essential in an RES-E scheme, rather than proposing an entire scheme itself.

4.1.4. C HOICE OF M ODELLING A PPROACH

The methodology and work presented herewith is fundamentally different from the afore-
mentioned works in two main aspects. One is a shift from a ‘policy’ view to a ‘design ele-
ment’ based view of RES-E support assessment. The second fundamental difference lies
in the methodological approach employed in this work, Agent-Based Modelling (ABM).
ABM is recognized as a methodology that provides a framework to model agents with
bounded rationality, their interactions with other agents, and the environment around
them, as Epstein (2007), Manson (2006), and North et al. (2013) have explained.
The ‘base model’employed, EMLab, consists of generation companies as agents who
individually make investment decisions. The investment decisions of the past affect
4.1. I NTRODUCTION 59

Figure 4.1: High level diagram of behaviour of agents and their interaction with environment in EMLab.
Adapted from Richstein (2015)

those of the future, and agents make decisions under imperfect foresight. Agents create
their own forecasts using regression techniques of past values of demand and fuel price
trends, much like in the real-world, to arrive at endogenous investment patterns. Such
real world representations help analyse how different designs of RES-E support affect in-
vestment incentives, and consequently affect the energy transition. The base model, on
which this work has been built, is described in detail in section 3.2, and with flowchart in
3.3, and is represented in Figure 4.1. The design elements, and consequently the RES-E
policies, that have been modelled as part of the current work, are described in Section
3.4.
This approach is markedly different from the aforementioned modelling methods
because of the following reasons. Firstly, since each agent makes individual investment
decisions based only with current knowledge of the system, we implement bounded ra-
tionality; this often leads to sub-optimal choices when assessed ex-post, much like re-
ality. Secondly, in equilibrium models, typically the policies are modelled close to how
they work in theory. It is implicitly assumed that the policy in place would achieve its
target, as modelled. However, this method does not help identify reasons that a policy
would not work as intended; interpretation is left to the analyst. Including uncertainties
and bounded rationalities in the model, helps pinpoint which micro decisions lead to
which macro outcomes in the model. Thirdly, unlike optimization models, the focus of
our model is not a final minimum cost state, but to analyse dynamics in the path of an
energy transition, while including specific uncertainties.
Such modelling takes us a step closer to representing the real world. The base model
has so far been applied to study long term dynamics of the electricity market in rela-
tion to security of supply and carbon trading, in various publications (Richstein, 2015;
Richstein et al., 2014; Bhagwat et al., 2014).
 60 4. R ENEWABLE SUPPORT INTERACTS WITH THE ELECTRICITY MARKET

4.2. E XPERIMENT D ESIGN

The Base Case Set The fundamental premise of this work is that design elements are
the building blocks which allow the policy analyst to create all possible types of RES-
E support schemes. Thus all combinations of three design elements introduced above,
where each design element can hold two values, lead to 23 RES-E policy scenarios. This
is shown in Table 4.1.
If one were to draw parallels between some of the scenarios and actually imple-
mented schemes, P_Ante would be akin to the German Feed-in-Tariff scheme, P_Post
to the German Feed-in-Premium, and Q_PostTS is comparable to the UK’s contract-for-
differences scheme, where ex-post contracts are allocated on a technology-specific ba-
sis, via auctions, and the SDE+ in the Netherlands is similar to Q_Post, where technology
neutral auctions are held for ex-post type of contracts. However, not all RES-E policy
4 scenarios exist currently or have been implemented in reality, so names for such policies
do not exist. Also, policies with the same names are implemented differently in different
countries. For this reason and to keep intact the relationship between each policy sce-
nario, and the design elements that it is composed of, we propose a naming convention
as provided in Table 4.1.

Table 4.1: Base Case Experiment Set - Naming Convention

RES-E Policy Sce- Design Element 1: Design Element 2: Design Element 3:

nario Name Warranty type Price Setting Tech Neutral vs Spe-
cific
P_Ante Price Warranty Ex_Ante Neutrality
P_Post Price Warranty Ex_Post Neutrality
P_AnteTS Price Warranty Ex_Ante Specificity
P_PostTS Price Warranty Ex_Post Specificity
Q_Ante Quantity Warranty Ex_Ante Neutrality
Q_Post Quantity Warranty Ex_Post Neutrality
Q_AnteTS Quantity Warranty Ex_Ante Specificity
Q_PostTS Quantity Warranty Ex_Post Specificity

Sensitivity Analysis The impact of the design element ex-ante vs ex-post inter-alia de-
pends on how well the expectations of producers’ electricity price match actual prices.
The development of electricity prices in a system dominated by CCGT technology is in
turn largely dependent on gas prices. In order to understand this relationship better, a
sensitivity analysis is executed for increasing and decreasing gas prices. The gas price
for the base scenario is set constant at the current3 approximate price of 4 Eur/GJ. The
Gas High scenario has an annual growth rate of 2% while the Gas Low scenario has one
of -2%.

Experiment Setup: Randomness and Repetitions Agent-based modelling in general,

and this model in particular, require multiple runs to arrive at statistically significant
3 June, 2016
4.2. E XPERIMENT D ESIGN 61

conclusions. This is because two runs of the same scenario are differentiated by ran-
domness in the following parameters such as a) randomised agent iteration in order to
prevent first-mover artefacts, b) stochastic demand growth trends, randomness in ini-
tial age of power plants, as the age is drawn from a uniform distribution between 0 and
the technical lifetime of a power plant, and finally c) randomness in initial power plant
ownership. After performing a simple descriptive statistical test for the variance of re-
sults, it was deemed that 40 repetitions were sufficient to obtain statistically significant
outcomes.

4.2.1. I NPUT D ATA : C ASE OF THE N ETHERLANDS

A single (isolated, uncongested) electricity market is considered, with four energy pro-
ducer companies, whose initial portfolio is based approximately on the existing gener-
ation mix in the Netherlands. However, to ensure focus on assessing RES-E design el-
ements, the model is simplified such that all conventional capacity in the Netherlands
4
is represented by the Combined Cycle Gas Turbine (CCGT) technology. Given recent
Dutch laws regarding the phasing out of coal, see Green (2015), and equivocal opin-
ions on nuclear technology, refer Association (2016), it is reasonable to assume that a
significant part of the conventional generation mix will be dominated by gas technolo-
gies. Along with CCGT, three renewable technologies are considered, and assumptions
regarding their characteristics are described in Table 3.1. The intermittent nature of re-
newable generation sources is represented by hourly availability factors, which are then
aggregated to segment-based4 availability factors. The data for hourly availability for the
renewable technologies is obtained from Pfenninger and Staffell (2016). The model runs
for 40 ticks, with each tick representing a year starting from 2014.
The targets and realistic potentials for renewable technologies have been set based
on data from Lako (2010) and Ragwitz et al. (2003), and extrapolated, as described in
A.1.3. Fuel prices of natural gas and electricity demand, are modelled as stochastic
trends, using a triangular distribution to determine the year-on-year growth rate. The
assumptions for modal growth rate, and its upper and lower bounds are summarized in
Table A.1. The initial load duration function is based on 2014 ENTSO-E data for Nether-
lands. A value of lost load of 2000 Eur/MWh has been used for this work (Anderson and
Taylor, 1986; Linares and Rey, 2013; Nooij et al., 2007).

4.2.2. C RITICAL R EVIEW OF M ODELLING A SSUMPTIONS

One assumption that impacts the analysis is that there are no interconnections or stor-
age in the system. This implies that as the share of renewable production increases, a
greater share of the energy generated will not be consumed, due to spillage5 . This leads
to the cost effectiveness of a subsidy reducing over time, as the share of renewable gen-
eration in the system increases, which would not occur as sharply in the presence of
storage or interconnections. Another important assumption is that the energy produc-
ers construct a market clearing for one time point in the future and extrapolate those
revenues for the lifetime of the plant. This implies that actual costs and benefits might
4 In order to represent variability of load across the year, the load duration curve is divided into segments; each
segment being a (load, time) pair value, and each segment is cleared separately.
5 Greater amounts of renewable energy will be generated when there is insufficient demand for it
 62 4. R ENEWABLE SUPPORT INTERACTS WITH THE ELECTRICITY MARKET

be very different from those expected. The next major assumption is that the regulator
agent has full knowledge of costs of technologies, and uses the same rates of return as
the energy producers. While this assumption may not hold in reality, it helps to isolate
and study the impacts of design elements better.

4.3. R ESULTS
This section comprises of two subsections: the first consists of the results as per the
performance indicators mentioned in the introduction to this chapter. The performance
indicators are effectiveness of policy, and social welfare and distributional implications.
The second consists of a discussion and interpretation in subsection 4.3.2, primarily in
4 terms of impacts of design elements. Condensing large sets of granular results to a few
key indicators is a challenging activity, and must be done carefully.

4.3.1. R ESULTS AND D ISCUSSION

E FFECTIVENESS OF P OLICY

Effectiveness of policy is measured using two indicators: cost effectiveness and target off-
set. Cost effectiveness is defined as total subsidy cost per MWh of renewable electricity
generated6 , summed across all 40 years, in Eur/MWh. It is then averaged across all 40
repetitions of the scenario. Target offset measures the difference between the actual re-
newable energy generation and the exogenously specified target. It is expressed as per-
centage, and then averaged across all years and 40 repetitions per scenario.

% % (r Gen t ,r ep −t ar g et t ,r ep )×100
r ep t t ar g et t ,r ep
t ar g etO f f set = (4.1)
n r ep × n t i ck

Figure 4.2 indicates these values for each scenario. The evolution of capacity in each
of the scenarios is shown in Figure 4.3.

6 All renewable energy technologies are considered

4.3. R ESULTS 63

Figure 4.2: Policy effectiveness measured in subsidy costs and target achievement

At the outset, it is to be noted that the target has been grossly under-achieved in
scenarios P_Ante and P_AnteTS. This is a consequence of the regulator agent’s short-
sightedness with respect to expectations of future electricity prices7 . It is for the same
reason that this is visible only in the ex-ante scenarios, as there is no need to com-
pute expected electricity price in the ex-post scenarios. The effect is exacerbated in the
technology-specific scenario, as a price warranty is calculated for each technology, while
in the technology-neutral scenario, a price warranty is only calculated for the marginal
technology. It is useful to note here that target achievement has little relation to the Av-
erage Subsidy Cost/Unit, as the latter is normalized with respect to generation in MWh.
The results indicate the following:
a) Quantity-warranty schemes are on average 4.5% more cost-effective and meet
their targets more consistently than their price-warranty counterparts. This is because
price-warranty schemes induce investment in technologies up to the point at which the
realistic potential of a technology is reached, and not the administrative target which
is lesser than the potential. Greater the amount of renewables in the portfolio, greater
the spillage 8 and lower the generation. Therefore, unless there is an interconnector
7 This is because when the regulator agent calculates the required price warranty, her expectation of revenue
from the electricity price is calculated by taking into account all the electricity plants that are expected at
that moment. However, after this calculation if investments do incur in the same year, due to which the
expected electricity price drops, the regulator does not make a reassessment of revenue expected from the
electricity market for the same year. Therefore, the regulator’s assessment of revenue from the electricity
market becomes higher than it actually is, and the corresponding price warranty becomes lower than it needs
to be, at the time of investment.
8 Spillage can be defined as renewable capacity generating more than the demand at a certain hour
 64 4. R ENEWABLE SUPPORT INTERACTS WITH THE ELECTRICITY MARKET

Figure 4.3: Capacity growth in GW per policy scenario with time (in years) on x-axis

to a region with complementary demands during hours of spillage, or the presence of

storage, or demand response, the higher the share of renewables, the lower the cost-
effectiveness.
b) Technology specific schemes are 60.3% more cost-effective than their technology
neutral counterparts. This is due to windfall profits to non-marginal technologies in the
technology neutral scenarios.
c) Ex-post schemes are 15.8% less cost effective than their ex-ante counterparts. In
the ex-ante schemes, the expectations of revenues from electricity market are higher
than actual, over a twenty year period. The subsidy in ex-ante schemes in the model
therefore tend to be lower than necessary. This result is sensitive to the future electricity
prices; impacts of high or low gas price scenarios can be observed in Figures B.1a and
B.1b.

S OCIAL W ELFARE AND D ISTRIBUTIONAL I MPLICATIONS

The distributional implications are presented in Figure 4.4, indicating the change in sur-
pluses for the consumer, producer, government, and total social surplus, for each sce-
4.3. R ESULTS 65

nario. The change is computed by comparing each scenario with a base case, where no
policy is implemented. Change in consumer expenditure9 , change in producer costs,
and change in government expenditure are used as proxies for calculating the changes
in consumer, producer, and government surpluses.

Figure 4.4: Change in Surplus for producer, consumer, government, and society (total) in 40 years

Overall, the results indicate that the greatest increase in social surplus occurs in the
scenario Q_AnteTS, where a technology-specific, quantity-warranty, ex-ante scheme is
implemented. The results will now be detailed per group. In all scenarios, consumer
surplus increases; this is primarily caused by a fall in the average electricity prices due
to the merit order effect. Government surplus is only affected by the amount of subsidy
spent. The main design element affecting government welfare is therefore technology
specificity. Surplus is more negative in technology neutral scenarios, compared to their
corresponding technology-specific counterparts due to the windfall profits mentioned
earlier.

9 The Consumer agent in the model only spends on electricity costs, the subsidy is assumed to be borne entirely
by the government for the sake of the model. In reality the cost burden is either borne by only the consumers
of electricity or all tax payers.
 66 4. R ENEWABLE SUPPORT INTERACTS WITH THE ELECTRICITY MARKET

Figure 4.5: Producer Suplus per Technology and Policy Scenario

Producer surplus is affected by costs (fixed and variable) and revenues (electricity
spot market revenue and RES-E subsidies) for various technologies. Figure 4.5 shows the
break up of producer surplus per technology and per policy scenario, for all 40 years.
In technology-neutral scenarios, as one would expect, producer surplus is high for non-
marginal renewable technologies. Furthermore, for a certain capacity of RES-E capacity,
the ex-ante scenarios show lower surpluses than their ex-post counterparts. This is again
due to the overestimating of revenue from the electricity market by either the producer
or the regulator. CCGT however shows a negative producer surplus in all scenarios10 .

The cost-benefit impacts of each policy scenario on a single technology, such as for
instance Wind Offshore, is illustrated in figure 4.6.

10 This is because fixed O&M and variable costs of CCGT are consistently higher than revenues from the elec-
tricity market. This is exacerbated by the fact that decommissioning of power plants is age based (40 years)
in the model, and not economic. In addition, reducing average electricity prices due to the merit order effect
also reduce their revenue.
4.3. R ESULTS 67

Figure 4.6: Cost, Revenue, Subsidy, and Profit for WindOffshore per MWh over 40 years

4.3.2. D ISCUSSION AND I NTERPRETATION

In this subsection, the results from the previous section are positioned in theory, and
discussed in terms of their relevance to the real-world.

Quantity-warranty vs price-warranty Quantity warranty schemes are more cost-effective

than price-warranty schemes, because price-warranty schemes induce investment in
technologies up to the point at which the total potential of a technology is reached. As
explained, this result in the model is a direct consequence of the lack of storage, demand
response, or interconnections. However, this indicates that control over quantity is ten-
uous at best under price warranty schemes, unless there are additional quantity-based
measures in place. Given this, at higher levels of penetration of RES-E, under pure price
warranty schemes, storage and/or demand response options hold utmost importance.

Technology-specificity vs technology-neutrality Theoretically as pointed out by Fais

et al. (2014), two effects are possible: the first is that expensive technologies are incen-
tivised before their time in technology-specific scenarios, therefore making technology
specificity more expensive, and the second is that cheap technologies do not get wind-
fall profits in technology-specific scenarios, therefore making those scenarios more cost-
effective. In the case of the Netherlands, it seems as if the second effect is much stronger
than the first, making the technology-neutral option more expensive. This corroborates
with the results of Fais et al. (2014), where technology neutral options incur almost twice
as much the subsidy costs as technology specific options. This effect would however not
be evident if the targets were much lower, making the marginal technology the cheapest
 68 4. R ENEWABLE SUPPORT INTERACTS WITH THE ELECTRICITY MARKET

one11 . Another factor which could impact this result is if technology cost reductions are
different than assumed.

Ex-ante vs ex-post Two effects could contribute to the impact of this design element:
the first is that there is a component of higher risk to the producer in the ex-ante scenar-
ios, therefore increasing their cost of capital, and consequently their subsidy costs. The
second effect is that higher (lower) expectations of future electricity price than reality
lead to lower (higher) subsidy costs in ex-ante (ex-post) scenarios. The results indicate
that the second effect overtakes the first. The isolated impact of the second effect can
be seen in Figure B.2a. In this scenario set, the same risk aversion of 11% is assumed in
both ex-ante and ex-post scenarios (r E p is reduced to zero in ex ante scenarios), under
constant gas prices. The ex-ante scenarios show an average of 4% decrease in subsidy
4 costs in same risk set compared to the base case set. This effectively quantifies the im-
pact of extra risk in ex-ante scenarios in the base case set. Ex-post scenarios in the same
risk scenario set are however 18% more expensive than ex-ante scenarios to the govern-
ment due to the merit order effect. A comparison between base case scenario set and
the same-risk scenario set is shown in Table B.2.
This design element is highly sensitive to expectations of future electricity prices,
which in turn depend greatly upon the merit-order effect of RES-E, and long term gas
price development. Even so, the absolute impact of this design element on policy cost
effectiveness or social welfare is at most half as significant as technology-specificity vs
neutrality. Therefore, while highly uncertain, it does not impact the socio-technical sys-
tem as much as technology-neutrality does.

4.3.3. A PPLICABILITY OF THE D ESIGN E LEMENT A PPROACH

By quantitatively demonstrating that mere design elements, irrespective of the RES-E
policy they belong to, have significant impacts on the energy system and on welfare dis-
tribution, the design element approach questions the current approach to policy making
and policy analysis in the realm of RES-E support in Europe. It takes the debate beyond a
choice between say, an auction or a feed-in-tariff, to ask how either should be designed
in order to achieve long term objectives of the system. While the concept of whether
renewable policies matter at all has been gaining traction of late in academic literature
(Winkler et al., 2016), it remains distant from ongoing policy discussions, as is elucidated
below.
The 2014 State Aid Guidelines proposed that competitive bidding, or auctions, should
be the main form of support Commission (2014) for utility scale renewable plants. This is
proposed in the place of the more popular price-based mechanisms in Europe. Compet-
itive bidding is modelled as ’quantity warranty’ in this work. This research interestingly
demonstrates that more than the feature of competitive bidding or quantity warranty,
the design element technology specificity, would incur far greater implications in terms
of welfare distribution in the Netherlands, over a period of 40 years.
Related to this, the fragmentation of the European internal electricity market due
to country-specific renewable support schemes, and security of supply policies is caus-
11 See Figure B.2b to observe results for a scenario set where the RES-E generation target remains constant at
10% of total consumption throughout the time-period
4.4. C ONCLUSIONS 69

ing increasing concern (Glachant and Ruester, 2013). Among the primary concerns of
the European Commission now, is to be able to promote renewable electricity without
causing unintended cross border impacts (Commission, 2015). A part of their strategy
to address this seems to be to promote competitive bidding in member states. However,
it is possible that even competitive bidding, when designed differently in neighbour-
ing states (for instance in terms of technology-specificity), could result in unintended
cross border effects. The design element method has the potential to provide insight
into which aspects of the policies need to be harmonised (or not); and if yes, to what
degree. This method allows the analyst to examine, element-by-element, which of them
lead to cross-border interactions between two neighbouring countries in the same elec-
tricity market.

4.4. C ONCLUSIONS 4
Most ongoing policy discussions relating to RES-E support schemes, both within and
outside of academia, compare existing policies. However, two seemingly different poli-
cies can be designed in a way that they have an equivalent effect on the market: for
instance, a tradable-green-certificate market with a long term contract is similar to a
tender. Conversely, two similar policies could have very different impacts on the sys-
tem, if designed slightly differently; for instance competitive bidding organized specific
to a technology would yield very different results from one that is technology neutral.
Therefore the core idea is that, it is the design features that form the vital component
of analysis, and not the policies in their entirety. We employ core design elements and
combine them to systematically arrive at a set of possible RES-E policy scenarios, consid-
ered complete with respect to the design elements, thus exploring the complete policy
design space. The design elements modelled are quantity warranty vs. price warranty,
technology specificty vs. neutrality, and ex-ante vs. ex-post price setting. We employ this
design element view in combination with agent based modelling to quantitatively assess
impacts of individual design elements on the socio-technical system.
The results demonstrate that design elements, irrespective of the RES-E policy they
belong to, do have significant impacts on the energy system and on welfare distribution,
and therefore that the approach is a useful one. The agent-based modelling framework
enables modelling of bounded rationalities in investment decisions, allowing the mod-
eller to incorporate real-world uncertainties in agents’ behaviour. An important uncer-
tainty in the real world is that of long-term electricity price development. The model in-
terestingly demonstrates that accounting for future electricity prices ex-ante in the sub-
sidy calculation may reduce the overall cost of subsidy by about 15%, since the actors are
likely to overestimate the future electricity price. This is a consequence of underestimat-
ing the impact of the merit order effect on expected electricity prices over the long-term.
Other significant results are that technology specificity could reduce the cost of subsidy
by upto 60%. Results regarding the design element, quantity vs price warranty corrob-
orate established literature: quantity warranty helps achieve targets better. The design
element configuration that leads to the highest increase in social welfare is the combi-
nation of quantity-warranty, ex-ante accounting for electricity prices, and technology-
specificity.
With regard to policy implications, the State Aid Guidelines of the European Com-
 70 4. R ENEWABLE SUPPORT INTERACTS WITH THE ELECTRICITY MARKET

mission promote competitive bidding to incentivize investment, while largely support-

ing technology neutrality. At the outset, our results corroborate with the choice of com-
petitive bidding. They however indicate that the feature technology specificity has a sig-
nificant implication on welfare impacts, subject to the assumption of regulator’s knowl-
edge of real costs being the same as the energy producer. Differences in such features
of RES-E policy between member states could lead to unintended cross border effects.
The design element method has the potential to provide insight into which aspects of
the policies need to be co-ordinated at the European level.

4
R EFERENCES 71

R EFERENCES
K. K. Iychettira, R. A. Hakvoort, P. Linares, and R. d. Jeu, Towards a comprehensive policy
for electricity from renewable energy: Designing for social welfare, Applied Energy 187,
228 (2017).

D. M. Newbery, Towards a green energy economy? The EU Energy Union’s transition to

a low-carbon zero subsidy electricity system – Lessons from the UK’s Electricity Market
Reform, Applied Energy 179, 1321 (2016).

European Commission and Directorate-General for Economic and Financial Affairs, En-
ergy economic developments: investment perspectives in electricity markets. (Publica-
tions Office, Luxembourg, 2015) oCLC: 948760359.

C. Batlle, I. J. Pérez-Arriaga, and P. Zambrano-Barragán, Regulatory design for RES-E sup- 4

port mechanisms: Learning curves, market structure, and burden-sharing, Energy Pol-
icy Modeling Transport (Energy) Demand and Policies, 41, 212 (2012a).

T. Couture and Y. Gagnon, An analysis of feed-in tariff remuneration models: Implications

for renewable energy investment, Energy Policy 38, 955 (2010).

Schmalensee, Evaluating Policies to Increase Electricity Generation from Renewable En-

ergy, Review of Environmental Economics and Policy 6, 45 (2012).

K. Neuhoff, J. Barquin, J. W. Bialek, R. Boyd, C. J. Dent, F. Echavarren, T. Grau, C. von

Hirschhausen, B. F. Hobbs, F. Kunz, C. Nabe, G. Papaefthymiou, C. Weber, and
H. Weigt, Renewable electric energy integration: Quantifying the value of design of mar-
kets for international transmission capacity, Energy Economics 40, 760 (2013).

I. E. A. (IEA), Deploying Renewables: principles for effective policies, Tech. Rep. (Paris,
2008).

P. del Rio and P. Linares, Back to the future? Rethinking auctions for renewable electricity
support, Renewable and Sustainable Energy Reviews 35, 42 (2014).

P. del Rio and P. Mir-Artigues, Combinations of support instruments for renewable elec-
tricity in Europe: A review, Renewable and Sustainable Energy Reviews 40, 287 (2014).

R. Haas, C. Panzer, G. Resch, M. Ragwitz, G. Reece, and A. Held, A historical review of

promotion strategies for electricity from renewable energy sources in EU countries, Re-
newable and Sustainable Energy Reviews 15, 1003 (2011).

B. Fais, M. Blesl, U. Fahl, and A. Voß, Comparing different support schemes for renewable
electricity in the scope of an energy systems analysis, Applied Energy 131, 479 (2014).

L. Dressler, Support Schemes for Renewable Electricity in the European Union: Producer
Strategies and Competition, Energy Economics , (2016).

J. Winkler, A. Gaio, B. Pfluger, and M. Ragwitz, Impact of renewables on electricity markets

– Do support schemes matter? Energy Policy 93, 157 (2016).
 72 R EFERENCES

W. H. Reuter, J. Szolgayová, S. Fuss, and M. Obersteiner, Renewable energy investment:

Policy and market impacts, Applied Energy 97, 249 (2012), energy Solutions for a Sus-
tainable World - Proceedings of the Third International Conference on Applied Energy,
May 16-18, 2011 - Perugia, Italy.

A. Held, M. Ragwitz, M. Gephart, F. ISI, M. Gephart, E. de Visser, C. Klessmann, and

Ecofys, Design features of support schemes for renewable electricity, Tech. Rep. (2014).

C. Batlle, P. Linares, M. Klobasa, J. Winkler, F. Isi, A. Ortner, and T. U. W. Eeg, Interac-

tions between RES-E support instruments and electricity markets, Tech. Rep. October
(Beyond 2020, 2012).

P. Capros, L. Paroussos, P. Fragkos, S. Tsani, B. Boitier, F. Wagner, S. Busch, G. Resch,

4 M. Blesl, and J. Bollen, Description of models and scenarios used to assess European
decarbonisation pathways, Sustainable Energy System Changes 2, 220 (2014).

C. Huber, T. Faber, R. Haas, G. Resch, J. Green, S. Olz, and S. White, Deriving optimal pro-
motion strategied for increasing hte share of RES-E in a dynamic European electricity
market, Tech. Rep. (2004).

M. H. Voogt, M. A. Uyterlinde, K. Skytte, M. Leonardi, and M. H. Whiteley, Renewable en-

ergy burden sharing. REBUS. Requirements and expectations of utilities and consumer
organisations in the European energy sector, Tech. Rep. (Energy research Centre of the
Netherlands ECN, Petten (Netherlands), 2001).

D. Most and W. Fichtner, Renewable energy sources in European energy supply and in-
teractions with emission trading, The Role of Trust in Managing Uncertainties in the
Transition to a Sustainable Energy Economy, Special Section with Regular Papers 38,
2898 (2010).

M. Ventosa, Baıllo, A. Ramos, and M. Rivier, Electricity market modeling trends, Energy
Policy 33, 897 (2005).

F. Olsina, F. Garcés, and H.-J. Haubrich, Modeling long-term dynamics of electricity mar-
kets, Energy Policy 34, 1411 (2006).

J. M. Epstein, Generative Social Science: Studies in Agent-Based Computational Modeling

(Princeton University Press, 2007).

S. Manson, Bounded rationality in agent-based models: Experiments with evolutionary

programs, International Journal of Geographical Information Science 20, 991 (2006),
cited By 40.

M. J. North, N. T. Collier, J. Ozik, E. R. Tatara, C. M. Macal, M. Bragen, and P. Sydelko,

Complex adaptive systems modeling with Repast Simphony, Complex Adaptive Sys-
tems Modeling 1, 1 (2013).

J. Richstein, Interactions between carbon and power markets in transition (Delft Univer-
sity of Technology, Netherlands, 2015).
R EFERENCES 73

J. C. Richstein, E. J. L. Chappin, and L. J. de Vries, Cross-border electricity market effects

due to price caps in an emission trading system: An agent-based approach, Energy Pol-
icy 71, 139 (2014).

P. Bhagwat, K. K. Iychettira, and L. J. De Vries, A modelling Approach to Analyze the Im-

plementation of Capacity markets in Germany, (Dusseldorf, 2014).

B. Green, Dutch lawmakers back coal phase-out, (2015).

W.-N. Association, Nuclear Power in the Netherlands, (2016).

S. Pfenninger and I. Staffell, Long-term patterns of European PV output using 30 years of

validated hourly reanalysis and satellite data, Energy (in review) (2016).

P. Lako, Technical and economic features of renewable electricity technologies, Tech. Rep. 4
ECN-E–10-034 (ECN, 2010).

M. Ragwitz, G. Resch, C. Huber, and S. White, Dynamic Cost Resource Curves, Tech.
Rep. ENG2-CT-2002-00607 (Energy Economics Group, Vienna University of Technol-
ogy, 2003).

R. Anderson and L. Taylor, The social cost of unsupplied electricity: A critical review, En-
ergy Economics 8, 139 (1986).

P. Linares and L. Rey, The costs of electricity interruptions in Spain. Are we sending the
right signals? Energy Policy 39 (2013).

M. d. Nooij, C. Koopmans, and C. Bijvoet, The value of supply security: The costs of power
interruptions: Economic input for damage reduction and investment in networks, En-
ergy Economics 29, 277 (2007).

E. Commission, Guidelines on State aid for environmental protection and energy 2014-
2020 (2014).

J.-M. Glachant and S. Ruester, The EU internal electricity market : done forever?, Working
Paper (2013).

E. Commission, Launching the public consultation process on a new energy market de-
sign, Tech. Rep. COM(2015) 340 final (Brussels, 2015).
 5
C ONGRUENCY BETWEEN NATIONAL
SUPPORT SCHEMES AND AN
INTERNATIONAL ELECTRICITY
MARKET

75
 5. C ONGRUENCY BETWEEN NATIONAL SUPPORT SCHEMES AND AN INTERNATIONAL
76 ELECTRICITY MARKET

5.1. I NTRODUCTION
The European Union’s Renewable Energy Source (RES) Directive (2009/28/EC) intro-
duced legally binding national targets to meet 20% of final energy consumption from
renewable sources (Commission, 2009). The implementation of the directive seems to
be largely on track to realizing the targets, and electricity from renewable sources flour-
ished in many countries. However, the lack of harmonisation between individual Mem-
ber States to achieve targets led to concerns about fragmentation of the electricity mar-
ket (Glachant and Ruester, 2013). There were also concerns about cost-effectiveness of a
nationally determined strategy from a pan-European perspective (Commission, 2016).
The latest package of measures released by the European Commission proposes a
single Europe-wide target of 27% renewables by 2030, without translating them into
national-level targets. In addition, it proposes that support of renewables is opened
to other Member States to ensure cross-border tradability and address fragmentation
of the internal market. In this regard, it makes bold strides by specifying that support
for at least 10% of the newly supported capacity in each year between 2021 and 2025,
and 15% between 2026 and 2030, should be open to installations in other member states
5 (Commission, 2016). This is a major departure from its earlier strategy, which paid much
importance to nationally-determined policies.
Significant amounts of intermittent RES-E in the energy mix have led to unintended
impacts. One of the most significant consequences is the so-called merit-order effect,
where the spot market electricity price reduces to the extent by which the renewable
electricity generation displaces demand along the merit order curve (Sensfuss et al.,
2008). It has been empirically demonstrated by a number of studies (Cludius et al., 2014;
Ederer, 2015; Gelabert et al., 2011; O’Mahoney and Denny, 2013; Traber and Kemfert,
2009, 2011; Weigt, 2009). It is also possible that the presence of a huge share of renewable
electricity will lead to substantial changes in import and export, leading to the spreading
of the merit order effect across national borders (Sensfuss et al., 2008). The problem ad-
dressed in this research is to understand such cross-border impacts between the Nether-
lands and Germany. We also seek to understand the role that design elements of RES-E
support could play in such cross-border effects.
Relatedly, the future of nuclear and coal-fired power plants appears bleak in the
Netherlands and Germany. In 2011 it was mandated that nuclear power was to be phased
out by 2022 in Germany. The German Climate Action Plan 2050 mentions that ”the fed-
eral German government in its development cooperation does not lend support to new
coal power plants”, and includes a commission for ”Growth, Structural Change, and Re-
gional Development” (Federal Ministry for the Environment, 2016). Similarly, the Dutch
parliament voted for a 55% cut in CO 2 emissions by 2030, which would require the clo-
sure of all of the country’s coal-fired power plants (Neslen, 2016). In Germany and the
Netherlands, such a trend limits the options of flexibility primarily to gas-fired power
generation. Biomass and hydro-power are other options, but less prevalent in these two
countries
The objective of this work is to analyse the long-term, cross-border welfare impacts
of different renewable support schemes in Germany and the Netherlands, while taking
into account the proposed phase-out of nuclear and coal power. The performance in-
dicators for the assessment are average wholesale electricity price trends, the costs of
5.1. I NTRODUCTION 77

subsidies, and the income transfer between the producers and consumers under various
schemes, and different interconnection capacities. As outlined in the previous chapters,
the RES-E schemes are identified by their constituent design elements: price or quantity
warranty, technology specificity or neutrality and ex-post vs. ex-ante price setting.

5.1.1. L ITERATURE REVIEW

O N HARMONISATION OF RES-E SUPPORT
During the early stages after the liberalisation of the electricity market, when the first
RES-E Directive was released in 2001, there was a vibrant academic discussion on the
most effective way to support renewable electricity in Europe. The objective of these
works, briefly described below, was to assess which RES-E support design for EU would
provide the most cost effective way of attaining RES-E targets.
Much of the quantitative work in comparing RES-E schemes was done from a purely
neoclassical economics perspective, with equilibrium models. Qualitative assessments
of RES-E support schemes however, such as from the policy analysis field, tend to be
broader in their theoretical foundations. For instance, Finon and Perez (2007) use a
transaction cost perspective to compare RES-E support instruments, and Jacobsson and
Lauber (2006) combine an ‘economics of innovation’ analysis (linking diffusion patterns
5
to actual policies) with a ‘politics of policy’ analysis (explaining the choice of policies in
the larger political context) to explain diffusion of renewable electricity technologies in
Germany. Others use empirical methods to perform assessments Johnstone et al. (2010);
Lewis and Wiser (2007); Haas et al. (2011).
Harmonisation can be defined as a top down implementation of common, binding
provisions concerning the support of RES-E throughout the EU (Bergmann et al., 2008).
The greatest merit of harmonisation of renewable support schemes, from an economist’s
perspective, is greater cost effectiveness from better resource allocation. And indeed this
was demonstrated by researchers who have conducted static theoretical analysis with
equilibrium models (del Rıo, 2005). Such results were corroborated by simulations that
indicate significant cost reductions in a harmonised scenario, as compared to a business
as usual (BAU) scenario. Huber et al. (2004) simulated the electricity market with the
Green-X model, and RES-E policies, by deriving endogenously changing supply and de-
mand curves, to generate market equilibriums, for every year (tick) in the model. Voogt
et al. (2001), using the REBUS model, indicated a 15% cost reduction in the harmonised
RES-E support scenario as compared to business as usual. The REBUS model is an Ex-
cel spreadsheet calculation tool which, given individual cost curves, and potentials of
renewable electricity options, computes the effects of an international burden sharing
tool.
In a recent work del Rio et al. (2017), the aspect of differences between support schemes
of neighbouring countries is analyzed. The authors model two main dimensions: the
degree of harmonisation of policies, and the RES-E support instruments. They find that
the differences in policy costs can be attributed more to the instruments rather than
the degrees of harmonisation. This result holds special relevance given that the recent
package of measures announced by the EC proposes strong measures towards "opening
of support schemes" to other Member States. 1 While the authors employ a detailed
1 While "opening of support schemes" does not amount to harmonisation, it does force MSs to invest in coun-
 5. C ONGRUENCY BETWEEN NATIONAL SUPPORT SCHEMES AND AN INTERNATIONAL
78 ELECTRICITY MARKET

representation of the costs and benefits of RES-E technologies in all EU countries, to

the knowledge of this author, they do not specifically model the electricity market sepa-
rately, and the ensuing cross-border price effects that occur due to the addition of RES-E
capacity in one country. These cross-border price impacts are important as they indicate
possible consequence on the costs of subsidy in the neighbouring country.
Others (Bruninx et al., 2013) model the impact of the German nuclear phase out on
Europe’s electricity generation via an optimisation model. In their results, nuclear energy
is replaced by coal and lignite. Since the publication however, most Central Western Eu-
ropean countries have announced a lignite phase out, and the future of coal seems bleak.
This work focuses solely on the Netherlands and Germany, and includes detailed repre-
sentations of their coal and nuclear phase out plans, and studies the impact of different
capacities of interconnection in the system.

5.2. E XPERIMENT DESIGN AND DATA

5.2.1. E XPERIMENT DESIGN
In order to study the cross border impacts, experiments were conducted using an agent
5 – based model of the electricity market with an endogenous investment algorithm, de-
scribed in Chapter 3. The impact of different renewable support schemes in one country
on the neighbouring country, under different levels of interconnection capacity between
them were tested. The renewable support scheme designs are characterised in term of
three main features, that are referred to as ‘design elements’: price vs. quantity war-
ranty, technology neutrality vs. specificity, and price setting ex-ante vs. ex-post. Each
renewable support scheme is composed by a combination of the three design elements,
leading to a total of 2 power 3 or 8 unique support schemes. In this experiment design,
for a certain interconnection capacity which remains constant throughout the simula-
tion, 8 scenarios are simulated: with each scenario testing a different support scheme
in Germany, while the support scheme in the Netherlands remains the same. This ex-
periment is run at three levels of interconnection capacity: 0MW, the current capacity
of 3950MW, and a high value of 10000MW. The interconnection capacities remain the
same throughout the simulation period which runs for 40 ticks, each tick representing
one year. The investments in any scenario are determined by decisions of the agents
involving cost benefit analysis incorporating the costs of the technology, the expected
revenue from the electricity market, and applicable subsidies. The following sections
outline the main results of the experiments.

5.2.2. D ATA
This section complements the data presented in Chapter 3. The data below is exclusive
to the experiments performed for this chapter.
Fuel costs For the purpose of the experiments here, the fuel costs are assumed to
remain constant at an average of the 2015 levels, through the length of the simulation.
The data sources and assumptions are listed in Table 5.1.
Technology targets and potentials At the official European level, targets for renew-
able electricity have not been set for individual member states beyond 2020. However

tries outside of their own in a bid to improve cost-effectiveness of investments.

5.2. E XPERIMENT DESIGN AND DATA 79

5
Figure 5.1: Experiment design for analysis of cross border effects

Table 5.1: Assumptions for fuel costs, and corresponding data sources

Fuel name Average Price In Unit Average Price Source

in 2015 in 2015 (con-
verted to
Eur/GJ)
Coal 61.619 US Dollars per Met- 1.896 (IMF) (2017)
ric Ton
Lignite 22.360 USD per short ton 1.460 Administration
(2016)
Uranium 0.005 US Dollars per KWh 1.303 Association
(2017)
Natural Gas 7.441 US Dollars per 6.361 (IMF) (2017)
MMBTU
 5. C ONGRUENCY BETWEEN NATIONAL SUPPORT SCHEMES AND AN INTERNATIONAL
80 ELECTRICITY MARKET

the European Union has set itself "a long-term goal of reducing greenhouse gas emis-
sions by 80-95%, when compared to 1990 levels, by 2050" (Commission, 2017). In view
of this, the targets have been set at a linearly increasing rate from 2015 levels in both
countries to 70% at 2050. The targets are shown pictorially in Figure A.4 and A.5 in the
appendix. The data points for ’realistic potentials’ at different years have been used to
linearly extrapolate trends for the whole time scope of the model. The data points and
their sources are mentioned in Table A.2.
Phasing out technologies As mentioned in the introduction, Germany announced
plans to phase out their nuclear power plants in 2022. Similarly, in 2015, Germany’s eco-
nomic ministry and energy companies agreed to take lignite fired power plants offline
(Environment, 2015). In the Netherlands too, a closure of all of the country’s coal-fired
power plants is expected to take place before 2030 (Neslen, 2016). The intial portfolio in
the simulation resembles the generation portfolios of the Netherlands and Germany in
2015 closely. The planned phasing out of coal and lignite is implemented in the model,
even if exact years for the decommissioning have not been proposed for all technologies.
Nuclear power plants in Germany are phased out in 2 stages of 5.GW each in 2018, and
5 2022. Lignite power plants are phased out in three stages: 8.18GW in 2020, 7.52GW in
2025, and 3.97GW in 2030. In the Netherlands, 1.1 GW of Lignite is assumed to be shut
down in 2020 (Maasvlakte 1 and 2 ), another 1.02 GW is assumed to be shut down in
2026, representing the closure of Centraale Maasvlakte 3. Furthermore, the model as-
sumes that there will be no additional investment in either Coal PSC, Lignite, or Nuclear
technologies in either country.

5.3. R ESULTS
5.3.1. I MPACTS OF INTERCONNECTION ON ELECTRICITY PRICES
The primary ‘medium’ so to say, by which electricity spot markets and renewable support
schemes interact is through the electricity spot market price. The fundamental mecha-
nism behind the interaction is as follows: the production of cheap variable renewable
electricity (VRE) causes a reduction in the average electricity prices. This has been doc-
umented and analysed by several studies (Sensfuss et al., 2008; Hirth, 2013). Depending
on the amount of interconnection with the neighbouring market, the relative size (ca-
pacity in MW) of the neighbouring price zone, and the design of the support scheme, the
reduction in electricity prices spread. That is to say, the average electricity price in the
neighbouring country also changes. And since subsidies are a difference between cost
of the technology and revenue from the electricity market, they are directly linked to the
electricity price, due to which, costs of renewable support schemes are also affected.
The spreading of the merit order effect between Germany and the Netherlands is
indicated in Figure 5.2. The figure shows the progression of average (annual) electric-
ity price in Germany and the Netherlands across time for each of the eight RES-E sup-
port schemes, at different interconnection capacities. At the outset, as one may ex-
pect, the electricity prices in both price zones converge as the interconnection capac-
ity increases, irrespective of support scheme design. At an interconnection capacity of
10GW, the prices have almost entirely converged under all scenarios of renewable sup-
port schemes.
5.3. R ESULTS 81

Interconnection 0 MW Interconnection 3950 MW Interconnection 10000 MW

50 Price Zone
Netherlands
Germany
Electricity Price [Eur/MWh]

40
RES−E Policy
P_Ante
P_AnteTS

30 P_Post
P_PostTS
Q_Ante
Q_AnteTS
20 Q_Post
Q_PostTS

0 10 20 30 40 0 10 20 30 40 0 10 20 30 40
Time [in years]

Figure 5.2: Average electricity price plot across different interconnection capacities, for each RES-E policy sce-
nario, in both the Netherlands and Germany

5
Since European guidelines indicate that support schemes have to be ‘market-based’,
subsidies are designed to cover only the part of renewable electricity costs that is not
covered by the electricity market. Therefore, as the revenue from the electricity mar-
ket decreases due to the spreading of the merit order effect, the cost of subsidies in the
neighbouring country increases. This is explained in detail in the following section.

5.3.2. E FFECTS ON COSTS TO CONSUMER

In this model, the costs to the consumer are driven primarily by two factors: the price
of electricity and the cost of subsidy. In Figure 5.2, we observe that costs of electricity
decrease in the Netherlands as interconnection capacity increases. This effect is partic-
ularly evident when the scenarios of medium (3950 MW) and high (10000 MW) inter-
connection capacities are compared. For Germany however, the electricity price level
remains more similar across interconnection levels, with the exception of certain peak
prices being absent as interconnection capacities increase.
The cost of the subsidy as mentioned above, is subject to the revenue that producers
earn from the electricity market, and to the costs of the technology. Figure 5.2 demon-
strates that the electricity prices decrease in the Netherlands as interconnection capacity
with Germany increases. Correspondingly, during the first 30 years, the subsidy costs in
the Netherlands increase in scenarios with highest interconnection. This trend can be
observed in Figure 5.3. The question that follows is why these trends appear, and how
these trends impact the overall welfare of the consumers, that is, how total costs of elec-
tricity, subsidy, and both together affect the consumer over the entire period of the sim-
ulation, under different scenarios. These results are explained in the following sections
in detail.

C OSTS OF S UBSIDY IN THE N ETHERLANDS

In Figure 5.3, a contradiction is apparent and deserves explanation. On the one hand
subsidy costs in the Netherlands increase as interconnection increases upto year 30 (appr.),
 5. C ONGRUENCY BETWEEN NATIONAL SUPPORT SCHEMES AND AN INTERNATIONAL
82 ELECTRICITY MARKET

Interconnection 0 MW Interconnection 3950 MW Interconnection 10000 MW

40

30

Germany
20
Subsidy Costs [Eur/MWh]

10

40

30

Netherlands
20

10

5
0 10 20 30 40 0 10 20 30 40 0 10 20 30 40
Time [in years]

P_Ante P_Post Q_Ante Q_Post

RES−E Policy
P_AnteTS P_PostTS Q_AnteTS Q_PostTS

Figure 5.3: Subsidy costs per MWh electricity generated in the Netherlands and Germany across time

however beyond year 30, subsidy costs decrease as interconnection capacity increases.
This is explained by the fact that beyond year 30, at low interconnection levels there is
substantial spillage in the system, i.e., the intermittent electricity production surpasses
demand frequently. This leads to an increase in costs per unit of production, thereby
increasing the subsidy costs. This is also corroborated by Figure B.5, and Figure B.6 in
the appendix, which show the spillage of RES-E production with time, per technology.
In the first phase of the simulation before year 30, where spillage does not play too great
a role, the following results are observed. In ‘Interconnection 10000’ scenarios, the sub-
sidy costs in the Netherlands are higher by a range of 10 - 2 Eur/MWh as compared to the
‘Interconnection 3950 MW’ scenarios (current levels), and by 18 - 2 Eur/MWh as com-
pared to the ‘Interconnection 0 MW’ scenarios. Since spillage substantially affects the
costs of subsidies beyond year 30, it is useful to see results separately until year 30, as in
Table 5.2, and for the full simulation as in Table 5.3.
The results demonstrate that greater interconnection capacities with Germany lead
to higher subsidy costs in the Netherlands, due to lowering of electricity prices, as long
as shares of RES-E production in the Netherlands remain lower than approximately 60%,
which is approximately the share of renewable electricity at year 30. As shares increase,
the level of interconnection plays a progressively higher role in the subsidy costs, due to
spillage.
The questions that follow are what factors these figures are sensitive to and why. The
amount of spillage is sensitive to the technology, the interconnection capacity, and the
5.3. R ESULTS 83

Table 5.2: Costs of electricity and subsidy, summed for the first 30 years [ticks], averaged across different RES-E
policy scenarios, in bn euros.

Country Interconnection Cost of Cost of Cost of Cost of

Capacity [in electricity electric- Subsidy Subsidy,
MW] in DE [in ity, SD, in in NL [in SD, in NL
bn Eur] DE [in bn bn Eur] [in bn
Eur] Eur]
Netherlands 0 159.95 2.46 54.60 1.26
Netherlands 3950 156.29 2.53 55.14 1.15
Netherlands 10000 124.17 1.92 63.72 2.27
Germany 0 537.03 9.94 316.65 37.53
Germany 3950 539.66 7.19 310.01 39.54
Germany 10000 551.04 4.81 308.46 37.94

Table 5.3: Costs of electricity and subsidy, summed for the full simulation period [40 years], averaged across
different RES-E policy scenarios, in bn euros.
5
Country Interconnection Cost of Cost of Cost of Cost of
Capacity [in electricity electric- Subsidy Subsidy,
MW] in DE [in ity, SD, in in NL [in SD, in NL
bn Eur] DE [in bn bn Eur] [in bn
Eur] Eur]
Netherlands 0 209.46 5.11 99.12 2.75
Netherlands 3950 205.39 5.06 92.32 2.36
Netherlands 10000 168.50 5.14 101.16 3.77
Germany 0 730.38 15.99 508.29 51.90
Germany 3950 732.43 7.17 494.42 51.35
Germany 10000 741.72 6.62 488.87 49.52

absence of storage in the system. The dependence of spillage on technology types in

each of the countries is indicated in Figures B.5 and B.6 . The type of support scheme,
and its design would also impact the results, as demonstrated in the previous chapter2
and explained further in the sections below. Other factors that could impact subsidy
costs, but whose sensitivity has not been modelled, due to resource constraints, are
the costs of technologies, their capacity factors, and geographical distribution of power
plants.

C OSTS OF S UBSIDY IN G ERMANY

The subsidy costs in Germany in this experiment depend primarily on the type of sup-
port scheme implemented in Germany. The impacts of design elements on subsidy costs
follow from the explanations provided in Chapter 4, and are briefly described below:
2 The type of support scheme here refers to the country’s own support scheme, and not that of the neighbouring
country’s.
 5. C ONGRUENCY BETWEEN NATIONAL SUPPORT SCHEMES AND AN INTERNATIONAL
84 ELECTRICITY MARKET

• Price warranty schemes vs Quantity warranty schemes: Price warranty schemes

lead to an investment up to the potential of the technology, rather than their tar-
gets, if no budget limit or quantity limits are exogenously set, and assuming the
regulator has perfect knowledge of future electricity prices. In this case, if the frac-
tion of VRE generation is higher than appr. 60% then spillage occurs, leading to
higher subsidy costs than their corresponding quantity warranty schemes. Quan-
tity warranty schemes lead to an investment up to the targets set by the regulator.
If the targets are set in a manner that there is no spillage in the system, then costs
of subsidy remain lower than their price-warranty counter parts. In this experi-
ment, both targets and potentials set for Germany across the 40 years are similar,
see Figure A.4 in the appendix. Therefore the difference in spillage between both
types of schemes is marginal.
• Technology specificity vs. technology neutrality: Technology neutral schemes are
more expensive than their technology specific counter-parts due to wind-fall prof-
its obtained by non-marginal technologies in the technology neutral schemes.
• Ex-post vs. ex-ante price setting: Ex post schemes are more expensive to the con-
sumer, as the regulator underestimate the reduction in electricity prices due to the
5 merit order effect.

Aside from impacts of support scheme design, other factors that impact subsidy
costs in Germany are i. the amount of interconnection and ii. the amount of spillage.
When capacity of interconnection increases the electricity price in the exporting zone
increases. An increase in the revenue from electricity prices implies a reduction in the
costs of the subsidy. On average, across all RES-E policy scenarios, as interconnection in-
creases, subsidy costs in Germany decrease, as is shown in both Tables 5.2 and 5.3. The
standard deviation of the costs of subsidy are quite high for Germany, primarily because
the technology neutral support schemes require much more subsidy than the technol-
ogy specific ones3 . This is evident in Figure B.4 which shows the sum of subsidy and
electricity costs through the full simulation for both countries.

T OTAL C OSTS TO C ONSUMER

The total costs to consumers; across the full period of the simulation, i.e., the sum of
electricity and subsidy costs through 40 years are indicated in Figure 5.4. Although for
the Netherlands the costs of subsidy increase and the costs of electricity decrease, the re-
duction in costs of electricity is higher than the increase in cost of subsidy. This implies
that the total cost for the Netherlands reduces as interconnection increases. This corrob-
orates with results that have been found by authors who studied the merit order effect,
for an uncongested region. In addition it is also evident from Figure 5.4 that the type
of design of the RES-E schemes in Germany does not directly affect costs in the Nether-
lands. Nonetheless, it can be expected that the quantity of RES-E generation in Germany,
determined by its target, has an impact on the electricity prices in the Netherlands and
consequently on its subsidy costs.
The results are not as straightforward for Germany, where the cost trend depends
not only on the interconnection capacities, but also on the type of renewable support
3 Due to high infra-marginal subsidy costs, i.e., high windfall profits for non-marginal technologies
5.3. R ESULTS 85

Netherlands

Plotting Total Costs to Consumers [in Billion Euros]

1000

500

Germany

1000

500

0
te S st tTS te S st tTS
An teT Po s An teT Po s
P_ An P_ Po Q_ An Q_ Po
P_ P_ Q_ Q_

Policy Scenario

Interconnection Level Interconnection 0 MW Interconnection 3950 MW Interconnection 10000 MW 5

Figure 5.4: Total cost to consumer including costs of electricity as well as subsidy over 40 years

scheme active in each scenario. As explained in the above sections, factors such as tech-
nology neutrality of support scheme, spillage, have an impact on the costs of subsidy
and electricity; the costs are reflected in Figure 5.4.

5.3.3. C OSTS AND R EVENUES TO P RODUCER

In this section the impacts of the various policy scenarios on the costs and revenues of
the producer are discussed. Under perfect competition, and long-term equilibrium, the-
ory predicts that profits are zero. However, in this model, neither is perfect information
assumed, nor is economic decommissioning implemented, and infra-marginal profits
play a significant role. In Figure 5.5, the profits of the producers summed across the
length of the simulation (40 years) are indicated per technology, per country, per RES-E
policy in Germany, and for different scenarios of interconnection capacity.
As mentioned in Chapter 3, economic decommissioning of power plants has not
been modelled; the plants are operational through out their technical lifetimes, irre-
spective of whether their long run marginal profits are negative. This has an impact on
CCGT plants as they are largely the marginal generating technology; their long-run prof-
its are negative. For pulverised super-critical coal based power plants (CoalPSC), their
low marginal costs lead to high infra-marginal rents, and positive profits. Since Ger-
many is the larger, exporting country, the profits of German CoalPSC plants increase, as
interconnection capacities increase, while the trend is opposite for the Dutch CoalPSC
plants. Nuclear and lignite plants, with their low short-run marginal costs, and flexible
nature, make higher infra-marginal profits as interconnection capacities increase. Un-
 5. C ONGRUENCY BETWEEN NATIONAL SUPPORT SCHEMES AND AN INTERNATIONAL
86 ELECTRICITY MARKET

Germany Netherlands
0.15
0.10

CCGT
0.05
0.00
−0.05
−0.10
0.15
0.10

CoalPSC
0.05
0.00
−0.05
−0.10
0.15
0.10

Lignite
0.05
Producer Profit in Billion Eur

0.00

5 −0.05
−0.10
0.15
0.10

Nuclear
0.05
0.00
−0.05
−0.10
0.15

Photovoltaic
0.10
0.05
0.00
−0.05
−0.10
0.15
0.10
0.05

Wind
0.00
−0.05
−0.10
0.15 WindOffshore
0.10
0.05
0.00
−0.05
−0.10
IntLow IntMedium IntHigh IntLow IntMedium IntHigh
Interconnection Level

P_Ante P_Post Q_Ante Q_Post

RES−E Policy
P_AnteTS P_PostTS Q_AnteTS Q_PostTS

Figure 5.5: Total profit to producer per policy scenario and per technology across 40 years
5.4. D ISCUSSION 87

forseen4 decommissioning by the German government, also exacerbate their low overall
profits at low interconnection capacities.
As for renewable technologies, the subsidies are designed to just cover costs, and the
regulator is assumed to have the same cost information as the energy producer. This is
reflected in near-zero profits for Solar PV. Since it is the marginal renewable technology
due to its higher costs, the technology neutral and technology specific RES-E policies
both result in near-zero profits. Wind power plants (on-shore) earn-infra marginal prof-
its under technology neutral policy scenarios.

5.4. D ISCUSSION
5.4.1. M ERITS AND LIMITATIONS OF THE MODEL
The popular aphorism, "All models are wrong, but some are useful" seems especially
pertinent when models seek to simulate the next several decades. The uncertainties
in modelling such a timescale are innumerable. While a few of them have been taken
into account, such as bounded rationality of agents regarding future electricity prices,
and uncertainty in demand growth, assumptions have been made regarding many other
parameters. For instance, if costs of variable renewable electricity technologies were to 5
drop at a faster rate than assumed in this study, the costs of subsidy would be even lower,
strengthening the idea that decrease in electricity costs will be greater than the increase
in subsidy costs to the consumer in the neighbouring country. Costs of gas, coal, and
uranium are impacted by global trends and difficult to predict in the long-term. In this
work they have been assumed to be constant at 2015 levels. While this assumption helps
the analyst focus solely on the variables of interest, such as levels of interconnection and
renewable support scheme designs, in reality fuel costs might have an impact on the
results.
Despite its many limitations the model provides several insights on cross-border ef-
fects caused due to national support schemes alongside an international electricity mar-
ket. Results show that given a certain interconnection capacity, the designs of renewable
support schemes and their targets do lead to unintended effects on subsidy costs in a
neighbouring country. However, these effects are primarily caused by virtue of the in-
stalled capacity of intermittent renewable electricity that a support scheme design in-
centivizes, as against the design elements themselves.
The results also indicate that at high shares of RES-E production, interconnections
play an important but ultimately limited role in lending flexibility to the system, and
consequently in ensuring that each unit of intermittent electricity is consumed in an
efficient manner. They indicate that the presence of flexibility has strong implications
on the targets of intermittent renewable generation the European Union or individual
countries can set for themselves, without having to suffer costs of spillage in the form of
increased subsidies. The targets to be set each country must take into account, ceteris
paribus, firstly its interconnection capacities with neighbouring countries. At a later
point in time when spillage of intermittent renewable generation becomes inevitable
in the absence of storage, countries must also take into account the amount of storage
available to them while setting their targets. An ex-ante price setting scheme ameliorates
4 at the time of investment
 5. C ONGRUENCY BETWEEN NATIONAL SUPPORT SCHEMES AND AN INTERNATIONAL
88 ELECTRICITY MARKET

Figure 5.6: EPEX-SPOT Data showing day ahead electricity prices and volume for 2015, for base load (grey) and
5 peak load (orange) in the DE/AT price zone. It also shows the 200 day average prices. (SPOT, 2015)

the impact of the merit order effect on subsidy costs, at the risk of renewable energy pro-
ducers incorporating their price risks into the cost of capital.

5.4.2. VALIDATION : C OMPARISON TO OBSERVED DATA

Validity, as described in section 3.6, is an investigation of the model’s adequacy for pur-
pose. One strategy to validate models is to check whether the simulation output fits
closely with observational data, while treating the simulation as a black box. Although
comparing output from the model for the next 30 years to actual observational data is
impractical, it is possible to use historical data as input and observe the outputs while
accounting for assumptions made.
In this chapter, data such as fuel prices, the expected decommissioning of plants
(nuclear and lignite) are inputs. Since cross border effects were the main purpose of the
model, we used constant fuel prices of 2015 levels, to reduce the variability due to less
important factors in the experiment. If the model works as expected, the outputs or elec-
tricity prices should match the prices that were observed in Germany and the Nether-
lands in 2015. The range of price levels under existing interconnection levels between
simulated and observed data are comparable. EPEX-SPOT data in Figure 5.6 shows that
the 200 day average price ranges lie between 30 and 40 Eur/MWh, whereas the actual
day-ahead prices vary from less than 0Eur/MWh upto 60 Eur/MWh (SPOT, 2015).
The simulation indicates comparable price ranges after year 7, at which point Lig-
nite power plants are phased out in Germany. Upto year 7 however, we see average
prices of between 16 Eur/MWh at zero interconnection capacity, upto 23 Eur/MWh at
10 GW interconnection capacity. The reason for much lower prices in Germany before
year 7 could be attributed to the fact that all interconnection capacities between Ger-
many and surrounding countries have not been modelled. The export to the Nether-
lands accounts for only a fraction of its total export. This is corroborrated by data from
5.5. C ONCLUSIONS 89

the ENTSO-E, which indicated that during 2015, the export to the Netherlands from Ger-
many accounted for an average of 29% [ranging between 16% and 45%] of its total ex-
ports (ENTSO-E, 2015).

5.4.3. D ISTORTION OF THE INTERNAL ENERGY MARKET

One could wonder whether the unintended effects are indeed a “distortion” of the inter-
nal energy market, as has been described in some key documents of the European Com-
mission (Commission, 2016) and (Commission, 2015). Similarly, some researchers have
termed such effects a “fragmentation” of the internal market due to national level poli-
cies (Glachant and Ruester, 2013). As far as national RES-E schemes are concerned, these
cross-border effects, while unintended, are but a fiercer manifestation of the merit order
effect, which affects VRE importing countries whose sizes are relatively smaller. This is
not necessarily an adverse effect, as the effect would have persisted even if countries
were isolated import of VRE, as shown above also brings with it a significant reduction
in electricity prices, whose benefits to the consumer outweigh the increase in subsidy
costs. As far as the simultaneous goals of consuming greater amounts of VRE, and hav-
ing a single internal electricity market are concerned, such a result is only reasonable, 5
and should be welcome!
The question of consistency or distortion then lies in whether the current institutions
ensure that the investment signal, and not just the “consumption (or production) signal”
is also commensurate with the principle of “low-hanging fruit”. RES-E support schemes
as they operate today predominantly incentivize cost-effective investments within na-
tional boundaries. While this arrangement has proved to be effective in inducing large
amounts of RES-E generation, from a pan-European perspective concentration of ef-
forts in certain member states has proved to be expensive. A popular example is Ger-
many’s expensive investment in electricity from photovoltaic panels in Germany itself,
while the same investment in Southern Europe could have gained from much higher ca-
pacity factors. The recently proposed RES-E directive (Commission, 2016) includes an
article which mandates that member states open support for RES-E to generators lo-
cated in other member states. The article specifies capacities as well: "Member States
shall ensure that support for at least 10% of the newly-supported capacity in each year
between 2021 and 2025 and at least 15% of the newly-supported capacity in each year be-
tween 2026 and 2030 is open to installations located in other Member States." The new
proposed scheme would help ameliorate large differences in cost effectiveness. The ef-
ficiency gains of such a tactic to members of the country investing in another country
however is constrained by interconnection capacities. Even sufficient interconnection
will only be useful up to an extent, beyond which having storage will become a necessity
to include greater amounts of RES. That limit, for the Netherlands and Germany put to-
gether seems to lie at 60% annual generation from variable renewable electricity in both
countries.

5.5. C ONCLUSIONS
The primary objective of this chapter was to investigate cross-border effects due to na-
tional RES-E support schemes, operating in an international electricity market. In order
 5. C ONGRUENCY BETWEEN NATIONAL SUPPORT SCHEMES AND AN INTERNATIONAL
90 ELECTRICITY MARKET

to carry out this research, EMLab Generation, a model comprising of both an electric-
ity market clearing module, and an endogenous investment algorithm was used. Price
zones based on Germany and the Netherlands were simulated, to test the effects of dif-
ferent support scheme designs in each country, and the common electricity market, on
distributional implications under differnent scenarios of interconnector capacities, over
a long term period of 40 years, starting from 2015.
As economic theory would predict it is found that subsidy costs increase in the smaller
country, due to spreading of the merit order effect from the larger neighbouring country.
Increasing the interconnection capacity between Germany and the Netherlands from
the current 3950 MW to 10000MW would lead to an increase of Eur 8.58 billion5 in Dutch
subsidy costs, under assumed cost structures until 2045. However, the increase in sub-
sidy costs remains lower than the reduction in costs of electricity (Eur 32.12 billion),
corroborrating earlier, single country analyses on the topic. Therefore, total cost in the
neighbouring country reduces as interconnection increases, as the invisible hand works
its magic. Another important insight is that as share of RES-E increases, interconnection
has limited impact on reducing spillage of RES-E while storage becomes increasingly im-
portant. The results show that targets for RES-E, depend not only on the targets of the
5 neighbouring countries, but also the capacity of storage that exists in the system.
Several policy recommendations follow from the results and analyses. The results
show that it is necessary that policy makers and researchers ponder over questions about
how subsidy costs should be allocated between countries over the long term when they
are so intricately connected with the electricity markets they operate in. This should
be part of the European Commission’s research agenda. Futher, it is evident that while
setting RES-E targets either for individual member states, or for several together, the ca-
pacity of flexible sources available to them should be explicitly taken into account. The
targets that can be set in a cost efficient manner are limited by the amount of intercon-
nection and storage available to that system.
This research also opens up several avenues for future work. The results indicate
that shares of RES-E greater than 60%, is not a cost efficient proposition without stor-
age. However this figure is accompanied by disclaimers of limited temporal resolution
in the model. In order to deduce a precise estimate of the point at which storage be-
comes essential, models with greater temporal resolution are necessary. Since costs of
RES-E technologies have been dramatically dropping over the last few years, a large sen-
sitivity analysis could help answer the question, "At what costs do subsidies become
redundant". Another interesting question is to evaluate cross border effects under other
institutional arrangements such as bilateral contracts, capacity markets, and perhaps
the EU-ETS.

5 At 2015 levels, without accounting for inflation

R EFERENCES 91

R EFERENCES
E. Commission, Directive 2009/28/EC, Official Journal of the European Union 52 (2009),
10.3000.

J.-M. Glachant and S. Ruester, The EU internal electricity market : done forever?, Working
Paper (2013).

E. Commission, Proposal for a DIRECTIVE OF THE EUROPEAN PARLIAMENT AND OF

THE COUNCIL on the promotion of the use of energy from renewable sources (recast),
Tech. Rep. COM(2016) 767 final (European Commission, Brussels, 2016).

F. Sensfuss, M. Ragwitz, and M. Genoese, The merit-order effect: A detailed analysis of

the price effect of renewable electricity generation on spot market prices in Germany,
Energy Policy 36, 3086 (2008).

J. Cludius, H. Hermann, F. C. Matthes, and V. Graichen, The merit order effect of wind
and photovoltaic electricity generation in Germany 2008–2016: Estimation and distri-
butional implications, Energy Economics 44, 302 (2014). 5
N. Ederer, The market value and impact of offshore wind on the electricity spot market:
Evidence from Germany, Applied Energy 154, 805 (2015).

L. Gelabert, X. Labandeira, and P. Linares, An ex-post analysis of the effect of renewables

and cogeneration on Spanish electricity prices, Energy Economics 33, S59 (2011).

A. O’Mahoney and E. Denny, Electricity prices and generator behaviour in gross pool elec-
tricity markets, Energy Policy 63, 628 (2013).

T. Traber and C. Kemfert, Impacts of the German Support for Renewable Energy on Elec-
tricity Prices, Emissions, and Firms, The Energy Journal 30, 155 (2009).

T. Traber and C. Kemfert, Gone with the wind? — Electricity market prices and incen-
tives to invest in thermal power plants under increasing wind energy supply, Energy
Economics 33, 249 (2011).

H. Weigt, Germany’s wind energy: The potential for fossil capacity replacement and cost
saving, Applied Energy 86, 1857 (2009).

B. a. N. S. Federal Ministry for the Environment, nature Conservation, Climate Action

Plan 2050: Principles and goals of the German government’s climate policy, (2016).

A. Neslen, Dutch parliament votes to close down country’s coal industry, (2016).

D. Finon and Y. Perez, The social efficiency of instruments of promotion of renewable en-
ergies: A transaction-cost perspective, Ecological Economics 62, 77 (2007).

S. Jacobsson and V. Lauber, The politics and policy of energy system transforma-
tion—explaining the German diffusion of renewable energy technology, Energy Policy
34, 256 (2006).
 92 R EFERENCES

N. Johnstone, I. Haščič, and D. Popp, Renewable Energy Policies and Technological Inno-
vation: Evidence Based on Patent Counts, Environmental and Resource Economics 45,
133 (2010).

J. I. Lewis and R. H. Wiser, Fostering a renewable energy technology industry: An inter-

national comparison of wind industry policy support mechanisms, Energy Policy 35,
1844 (2007).

R. Haas, C. Panzer, G. Resch, M. Ragwitz, G. Reece, and A. Held, A historical review of

promotion strategies for electricity from renewable energy sources in EU countries, Re-
newable and Sustainable Energy Reviews 15, 1003 (2011).

J. Bergmann, C. Bitsch, V. Behlau, and S. Jensen, A European harmonised policy to pro-

mote RES-electricity - sharing costs and benefits, Tech. Rep. D17 (Vienna University of
Technology, 2008).

P. del Rıo, A European-wide harmonised tradable green certificate scheme for renewable
electricity: is it really so beneficial? Energy Policy 33, 1239 (2005).
5
C. Huber, T. Faber, R. Haas, G. Resch, J. Green, S. Olz, and S. White, Deriving optimal
promotion strategies for increasing the share of RES-E in a dynamic European electricity
market, Tech. Rep. (2004).

M. H. Voogt, M. A. Uyterlinde, K. Skytte, M. Leonardi, and M. H. Whiteley, Renewable en-

ergy burden sharing. REBUS. Requirements and expectations of utilities and consumer
organisations in the European energy sector, Tech. Rep. (Energy research Centre of the
Netherlands ECN, Petten (Netherlands), 2001).

P. del Rio, G. Resch, A. Ortner, L. Liebmann, S. Busch, and C. Panzer, A techno-economic

analysis of {EU} renewable electricity policy pathways in 2030, Energy Policy , (2017).

K. Bruninx, D. Madzharov, E. Delarue, and W. D’haeseleer, Impact of the German nuclear

phase-out on Europe’s electricity generation—A comprehensive study, Energy Policy 60,
251 (2013).

I. M. F. (IMF), IMF Cross Country Macroeconomic Statistics, (2017).

U. E. I. Administration, Annual Coal Report 2015, Tech. Rep. (US Department of Energy,
2016).

W. N. Association, Nuclear Power Economics | Nuclear Energy Costs, (2017).

E. Commission, 2050 Energy strategy - European Commission, (2017).

D. Environment, The end of lignite coal for power in Germany, (2015).

L. Hirth, The market value of variable renewables: The effect of solar wind power variabil-
ity on their relative price, Energy Economics 38, 218 (2013).

E. SPOT, EPEX SPOT SE: Day-Ahead Auction, (2015).

R EFERENCES 93

ENTSO-E, Country Data Packages, (2015).

E. Commission, Launching the public consultation process on a new energy market de-
sign, Tech. Rep. COM(2015) 340 final (Brussels, 2015).

5
 6
RES-E POLICY DESIGN IN
PERSPECTIVE OF THE ENERGY
TRANSITION

95
 96 6. RES-E POLICY DESIGN IN PERSPECTIVE OF THE ENERGY TRANSITION

6.1. I NTRODUCTION
In auctions for wind offshore projects organised in Germany in April 2017, three of the
four winning projects bid at zero subsidy. In such auctions, we see mounting evidence of
dramatic reductions in technology costs, caused in part (for wind offshore) by advance-
ments in turbine technologies, standardization and optimisation of designs. At the same
time, governments have become increasingly optimistic that zero subsidy becomes the
norm; for instance, the Dutch government announced plans to organise subsidy-free
auctions in 2026, and the UK is not too far behind (Wind, 2017). Despite this optimism
from the wind offshore industry and governments, the question we ask is whether such
optimism is justified - not just in 2026, but also in the longer-term perspective of three
to four decades in the future, given the current institutional settings that we operate in.
In Europe, the energy transition comprises a multitude of policies and strategies
as multiple objectives simultaneously compete for attention: the EU ETS, demand re-
sponse measures, capacity mechanisms etc are a few. The recent package of measures
released by the European Commission, ‘Clean Energy for All Europeans’, further pro-
motes the objective of decarbonisation by proposing strict measures to ensure that at
least 27% of all energy consumed comes from renewable sources by 2030, while steadily
moving towards reducing emissions by 80-95% in Europe by 2050.Dramatically decreas-
ing costs of solar and wind generation technologies have aided the transition further, not
just in Europe, but the world over: costs of solar PV are half as much as they were in 2010,
6 and are expected to fall by a further 60% over the next decade (IRENA, 2016).
As has been a recurring theme in the preceding chapters, there have been unin-
tended effects caused by the huge influx of intermittent renewable electricity into the
electricity market. The first of these unintended effects is the "merit-order" effect, due
to which renewable electricity generation has a strong price reduction effect on whole-
sale electricity spot prices (Sensfuss et al., 2008). A second effect, which directly relates
the presence of variable renewable electricity (VRE) to its economic value, is the idea of
the "market value of VRE", as defined and quantified by Hirth (2013). Due to the simple
fact that VRE1 is inherently variable with time and has limited predictability, its eco-
nomic value in the spot market is heavily influenced by its own existence. As the share of
VRE increases, its economic value reduces. Due to this reason, it is questionable whether
support for renewables can be phased out as their share increases.
Three elementary aspects of the energy transition are especially important: the elec-
tricity spot market, the EU ETS, and finally, flexible resources. The electricity spot mar-
ket represents a set of institutions which form the essence of a well-functioning, liber-
alized market for electricity. They signal the value of electricity as a commodity, and
in principle, incentivize investment in the most cost-efficient generation sources. The
relationship of the cost of subsidies to VRE with the price of electricity is therefore, fun-
damental to the assessment of the economic viability of VRE. The EU-ETS is a pillar of
the European Union’s energy strategy. It was envisioned as a tool to price carbon-dioxide
emissions, and therefore incentivize low-carbon sources of electricity generation in the
power sector as well. It is, therefore, imperative to assess how this ETS mechanism will
impact the costs of subsidies as the share of RES-E in the system increases. Finally, tem-

1 For the purposes of this chapter, VRE and RES-E are used interchangeably
6.2. S UBSIDY COSTS AND ELECTRICITY PRICES 97

porally flexible resources such as storage, and locational-based flexible resources such as
interconnection are resources essential for controlling the variability, and consequently
the market value of VRE. The need for flexibility, and its impact on the possible cost of
subsidy is discussed.
The aim of this chapter is therefore to touch upon these three aspects of the energy
transition. In light of the intricate relationship between the subsidy designs and the in-
ternal electricity market, the first section comprises a discussion on the components that
drive RES-E subsidies. The second section of this chapter is a deliberation on the role the
EU-ETS could play in the interaction between subsidy costs and electricity price. The
third section reviews scientific literature regarding the role of flexible resources, such as
interconnection and storage, on the market value VRE, and juxtaposes it with results ob-
tained from the preceding chapters. Finally, the chapter culminates with a comparison
between factors which could influence subsidy costs over the short term versus the long
term.

6.2. S UBSIDY COSTS AND ELECTRICITY PRICES

6.2.1. D RIVERS OF RES-E SUBSIDY
The standard method to make an investment decision is to carry out a cost-benefit anal-
ysis, which culminates in a net present value (NPV) calculation. An NPV calculation
takes into account all cashflows during each year of the project operation, and discounts 6
it to the present value. The subsidy, similarly, is calculated as the difference between
the costs and revenues for a given project. As long as the costs are greater than the rev-
enue from the electricity market, subsidy programmes will be necessary to incentivize
investment. Through the course of this PhD project, technology costs have dropped
significantly in the industry, and there have been interesting analyses on factors influ-
encing revenues from the spot market price. These issues have been discussed in this
subsection.

C OSTS
The cost comprises three quantities: the technology costs, the operation and mainte-
nance costs, and the cost of capital. Technology costs have seen a rapid decline over the
last few years for solar, wind, and wind-offshore technologies.
In the case of solar, global utility-scale PV system costs have reduced by 62% between
2009 and 2015 (IRENA, 2016). The potential for further cost reduction is expected to be
large: the global average of installed-costs could reduce by 57% in 10 years from 2015 to
2025 (ire, 2016). The utility scale cost trends for solar are presented in Figure A.6.
Multiple sources report a similar cost decline in the wind industry: in the order of
60% between 2009 and 2016. For instance, Bloomberg New Energy Finance finds that in
Europe, the costs of building an offshore wind farm has fallen by 22% solely over 2016-17
(Finance, 2017a). They expect levelized costs of offshore wind to fall by a further 71% by
2040 (Finance, 2017b). A report from the Lawrence Berkeley National Laboratory finds
that wind turbine prices in the USA have fallen between 20% and 40% from 2008 until
2016 (Wiser et al., 2016).
 98 6. RES-E POLICY DESIGN IN PERSPECTIVE OF THE ENERGY TRANSITION

Figure 6.1: Comparison of electricity traded volumes in day-ahead, forward, OTC markets in the third quarter
of 2016. Source for Energy (2016)

R EVENUES
In western Europe, substantial shares of electricity are either traded on organised trad-
ing platforms such as the spot market or bilaterally on the so-called over-the-counter
(OTC) markets. In different countries the shares traded in each market might be differ-
6 ent. For an indication, see Figure 6.1. The day-ahead markets such as the EPEX-Spot
in the central western Europe region are transparent with data, and therefore provide a
useful source for analysing revenues from the electricity market. In this section, we seek
to understand what the current drivers of spot market electricity prices are.
In the body of literature on electricity price drivers, most existing studies are em-
pirical, and focus exclusively on the merit order effect of renewables; for instance Sens-
fuss et al. (2008); Cludius et al. (2014); Ederer (2015); Gelabert et al. (2011); O’Mahoney
and Denny (2013); Traber and Kemfert (2009, 2011); Weigt (2009); Würzburg et al. (2013).
Bublitz et al. (2017) provide a concise and insightful literature review of such studies. In
addition, unlike the aforementioned studies, they determine and compare price drivers
with the objective of ascertaining whether the merit order effect is indeed the biggest
cause of the price drop in the German electricity market (Bublitz et al., 2017). They
use two methods: a linear regression model and an agent-based model to test price
effects from different factors on the spot market price. Contrary to popular opinion,
they find that between 2011 and 2015 the impact of carbon and coal prices was twice as
high as the impact of renewable expansion. Other studies look at effects of other factors
such as the influence of neighbouring countries (Dehler et al., 2016), coal and gas prices
(O’Mahoney and Denny, 2013; Dehler et al., 2016), the nuclear moratorium (Thoenes,
2014), and of demand and fuel prices (Hirth, 2016a).
All the above studies are ex-post analyses and use empirical data. They have little
to say about future price trends or determinants of electricity prices in the presence of
high shares of renewable electricity generation. To gain insight into the drivers of elec-
tricity prices decades into the future, long-term energy models are commonly employed.
Given the uncertainties involved in such modelling, it is worth remembering Hamming’s
argument on computation: the purpose of computation is insight, not numbers (Ham-
6.2. S UBSIDY COSTS AND ELECTRICITY PRICES 99

ming, 1962). As a strategy to gain insight under uncertainty, scenario analysis has gained
immense importance since Rand Corporation first used it in the 1940s. For this disser-
tation, scenarios with at least 70% renewable electricity generation by 2055 are consid-
ered. There are few studies modelling such scenarios specifically for Germany and the
Netherlands which account for factors such as uncertainties in carbon prices, coal and
gas prices, demand growth, etc. on the electricity prices.
Simulations conducted for this dissertation provide some useful insights. In the fol-
lowing section, we look at results from the model to understand what could be major
electricity price drivers in the long-term period of 20-30 years, as the share of intermit-
tent renewable energy increases in the energy mix.

6.2.2. M ODEL RESULTS ON THE RELATIONSHIP BETWEEN SUBSIDY COSTS

AND ELECTRICITY PRICES
In Chapter 4 of this dissertation, a simple representation of the Netherlands is modelled
with only four technologies: Solar PV, wind, wind offshore, and CCGT (Combined Cycle
Gas Turbine). The results presented in Chapter 4 were mainly concerned with the ef-
fects of different RES-E support designs on social welfare. In this chapter, we focus on
the part of the results that shows the relationship between electricity prices and subsidy
costs under different gas price trends. We use the RES-E support scheme with the fol-
lowing configuration: quantity warranty, technology specificity, and ex-post price setting.
This configuration most closely represents the support scheme currently implemented 6
in the Netherlands, as shown in Chapter 4. In particular, we study three trends: a) Con-
stant, where the gas price remains constant at 4 Eur/GJ through the entire period of the
simulation, b) High, where, starting from 4 Eur/GJ, the gas price increases at a mode of 4
% percent at each tick, and c) Low, where, starting from 4 Eur/GJ, the gas price decreases
at a mode of 2 % at each tick. The progression of the trends in gas prices for each of the
above cases is indicated in figure 6.2.

O BSERVATIONS
Trends in electricity prices (averaged per year) and subsidy costs per year, both expressed
in Eur/MWh, across the length of the simulation period - 0 to 40 years - are shown in Fig-
ure 6.3. The electricity price trends are, as expected, primarily impacted by gas prices.
However, interestingly, they are not impacted as much by the amount of renewable elec-
tricity generation in the mix. Figure 6.3 also demonstrates that the subsidy cost for solar
PV reduces over time, with the rate of reduction differing between scenarios depend-
ing on the electricity price trends. The scenarios differ only in their gas trends, which
are indicated in figure 6.2 as mentioned above. The corresponding fractions of renew-
able electricity generated are indicated in figure 6.5. At the outset, the following major
observations can be made with regard to the relationship between electricity price and
subsidy costs under different gas price trend scenarios:
• In the gas price trend: Constant scenario, a clear pattern does not emerge. There is
no direct relationship between electricity price and subsidy cost.
• In the gas price trend: High scenario, as electricity prices increase, subsidy costs
decrease for all technologies. This can be seen in figures 6.3 and 6.4. This relation-
ship is maintained as the share of renewable electricity in the mix increases.
 100 6. RES-E POLICY DESIGN IN PERSPECTIVE OF THE ENERGY TRANSITION

6 Figure 6.2: Assumptions of gas prices made for Constant, Low, High scenarios.

• In the gas price trend: Low scenario, as electricity prices decrease and as the share
of renewable electricity increases, the subsidy costs do increase, but only slightly
more than in the gas price trend: Constant scenario. This increase is more clearly
visible in Figure 6.3 than in Figure 6.4. In addition, it is also evident that when
electricity prices fall, they cause modest increases in subsidy costs; but that rising
electricity prices cause larger reductions in subsidy costs.

Factors other than gas prices also impact the electricity price and subsidy cost within
the model. The subsidy costs are influenced by fixed costs of the technology, their capac-
ity factors, and the absence of storage or interconnection in the system. This last factor
leads to spillage of intermittent renewable electricity production. As the share of VRE
generation increases, there are more frequent instances of VRE generation being higher
than the demand, which leads to curtailment of RES-E or spillage. Increased spillage
manifests itself as an increase in the per unit (MWh) fixed operation and maintenance
cost of the technology.
The electricity prices in the model are also influenced by low granularity of time pe-
riods within a tick: all 8760 hours in a year are modelled in terms of 20 load segments as
described in Chapter 3 of this dissertation.

I NTERPRETATION
The main effects observed in each of the three scenarios investigated above are explained
here.
6.2. S UBSIDY COSTS AND ELECTRICITY PRICES 101

Figure 6.3: Electricity price and subsidy cost trends [in Eur/MWh] across time. The shaded region indicates
the 95% confidence interval. Source: Own illustration.

Figure 6.4: Electricity price (yearly average) versus subsidy costs [in Eur/MWh] under different gas price trend
scenarios for different technologies. Source: Own illustration.
 102 6. RES-E POLICY DESIGN IN PERSPECTIVE OF THE ENERGY TRANSITION

Figure 6.5: Fraction of electricity generated by renewables per tick, and under the three gas price scenarios

Gas price trend: Constant We observe that when the gas price is constant, there is little
impact on the electricity prices, despite an increase in the share of renewables. The impli-
cation of this observation is that the merit order effect is not apparent in the electricity
price trends of these runs. One reason for this is that, for these runs, the Netherlands
is modelled as an isolated system with no interconnections or storage. The absence of
spatial or temporal flexibility in the system from these sources means that there are very
few instances in the model when the demand can be met entirely by intermittent renew-
6 able electricity, and since CCGT is the only flexible technology in the model, all flexibil-
ity is derived from the CCGT. This in turn ensures that the marginal cost of those plants
remain a prominent component in the electricity prices, even as the share of VRE in-
creases.

Gas price trend: High In the gas price high scenarios, under all conditions, the subsidy
costs reduce with increasing electricity prices, for all technologies. Even in the gas price
constant scenarios, when the electricity prices remain almost constant, after the first 7-8
years, the per-unit subsidy costs for solar and wind offshore reduce with time, albeit at
different rates. This is shown in Figure 6.3. This reduction is almost entirely a function
of the reducing cost curves of the technologies, combined with their capacity factors.
The cost curves are presented in Figures A.2 and A.3. For wind onshore, the reduction in
fixed costs takes place at a far slower rate than for PV and wind offshore. Furthermore,
the assumed average capacity factor for wind onshore is half as much as that of wind
offshore. This means that, as the share of renewables increases, curtailment for wind
onshore is much higher than that of wind offshore. This curtailment manifests itself as
increasing per-unit subsidy costs for wind onshore technology in the gas trend constant
scenarios. Interestingly enough, even this effect is trumped for wind offshore by the
relatively high electricity price in the gas trend high scenarios.
The observations suggest that high electricity prices have a significant and direct role
to play in reducing subsidy costs, especially in scenarios where fixed costs of technolo-
gies do not reduce with time, or when the absence of transmission and storage lead to
high amounts of curtailment.
6.3. EU-ETS IN RELATION TO S UBSIDY C OSTS 103

Gas price trend: Low Conversely, observations from Figure 6.3 indicate that low elec-
tricity prices do not directly translate to high subsidy costs as the share of renewables in-
crease. Instead, in the gas trend low scenario, when the electricity prices are relatively
low, the subsidy costs depend far more on the rate of reduction in the costs of technol-
ogy, relative to electricity prices. Therefore, if the rate of reduction in technology costs is
lower than the rate of reduction in electricity prices, as in the case of wind onshore, then
subsidy costs increase with time. The increase in per-unit costs are also bolstered by
the curtailment of wind onshore production as the share of VRE increases. In contrast,
if the rate of reduction of fixed technology costs is greater than the rate of reduction in
electricity prices, subsidy costs reduce.

6.2.3. I NSIGHTS
There are large uncertainties associated with cost developments of renewable technolo-
gies, amounts of interconnection, advancements in storage technologies, and their im-
pact on electricity prices. Therefore, the results above are not predictive, but rather pro-
vide insights on the relationship between electricity prices and subsidy costs under var-
ious conditions.
Despite the fact that the merit-order effect was not apparent in the results, the ob-
servations still provide interesting insights on different factors impacting subsidy costs
under various electricity price scenarios. The most interesting outcome, to the author,
is the difference between the factors affecting subsidy costs under low or high electricity 6
prices. Crucially, over a long-term period of 40 years, high electricity prices [appr. 30-
50 Eur/MWh] have a more substantial role to play in reducing subsidy costs, than low
electricity prices have in increasing it. Another important insight is that decreasing av-
erage electricity prices is only a problem for the attractiveness to invest in RES-E if the
technology costs do not reduce at a rate faster than the electricity prices.

6.3. EU-ETS IN RELATION TO S UBSIDY C OSTS

In the previous section, model results demonstrated that when average electricity prices
are low, the impact on subsidy costs depends more on the cost curves of the technolo-
gies themselves. The objective of this section is to understand how (a) the EU-ETS and
RES support interact in the presence of an electricity market and (b) the EU-ETS could
influence the phasing out of subsidy costs for RES-E technologies, especially under low
average electricity prices.
The European Union’s Emissions Trading Scheme (EU-ETS) has been one of the pri-
mary policy instruments in the Union’s transition towards a low carbon economy. The
electricity sector has been the largest sector of the economy contributing to emissions
within the EU-ETS framework. While the share of EU emissions from the electricity sec-
tor has been gradually reducing since 1990, the total amount of electricity consumed
is set to increase, as other sectors such as transport and heating tend towards electrifi-
cation (European Commission and Direction générale de la mobilité et des transports,
2014). Therefore, policies in the electricity sector and the ETS framework are intricately
interrelated and the interactions between them have been justifiably analysed and de-
bated in the literature.
 104 6. RES-E POLICY DESIGN IN PERSPECTIVE OF THE ENERGY TRANSITION

This section comprises a brief discussion highlighting the major points in the afore-
mentioned debate. It further situates the results obtained in this dissertation within the
context of this debate. One notable discussion is the efficiency versus cost-effectiveness
debate, which is addressed in the first sub-section. There have been vigorous debates on
the economic logic behind simultaneously implementing renewable electricity support
schemes as well as the emission trading scheme. Another discussion, more directly rel-
evant to this thesis, is regarding the price effects of renewable support schemes on ETS
prices, and conversely, the effect of ETS prices on RES-E subsidy costs. This is addressed
in the second subsection.

6.3.1. E FFICIENCY AND COST- EFFECTIVENESS OF THE EU-ETS AND RES-

E?
If the sole objective of the many policy instruments was to reduce carbon emissions,
then, from the perspective of theoretical neoclassical economics, a combination of RES-
E subsidies and the EU-ETS system is inefficient. Central to the argument is the idea
that RES-E support policies do not have any impact on altering the CO2 emissions cap,
and hence do not result in emissions reductions. Further, such policies would con-
tribute to reductions in the European Emission Allowances (EUA) price, and shift emis-
sions to other sectors and other European countries instead (Böhringer and Rosendahl,
2010; del Río and Labandeira, 2009; Fischer and Preonas, 2010; Frondel et al., 2010, 2008;
6 Sijm, 2005). Such arguments have been demonstrated with analytical models, many of
which use partial equilibrium modelling (Amundsen and Mortensen, 2001; De Jonghe
et al., 2009; Fankhauser et al., 2010; Morthorst, 2003). Computational general equilib-
rium models have also demonstrated similar results (Paltsev et al., 2009; Morris, 2009;
Böhringer et al., 2009), as have simulations and numerical methods (Traber and Kem-
fert, 2009; De Jonghe et al., 2009; Linares et al., 2008).
Others, such as Lehmann and Gawel (2013), argue that such a thesis assumes that the
only market failure which exists is that the true cost of carbon emissions is not reflected
in the prices of commodities. They argue that this assumption does not account for sev-
eral market failures, and factors such as lock-in and path-dependencies. They contend
that due to such reasons, some of which are elaborated below, the ETS alone is not a
cost-effective solution in the long run. The following points have been extracted from
Lehmann and Gawel (2013); Sonnenschein (2016)

• Positive externalities which arise from knowledge spillovers are not captured in the
above argument. Several authors demonstrate that the presence of both policies
simultaneously is justified when learning spillovers are accounted for (van Ben-
them et al., 2008; Fischer and Newell, 2008; Kalkuhl et al., 2012; Kverndokk and
Rosendahl, 2007).
• The true social costs of non-renewable energy sources cannot be completely in-
ternalised. From an economic perspective the cost of marginal damages from one
tonne of CO2 is subject to considerable uncertainty. Estimates range from zero to
€300 per tonne of CO2 . Furthermore, the emissions gap is the result of political
negotiation rather than efficiency considerations.
• Participants in the EU ETS are subject to policy-induced uncertainties caused by
6.3. EU-ETS IN RELATION TO S UBSIDY C OSTS 105

price fluctuations and the political nature of setting of the cap. Such uncertainties
are detrimental to investments, which are characterised by high initial costs and
long life-times.
• The path-dependent nature of investments in the electricity sector mean that sub-
optimal decisions taken today could render infrastructure changes towards a low
emission system much, much more expensive in the future. This is known as the
carbon lock-in.
In the policy-setting arena of the European Commission, the debate seems to have
been decisively resolved. Both the EU-ETS and the renewable electricity support schemes
coexist and are periodically revised in order to account for developments in abatement
technologies and related costs.

6.3.2. I NTERACTING PRICE EFFECTS

Renewable electricity support schemes and the EU-ETS, as currently instituted in sev-
eral countries in Europe, are interdependent. On the one hand, subsidies which lead to
expansions in RES-E generation could reduce demand for allowances, and consequently
reduce the emissions allowance (EUA) price over the long term (Hintermann et al., 2016).
On the other hand, the emissions allowance (EUA) price or a carbon tax would increase
electricity prices, which consequently reduce subsidy costs. As Linares et al. (2008) ex-
plain, the final cost to society depends on the interactions between these instruments.
In this section, literature which provide quantitative estimates of such interactive effects 6
have been reviewed. From this brief literature review, it is found that most "interaction"
literature is primarily focussed on quantifying the extra costs due to the introduction
of RES-E support along with the EU-ETS. Fewer studies discuss results in a manner in
which the converse effect, that is, the effect of EU-ETS on reducing RES-E subsidy costs
are clear.
In Table 6.1, papers which quantify impacts of renewable electricity on EUA prices
using numerical models, simulation models or empirical methods are presented. Most
papers on the topic use numerical models to quantify effects. The results of these nu-
merical models are also not immediately comparable, as some of them present a drop in
EUA prices, while others present a reduction in the demand for EUA allowances, and in
many cases all assumptions are not clearly indicated. The scopes of the papers also differ
in terms of location and time period. One interesting insight is the substantial difference
between the estimations of numerical models and that of econometric analysis. While
the numerical models present results ranging from −1.32 Eur/tCO2 to −100 Eur/tCO2 ,
or at least a −22% change in the EUA price, the econometric analysis attributes a much
more modest −0.11 to −0.14% drop in ETS prices to the presence of renewable electric-
ity. Interestingly, the paper using econometrics Koch et al. (2014) finds that dummy vari-
ables employed to simulate selected policy events, including the ban on using certain
CERs and the announcement of an energy efficiency directive amongst others explain
44% of the price variation.
In Table 6.2, papers which quantify the impact of EUA prices on reducing subsidy
costs are presented. Both papers presented here are based on numerical modelling. To
the knowledge of this author there are no papers analysing this effect based solely on
empirical data. In the first paper Linares et al. (2008), a scenario which has an increasing
 106 6. RES-E POLICY DESIGN IN PERSPECTIVE OF THE ENERGY TRANSITION

Table 6.1: Influence of renewable electricity on carbon price

Study Methodology Regional and Tem- Objective Key Findings Reference case
poral Scope
Linares Analytical and sim- Spain, 2004 to 2020 Impact of renew- Reduction in EUA No RES-E Sup-
et al. (2008) ulation research able support pol- price per ton of CO2 : port
icy (Tradable Green 2008 to 2012: 1.32
Certificate) on EUA Eur/tCO2 (22%)
price 2013 to 2017: 5.28
Eur/tCO2 (42%)
2018 to 2020: 7.34
Eur/tCO2 (33%)
Weigt et al. Unit commitment Germany, 2006 to Reduction in de- Reduction by 33-57 No RE Policy
(2012) model 2010 mand for EUA Mton (10-16%) between
allowances due to 2006 and 2010
RE deployment in
German electricity
sector
Van den Partial equilibrium 2 EU Member The impact of RES- Reduction in EUA price No RE Policy
Bergh et al. model States in South- E deployment on per ton of CO2 : 15
(2013) ern and Western the EUA price and EUR/tCO2 in 2007, 46
Europe, 2007 to the CO2emissions EUR/tCO2 in 2008, 100
2010 in the European EUR/tCO2 in 2010
electricity sector
Koch et al. Econometric ex- EU, 2008-2013 To what extent can 0.11% to 0.14% (5% sig- No RES-E sup-
(2014) post analysis EUA price drop nificance level) port
be attributed to
renewable policies

6 quota for renewables is compared with a similar scenario which has the EU ETS system
added to the RES quota. Here, the reduction in subsidy costs ranges from 0 to appr. 12%.
However, since assumptions regarding the progression of renewables quota and of the
ETS caps through time are not clearly specified, it is difficult to assess the impact of ETS
prices on the subsidy costs as the quota increases. In the second paper De Jonghe et al.
(2009) however, the lowering of subsidy costs is clearly indicated at various assumptions
of RES-E quota as well as ETS allowance caps. All in all, it is evident from the analysis that
as the share of renewables increase, the EUA price itself, and the impact of EUA prices on
subsidy costs reduce. This is clearly indicated in Figure 6.6, from De Jonghe et al. (2009).
Of course, it is possible that if the cap were set to be increasingly stringent to ensure a
high CO2 price, the price could have an impact on the subsidy costs. However, such an
approach would defeat the purpose of a quantity-based instrument, while employing a
CO2 tax could achieve the same goal in a direct manner.

6.3.3. I MPLICATIONS FOR MODEL RESULTS

In the previous section, model results demonstrated that when average electricity prices
are low, the impact on subsidy costs depends more on the cost curves of the technolo-
gies themselves. The objective of this section was to understand how the EU-ETS could
influence the phasing out of subsidy costs for RES-E technologies, especially under low
average electricity prices. Findings from the literature reviewed indicate that, while EUA
prices could indeed reduce subsidy costs in the short term (0-5 years), they are far less
likely to help phase out RES-E subsidies in the longer term (20-30 years) as the share of
RES-E significantly increases. Of course, this is true if the electricity sector still forms
a significant share of the sectors included in the ETS in the long-term. If other carbon
intensive sectors drive the marginal CO2 price, this implication might not hold.
6.4. I MPACT OF FLEXIBILITY ON THE ECONOMIC VALUE OF VRE 107

Table 6.2: Influence of carbon price on renewable electricity subsidy

Study Methodology Regional and Tem- Objective Key Findings Reference case
poral Scope
Linares Analytical and sim- Spain, 2020 Impact of EUA Subsidy costs reduce No EU ETS
et al. (2008) ulation research price for emissions from 56.59 Eur/MWh
cap of 80Mt on VRE to 50.82 Eur/MWh
subsidy (by 10%) when ETS is
introduced
De Jonghe Simulation model France, Germany, Impact of CO2 For a relatively high No EU ETS
et al. (2009) and Benelux quota on RES-E quota on renewables
subsidy cost (>40%), the certificate
price is only depen-
dent on this quota, and
is independent from
the restriction on CO2
emissions.

6
Figure 6.6: (a) Tradable Geeen Certificate price as a proxy for subsidy cost and (b) allowance price for combina-
tions of TGC and CO2 allowances. Note that the X-axis has a reversed direction in panel (b). Source: De Jonghe
et al. (2009)

Nonetheless, even if the quota on CO2 reduction were to be increased as the share
of RES-E increased, the impact of a marginal increase in the quota on EUA prices and
RES-E subsidy costs would be far less at higher RES-E shares than at lower shares. This
is because carbon-intensive generation technologies themselves will set the marginal
price less often in a system with greater RES-E. If the main objective were to phase out
subsidy costs based on spot market prices, a tax on CO2 would be a more effective and
straightforward means to achieve it at high shares of RES-E in the generation mix.

6.4. I MPACT OF FLEXIBILITY ON THE ECONOMIC VALUE OF VRE

In the previous chapter it became apparent that both interconnection and storage, or
the lack of it, have significant implications on the per-unit costs of intermittent RES-E
generated. In particular, the lack of storage, or temporal flexibility in the system, mani-
fested itself as a large amount of curtailment of intermittent RES-E, which led to greater
proportions of the costs having to be recovered in shorter periods, thereby increasing
the per-unit costs of intermittent electricity generated. In this section, we juxtapose this
outcome with comparable literature dealing with the impacts of flexibility on the subsidy
costs of renewables.
Several works on flexibility focus on system cost minimisation under high penetra-
tion of renewables or a completely decarbonised system, with different options for flexi-
 108 6. RES-E POLICY DESIGN IN PERSPECTIVE OF THE ENERGY TRANSITION

bility (Brouwer et al., 2016; Budischak et al., 2013; De Jonghe et al., 2011; Jacobson et al.,
2015; Steinke et al., 2013). Other papers discuss the "integration costs of RES-E", where
they estimate, for instance, “an increase in power system operating costs” due to the
presence of VRE (Milligan and Kirby, 2009). While these are important questions in
themselves, they answer questions related to the ’ideal’ costs, or describe the changes
possible to system costs when renewables are integrated into the system and more flex-
ibility is required. They do not, however, provide insights on the effects of flexibility on
the economic value of variable renewables electricity itself. In the researcher’s effort to
assess the possibilities of phasing out of RES-E subsidy, this aspect is key, as it evaluates
the role of flexibility in the economic viability of VRE.
In a landmark paper, Hirth (2013), defined and quantified the concept of "market
value of VRE". Central to this concept is the idea that the availability of the primary
energy source, wind or solar irradiation, fluctuates over time in a manner that has limited
predictability. This, as Hirth describes, "affects the economics of power generation either
by increasing costs, for instance of balancing, or more significantly, reducing the value of
the power generated on the spot market." Several papers have quantified this reduction
in market value of renewables with an increase in their share. To the knowledge of the
author, only two papers however, clearly quantify the impact of flexibility stemming this
reduction of the economic value of intermittent renewable electricity. Their results are
presented below.
The first paper is a work from the Lawrence Berkeley National Laboratory (Mills and
6 Wiser, 2015), which evaluates options to stem the reduction in the value of wind and PV
using a long-run equilibrium investment and dispatch model and simulating a region
loosely based on California in 2030. Interestingly, for this specific case, the authors find
that the largest increase in the value of wind comes from geographical diversity, while
the largest increase in the value of solar PV comes from low cost bulk storage.
The second paper, again by Hirth (2016b) comprises both empirical and numerical
analyses, to quantify how the flexibility provided by hydropower mitigates the reduc-
tion in market value of wind energy in Sweden and Germany. He finds that, as the share
of wind increased from 0% to 30%, hydropower mitigates the value drop by a third. As
a consequence, a unit of wind energy (in MWh) is 12%-29% more valuable in Sweden
than in Germany. The use of low wind speed turbines, carbon pricing, and greater hy-
dropower capacity are found to be able to accentuate the value added by flexible hy-
dropower. The author also finds that the positive impact of hydro-flexibility on the value
of wind tapers off as wind penetration increases beyond 20%.
It follows from the above literature that the specific features of a system play a sig-
nificant role in determining which source of flexibility might work the most to increase
the value of VRE. In conclusion, as the share of VRE increases in any electricity system,
the decision to invest in a VRE plant will depend on benefits accrued not just from the
primary resource itself, but also from the types and sources of flexibility in the system.

6.5. D ISCUSSION
In this section, insights from the three strands of literature discussed in this chapter are
brought together to expound on factors that influence RES-E subsidy costs. The time
scales that are analysed in this dissertation are spread over the long term (appr. 30 years),
6.5. D ISCUSSION 109

whereas a lot of the literature is focussed on past, current, or shorter future time scales.
The discussion therefore culminates in a comparison between factors which could influ-
ence subsidy costs over the short term versus the long term, when the share of RES-E is
expected to be substantially higher. An indicative representation is shown in Figure 6.7.
Costs have reduced dramatically over the last decade and are expected to reduce in
the near future, as explained in section 6.2. Learning curves or experience curves are
commonly used to forecast cost projections for new technologies. This is based on the
notion that the more times a task is performed, the less investment in effort is required
for each subsequent iteration2 . A common formulation of this effect is that unit costs de-
crease by a constant percentage for each doubling of experience or production. There-
fore, as technologies mature, cost reductions are also expected to decrease with every
extra unit of production (McDonald and Schrattenholzer, 2001). Based on this theory it
is reasonable to predict that further cost reductions will be far lesser in about 30 years
from now for technologies such as PV and wind.
Given the merit order effect, revenue from electricity spot prices would be lower in a
system with more RES-E generation than less. In a system with substantially higher RES-
E generation, the time during which coal/gas would set the marginal price would reduce,
as compared to a system with lower shares of RES-E generation. Therefore, a unit change
in fuel prices such as those of coal or gas, will have a lower impact on the revenue from
the electricity spot market for RES-E technologies, than it does today. This is true under
the assumption that coal and gas will still be part of the merit-order that far into the
future. A similar argument holds for the role of carbon prices in reducing subsidy costs 6
in the future. As the share of RES-E increases, and carbon-based generation decreases,
the influence of carbon prices on the spot market decreases.
Interconnection, as we know, provides spatial flexibility, while storage provides tem-
poral flexibility. At high shares of RES-E, weather conditions and consequently primary
energy supply across different locations should be diverse enough to supply demand.
In large parts of Europe, where seasonal weather and demand changes are drastic and
simultaneous, for instance, the supply of solar energy reduces substantially during the
winter, while demand remains the same or increases. Therefore, while interconnection
does help provide flexibility at low shares of RES-E and in the near future, it is unlikely
that it will be sufficient to increase the market value of VRE at high shares (>60%) of RES-
E in the long term. This was also a key finding in section 5.3.2, where we found that
interconnection plays a limited role in lending flexibility to the system, whereas stor-
age will play an increasingly important role as RES-E shares increase. These findings are
pictorially represented in Figure 6.7.

2 The concept is commonly referred to as the learning curve effect.

 110 6. RES-E POLICY DESIGN IN PERSPECTIVE OF THE ENERGY TRANSITION

Figure 6.7: Impacts of various factors on subsidy costs in the short and long terms
6.6. C ONCLUSION 111

6.6. C ONCLUSION
The main research objective for this chapter was to identify the conditions under which
RES-E subsidies can be phased out. Given that some European countries are considering
a system in which subsidies will no longer be offered to RES-E technologies, this is a top-
ical question. This is also a vital topic in the context of the objective of this dissertation,
which is to help understand design of RES-E subsidies in the European institutional set-
ting. The objective was accomplished primarily by a literature review which examined
three major aspects of the energy transition in Europe: the cost and revenue drivers of
RES-E technologies, the EU ETS, and finally, the role of flexibility. In the course of this
literature review, work from the earlier chapters on designing RES-E subsidies was put in
perspective of the larger energy transition in Europe.
We find that at relatively high average electricity prices, the impact of electricity prices
on subsidy costs is direct and clear; as prices increase, subsidy costs decrease. How-
ever, as electricity prices on the spot market reduce, the change in subsidy cost trends
depends far more on the cost curves of each RES-E technology. With respect to the EU-
ETS, it is clear that, even if EUA prices play a significant role on the electricity prices
today, their influence on the electricity prices, and the RES-E subsidy costs will reduce
as the share of RES-E generation increases with time. From Sections 6.4 and 5.3.2, it is
clear that at high shares of RES-E, interconnection will play a limited role in providing
flexibility while storage (long-term, seasonal) will play a pivotal role. Consequently, the
costs of subsidy for RES-E at high shares are substantially dependant on the presence of
storage in the system. The chapter concludes with a discussion on which factors impact
6
subsidy costs the most in the short term, versus the long term, under certain institutional
settings.
The analysis is largely qualitative, and opens up several avenues for future research.
One interesting path is to analyze the costs at which auctions for long-term contracts are
better than the spot market at creating competition in RES-E and low-carbon flexibility
technologies. Another path for future research could be to use a high-resolution model
to compute the point at which the extra costs of subsidy due to spillage could be higher
than the cost of a certain amount of (large-scale)storage.
 112 R EFERENCES

R EFERENCES
O. Wind, UK to Set Sail to Subsidy-Free Offshore Wind Farms, (2017).

IRENA, Solar PV in Africa: Costs and Markets, Tech. Rep. (2016).

F. Sensfuss, M. Ragwitz, and M. Genoese, The merit-order effect: A detailed analysis of

the price effect of renewable electricity generation on spot market prices in Germany,
Energy Policy 36, 3086 (2008).

L. Hirth, The market value of variable renewables: The effect of solar wind power variabil-
ity on their relative price, Energy Economics 38, 218 (2013).

REmap: Roadmap for a Renewable Energy Future., Tech. Rep. (IRENA, Abu Dhabi, 2016).

B. N. E. Finance, Wind Power Blows Through Nuclear, Coal as Costs Drop at Sea,
Bloomberg.com (2017a).

B. N. E. Finance, New Energy Outlook 2017, Tech. Rep. (Bloomberg Finance L.P., 2017).

R. Wiser, M. Bolinger, G. Barbose, N. Darghouth, B. Hoen, A. Mills, J. Rand, D. Millstein,

K. Porter, and R. Widiss, 2015 Wind Technologies Market Report, Tech. Rep. LBNL-
6 1005951 (Lawrence Berkeley National Laboratory, 2016).

M. O. for Energy, Quarterly report on European electricity markets, Tech. Rep.

(Directorate-General for Energy, European Commission, Brussels, 2016).

J. Cludius, H. Hermann, F. C. Matthes, and V. Graichen, The merit order effect of wind
and photovoltaic electricity generation in Germany 2008–2016: Estimation and distri-
butional implications, Energy Economics 44, 302 (2014).

N. Ederer, The market value and impact of offshore wind on the electricity spot market:
Evidence from Germany, Applied Energy 154, 805 (2015).

L. Gelabert, X. Labandeira, and P. Linares, An ex-post analysis of the effect of renewables

and cogeneration on Spanish electricity prices, Energy Economics 33, S59 (2011).

A. O’Mahoney and E. Denny, Electricity prices and generator behaviour in gross pool elec-
tricity markets, Energy Policy 63, 628 (2013).

T. Traber and C. Kemfert, Impacts of the German Support for Renewable Energy on Elec-
tricity Prices, Emissions, and Firms, The Energy Journal 30, 155 (2009).

T. Traber and C. Kemfert, Gone with the wind? — Electricity market prices and incen-
tives to invest in thermal power plants under increasing wind energy supply, Energy
Economics 33, 249 (2011).

H. Weigt, Germany’s wind energy: The potential for fossil capacity replacement and cost
saving, Applied Energy 86, 1857 (2009).
R EFERENCES 113

K. Würzburg, X. Labandeira, and P. Linares, Renewable generation and electricity prices:

Taking stock and new evidence for Germany and Austria, Energy Economics Supple-
ment Issue: Fifth Atlantic Workshop in Energy and Environmental Economics, 40,
Supplement 1, S159 (2013).

A. Bublitz, D. Keles, and W. Fichtner, An analysis of the decline of electricity spot prices in
Europe: Who is to blame? Energy Policy 107, 323 (2017).

J. Dehler, F. Zimmermann, D. Keles, and W. Fichtner, Der Einfluss der Nachbarländer auf
den Schweizer Strommarkt, (Graz, Austria, 2016).

S. Thoenes, Understanding the Determinants of Electricity Prices and the Impact

of the German Nuclear Moratorium in 2011, The Energy Journal 35 (2014),
10.5547/01956574.35.4.3.

L. Hirth, What Caused the Drop in European Electricity Prices?, SSRN Scholarly Paper ID
2874841 (Social Science Research Network, Rochester, NY, 2016).

R. Hamming, Numerical methods for scientists and engineers (McGraw-Hill, New York,
1962).

European Commission and Direction générale de la mobilité et des transports, EU en-

ergy, transport and GHG emmissions trends to 2030: reference scenario 2013 (Office 6
for official publications of the european communities, Luxembourg, 2014) oCLC:
889309007.

C. Böhringer and K. Rosendahl, Green promotes the dirtiest: On the interaction between
black and green quotas in energy markets, Journal of Regulatory Economics 37, 316
(2010).

P. del Río and X. Labandeira, Barriers to the introduction of market-based instruments in

climate policies: an integrated theoretical framework, Environmental Economics and
Policy Studies 10, 41 (2009).

C. Fischer and L. Preonas, Combining Policies for Renewable Energy: Is the Whole Less
than the Sum of its Parts? SSRN Electronic Journal (2010), 10.2139/ssrn.1569634.

M. Frondel, N. Ritter, C. M. Schmidt, and C. Vance, Economic impacts from the promo-
tion of renewable energy technologies: The German experience, Energy Policy 38, 4048
(2010).

M. Frondel, N. Ritter, and C. M. Schmidt, Germany’s solar cell promotion: Dark clouds
on the horizon, Energy Policy 36, 4198 (2008).

J. Sijm, The interaction between the EU emissions trading scheme and national energy
policies, Climate Policy 5, 79 (2005).

E. S. Amundsen and J. B. Mortensen, The Danish Green Certificate System: some simple
analytical results, Energy Economics 23, 489 (2001).
 114 R EFERENCES

C. De Jonghe, E. Delarue, R. Belmans, and W. D’haeseleer, Interactions between measures

for the support of electricity from renewable energy sources and CO2 mitigation, Energy
Policy 37, 4743 (2009).

S. Fankhauser, C. Hepburn, and J. Park, COMBINING MULTIPLE CLIMATE POLICY IN-

STRUMENTS: HOW NOT TO DO IT, Climate Change Economics 01, 209 (2010).

P. Morthorst, National environmental targets and international emission reduction in-

struments, Energy Policy 31, 73 (2003).

S. Paltsev, J. Reilly, H. Jacoby, and J. Morris, The cost of climate policy in the United States,
Tech. Rep. 173 (2009).

J. F. Morris, Combining a renewable portfolio standard with a cap-and-trade policy :

a general equilibrium analysis, Ph.D. thesis, Massachusetts Institute of Technology,
Cambridge, Massachusetts (2009).

C. Böhringer, A. Löschel, U. Moslener, and T. F. Rutherford, EU climate policy up to 2020:

An economic impact assessment, Energy Economics 31, S295 (2009).

P. Linares, F. J. Santos, and M. Ventosa, Coordination of carbon reduction and renewable

energy support policies, Climate Policy 8, 377 (2008).
6
P. Lehmann and E. Gawel, Why should support schemes for renewable electricity com-
plement the EU emissions trading scheme? Energy Policy Special Section: Transition
Pathways to a Low Carbon Economy, 52, 597 (2013).

J. Sonnenschein, Conditions for the Cost Effective Combination of Emissions Trading and
Renewable Energy Support Policies, Energy Procedia 88, 133 (2016).

A. van Benthem, K. Gillingham, and J. Sweeney, Learning-by-Doing and the Optimal

Solar Policy in California, The Energy Journal Volume 29, 131 (2008).

C. Fischer and R. G. Newell, Environmental and technology policies for climate mitiga-
tion, Journal of Environmental Economics and Management 55, 142 (2008).

M. Kalkuhl, O. Edenhofer, and K. Lessmann, Learning or lock-in: Optimal technology

policies to support mitigation, Resource and Energy Economics 34, 1 (2012).

S. Kverndokk and K. E. Rosendahl, Climate policies and learning by doing: Impacts and
timing of technology subsidies, Resource and Energy Economics 29, 58 (2007).

B. Hintermann, S. Peterson, and W. Rickels, Price and Market Behavior in Phase II of

the EU ETS: A Review of the Literature: Appendix Table 1, Review of Environmental
Economics and Policy 10, 108 (2016).

H. Weigt, E. Delarue, and D. Ellerman, CO2 Abatement from Renewable Energy Injections
in the German Electricity Sector: Does a CO2 Price Help?, Tech. Rep. CEEPR 2012-003
(MIT Center for Energy and Environmental Policy Research, 2012).
R EFERENCES 115

K. Van den Bergh, E. Delarue, and W. D’haeseleer, Impact of renewables deployment on

the CO2 price and the CO2 emissions in the European electricity sector, Energy Policy
63, 1021 (2013).

N. Koch, S. Fuss, G. Grosjean, and O. Edenhofer, Causes of the EU ETS price drop: Reces-
sion, CDM, renewable policies or a bit of everything?—New evidence, Energy Policy 73,
676 (2014).

A. S. Brouwer, M. van den Broek, W. Zappa, W. C. Turkenburg, and A. Faaij, Least-cost

options for integrating intermittent renewables in low-carbon power systems, Applied
Energy 161, 48 (2016).

C. Budischak, D. Sewell, H. Thomson, L. Mach, D. E. Veron, and W. Kempton, Cost-

minimized combinations of wind power, solar power and electrochemical storage, pow-
ering the grid up to 99.9% of the time, Journal of Power Sources 225, 60 (2013).

C. De Jonghe, E. Delarue, R. Belmans, and W. D’haeseleer, Determining optimal electric-

ity technology mix with high level of wind power penetration, Applied Energy 88, 2231
(2011).

M. Z. Jacobson, M. A. Delucchi, M. A. Cameron, and B. A. Frew, Low-cost solution to the

grid reliability problem with 100% penetration of intermittent wind, water, and solar
for all purposes, Proceedings of the National Academy of Sciences 112, 15060 (2015). 6
F. Steinke, P. Wolfrum, and C. Hoffmann, Grid vs. storage in a 100% renewable Europe,
Renewable Energy 50, 826 (2013).

M. R. Milligan and B. Kirby, Calculating wind integration costs: Separating wind energy
value from integration cost impacts, Tech. Rep. (National Renewable Energy Labora-
tory Colorado, 2009).

A. D. Mills and R. H. Wiser, Strategies to mitigate declines in the economic value of wind
and solar at high penetration in California, Applied Energy 147, 269 (2015).

L. Hirth, The benefits of flexibility: The value of wind energy with hydropower, Applied
Energy 181, 210 (2016b).

A. McDonald and L. Schrattenholzer, Learning rates for energy technologies, Energy Pol-
icy 29, 255 (2001).
 7
C ONCLUSIONS AND REFLECTIONS
In this chapter, key findings of the thesis are summarised, and the findings are reflected
upon. In Section 7.1, answers to research questions are presented. The following section
7.2 presents a reflection on the usefulness and limitations of the approach, the model,
and the results presented. Section 7.3 then describes avenues for future work.

7.1. A NSWERS TO R ESEARCH QUESTIONS

As the energy transition in Europe gathers steam, policy makers grapple with questions
regarding the most appropriate designs of the institutions that drive the transition. Pol-
icy makers confront inter-alia, multiple objectives, multiple stakeholders, dynamically
changing costs, and advancements in technology. As the share of renewable electricity
from solar PV and wind increases in the system, impacts such as the merit-order effect,
reducing market value of intermittent generations, decrease the wholesale market-based
revenues for power plants. Such effects question the design of current institutions such
as the electricity spot market as well as renewable support schemes over the long term.
In this context, the main research question posed in Chapter 1 of the thesis is the follow-
ing:

How do national renewable electricity support schemes interact with the electricity
market over the long term (20-30 years) as the European Union transitions to a decar-
bonized energy system?

Using an approach based on theoretical foundations of institutional analysis, and

design theory, in this dissertation RES-E support schemes are broken down into their
design elements from a welfare economics perspective. A modelling framework is then
introduced, using the modelling paradigm of agent-based modelling, by which RES-E
schemes are modelled in EMLab Generation. EMLab Generation mainly comprises an
electricity market clearing algorithm and an endogenous investment algorithm, and al-
lows the researcher to study long-term dynamics in the electricity sector. Most litera-
ture tackling similar questions and using quantitative models, employ optimization tools

117
 118 7. C ONCLUSIONS AND REFLECTIONS

which assume perfect foresight or known probabilities of imperfect foresight 1 , or oper-

ate under the assumption of a long-term equilibrium. In contrast, this approach incor-
porates true uncertainty to model regulatory and investment decisions. This way the
long-term model is a step closer to reality.
The model shows that the way electricity prices are taken into account while de-
signing subsidies makes a substantial difference to welfare distributions. When multiple
countries are considered, the spreading of the merit order effect from a larger country to
a smaller one, would cause the smaller country’s subsidy costs to increase. This indicates
that even if RES-E support can be designed to be "market-based" at the operational level,
designing them to be "market based" at the investment level is far more complicated. It
indicates that RES-E targets of countries should take into account the targets of neigh-
bouring countries, the amount of flexibility in terms of interconnection and storage in
the sytem. Finally, the impact of flexibility on costs of RES-E subsidies is also discussed.
These issues have been explored through the various research questions that follow.

How can policy design options for RES-E support in Europe be systematically and
comprehensively explored and modelled?
Using a combination of design theory, institutional analysis, and agent-based mod-
elling (ABM), we provide a method to systematically explore policy design options for
RES-E support in Europe. This is done firstly by identification of the design elements2 of
a policy or set of policies, and secondly by evaluating the impact of each design element
on the socio-technical system using an agent-based model.
Given a certain frame of analysis, we propose that it is theoretically possible to iden-
tify the complete policy design space. Crucially, this aspect potentially opens up to the
7 policy analyst new avenues for intervention, and allows her explore, given a range of un-
certainties, which element(s) of intervention is(are) the most vital to achieve the goals of
the community. The applicability of the approach is demonstrated by representing and
differentiating between six renewable electricity support schemes from Western Europe
in terms of the design elements. The applicability of the modelling framework using
ABM, and consequently of the Design Element Approach, is demonstrated by evaluating
the long-term, dynamic impact of three design elements: price warranty versus quan-
tity warranty, technology neutrality versus specificity, and accounting for the electricity
market price ex-ante versus ex-post on the Dutch electricity sector.
It is important to note here that claims of completeness of the design space come
with limitations. For instance, if the energy producer agent were to assume multiple
identities, such as being politically active and strongly pushing for local autarky, the de-
sign elements would be different. The design framework published here therefore per-
tains mainly to an analysis which lies within the scope of welfare economics, although
founded firmly within an institutional framework and empirical experience. Other lim-
itations of the approach include its computationally intensive nature, and the need to
prudently select the most important design elements necessary for the analysis.

1 As in stochastic optimisation
2 Design elements of renewable electricity support policies are defined as a closed set of components which
a renewable electricity policy is comprised of, such that each component now forms the smallest level of
analysis.
7.1. A NSWERS TO R ESEARCH QUESTIONS 119

How do RES-E support policy design elements interact with a single isolated elec-
tricity market and what social welfare implications do they actualise?
We employ this design element view in combination with agent based modelling to
quantitatively assess impacts of individual design elements on the socio-technical sys-
tem. The results demonstrate that design elements, irrespective of the RES-E policy to
which they belong, have significant impacts on the energy system and on welfare dis-
tribution, and therefore that the approach is a useful one. The agent-based modelling
framework enables modelling of bounded rationalities in investment decisions, allow-
ing the modeller to incorporate real-world uncertainties in agents’ behaviour. An impor-
tant uncertainty in the real world is that of long-term electricity price development. The
model interestingly demonstrates that accounting for future electricity prices ex-ante
in the subsidy calculation may reduce the overall cost of subsidy by about 15%, since
the actors are likely to overestimate the future electricity price. This is a consequence
of underestimating the impact of the merit order effect on expected electricity prices
over the long-term. Other significant results are that technology specificity could reduce
the cost of subsidy by upto 60%. Results regarding the design element, quantity vs price
warranty corroborate established literature: quantity warranty helps achieve targets bet-
ter. The design element configuration that leads to the highest increase in social welfare
is the combination of quantity-warranty, ex-ante accounting for electricity prices, and
technology-specificity.
With regard to policy implications, the State Aid Guidelines of the European Com-
mission promote competitive bidding to incentivize investment, while largely support-
ing technology neutrality. At the outset, our results corroborate with the choice of com-
petitive bidding. They however indicate that the feature technology specificity has a sig- 7
nificant implication on welfare impacts, subject to the assumption of regulator’s knowl-
edge of real costs being the same as the energy producer. Differences in such features
of RES-E policy between member states could lead to unintended cross border effects.
The design element method has the potential to provide insight into which aspects of
the policies need to be co-ordinated at the European level.

How do RES-E support policy design elements interact with an interconnected,

congested electricity market? The primary objective of chapter 5 is to investigate
cross-border effects due to national RES-E support schemes, operating in an interna-
tional electricity market. Using EMLab Generation, regions based on Germany and the
Netherlands were simulated, to test the effects of different support scheme designs in
each country, and the common electricity market, on distributional implications under
different scenarios of interconnector capacities.
As theory would predict, it is found that subsidy costs increase in the smaller country,
as electricity prices decrease, as the merit order effect spreads from the larger neighbour-
ing country. However, the increase in subsidy costs remains lower than the reduction in
costs of electricity, corroborating earlier, single country analyses on the topic. Therefore,
total costs in both countries reduce as interconnection increases, and the invisible hand
works its magic. As the share of RES-E increases, interconnection has limited impact on
reducing spillage of RES-E while storage becomes increasingly important.
 120 7. C ONCLUSIONS AND REFLECTIONS

The results show that setting targets for RES-E, should not only take into account
the targets of the neighbouring countries, but also the capacity of storage that exists in
the system. The results also raise interesting questions about how subsidy costs should
be allocated between countries over the long-term, especially when subsidy costs are so
intricately connected with the electricity markets they operate in.

How could major developments in the energy transition such as RES-E technology
cost trends, the EU ETS, and flexibility influence RES-E support?
The objective was accomplished primarily by a literature review which examined
three major aspects of the energy transition in Europe: the cost and revenue drivers of
RES-E technologies, the EU ETS, and finally, the role of flexibility. In the course of this
literature review, work from the earlier chapters on designing RES-E subsidies was put in
perspective of the larger energy transition in Europe.
We find that the impact of high average electricity prices on subsidy costs is direct
and clear; i.e., as prices increase subsidy costs decrease. However, as electricity prices
on the spot market reduce on average, the change in subsidy cost trends depend far
more on the cost curves of each RES-E technology. With respect to the EU-ETS, it is
clear that, even if EUA prices play a significant role on the electricity prices today, their
influence on the electricity prices, and the RES-E subsidy costs will reduce as the share
of RES-E generation increases with time. From Sections 6.4 and 5.3.2, it is evident that
at high shares of RES-E, interconnection will play a limited role in providing flexibility
while storage (long-term, seasonal) will play a pivotal role. Consequently the costs of
subsidy for RES-E at high shares are substantially dependent on the presence of storage
in the system.
7
7.2. R EFLECTIONS
As computer simulation methods are gaining more and more prominence across disci-
plines, ideas about their trustworthiness for generating new knowledge are being dis-
cussed in works on philosophy of science (Gilbert and Troitzsch, 2005; Parker, 2008;
Winsberg, 1999, 2001; Epstein, 1999). The reader’s confidence in a computer simula-
tion is instilled by evaluating the claim that a simulation is indeed fit for its purpose. The
science that deals with such evaluations is called the Epistemology of Computer Simula-
tions (EOCS) Winsberg (2015). The idea of external validation of a simulation is explored
in the subsequent paragraphs. The stage is set by a note on the role of simulations in sci-
ence. The uncertainties and assumptions involved in the model (EMLab) and how they
relate to the real world are explored in the following subsection. Finally, the usefulness
of the framework, model, and results are reflected upon.

7.2.1. R OLE OF SIMULATION IN SCIENCE

The study of epistemology of (computer) simulations concerns itself with questions re-
lated to the nature and purpose of simulations. How are experiments different from sim-
ulation models? How should their reliability and validity be evaluated? What is the role
of simplifying assumptions in simulations? In relation to evaluating a simulation against
its fitness of purpose, a widely popular approach is to verify and validate. While it is a
7.2. R EFLECTIONS 121

pragmatically useful approach, certain philosophers question whether verification and

validation are indeed entirely independent of each other, and whether such a distinction
could misrepresent the messiness of simulationists’ practice.
Winsberg, who has written extensively on the topic, suggests that the idea that sim-
ulation is a new mode of science, methodologically lying in between theory and exper-
iment, is a good place to start (Winsberg, 2010). Deb Dowling, a sociologist of science,
characterises simulation as being similar to theory because it involves "manipulating
equations" and "developing ideas". However, she also suggests that it is like an experi-
ment given that it involves "fiddling with machines", "trying things out", and "watching
to see what happens" (Dowling, 1999). The idea that there are more activities between
experiment and theory has been introduced several decades ago in 1983 by (Hacking,
1983). Hacking repudiates the traditional dichotomy of theory and experiment. He urges
that it is replaced by the following characterizations: speculation, calculation, and exper-
iment. By speculation he refers to the activity of laying out basic theoretical principles
such as Maxwell’s equations, or Newton’s laws of motion to name a few. The activity
he calls ’calculation’ refers to making theoretical principles apply to the local, concrete
systems that make up the real world. Winsberg suggests that simulation is most similar
to the second activity, one which often takes the researcher beyond merely theoretical
principles themselves.
Describing theories as Hacking does is a refreshing point of view to any researcher
who has had to explain the value of agent-based models to firm believers of the neoclas-
sical school of economic theory. It is a reminder that the idea of a long-term equilibrium
in energy markets is merely speculation, and never a truth. One that must be scrutinised
time and again, in order to either confirm or deny its credibility. To the author of this
work, agent based models are a powerful method for performing such evaluations, and
in creating and testing new speculations.
7
Even so, how does one evaluate the credibility or validity of such a method? Wins-
berg is of the opinion that the credibility of simulations "comes not only from the cre-
dentials supplied to it by the governing theory, but also from the antecedentally estab-
lished credentials of the model-building techniques used to make it." This is because if
it is successful in making a variety of predictions or interventions, it means that there is
something about it that is "in some way latching on to the real structure of the world".
Furthermore, Parker suggests that for model evaluation to have rigour and structure, it is
necessary that there is explicit argumentation about the capacity of a model to provide
evidence for the concerned scientific hypotheses, and not simply superficial compar-
isons of simulation results and empirical data.
Following from this discussion, in the sections below we explore how the model pre-
sented relates to the real world, how the assumptions made in the model affect results
and under what conditions. We then discuss the usefulness of the scientific framework
used, the model itself, and the results.

7.2.2. H OW DOES THE MODEL RELATE TO THE REAL WORLD ?

One of the challenges that are commonly faced by researchers of long term electricity
market models is ensuring a "suitable" tradeoff between fine temporal resolution and
computational tractability. For EMLab, although the input is hourly data, during the
 122 7. C ONCLUSIONS AND REFLECTIONS

market clearing process, the data is aggregated to 20 segments on both the demand and
supply side. At high shares of intermittent RES-E generation, it becomes is important to
represent the market clearing with a fine resolution, as otherwise the impact of renew-
ables on the market price will be under-represented.
Interconnections are represented in a limted manner in the model. In Chapter 4,
The Netherlands is represented as an isolated system, and in Chapter 5, only the inter-
connection between Germany and the Netherlands is represented. In reality the trans-
mission network is far more complex. As spatial heterogeneity in supply and demand
contributes to balancing, its absence in Chapter 4 leads to an inflated presence of CCGT
in the merit-order, increasing average electricity prices. Similarly in Chapter 5, if the in-
terconnections had been represented with greater accuracy, while taking into account
the electricity markets of the neighbouring countries, the effect of Germany’s electricity
prices on the Dutch RES-E subsidy costs would be lower than what has been presented.
The subsidy costs to the Netherlands would depend more on the interconnection capac-
ities and the supply curves of the other neighbouring zones.
Market mechanisms outside of the day-ahead market have not ben analyzed: for in-
stance, intra-day markets, or balancing markets. Interviews with officials from the EPEX-
SPOT indicates that the volumes traded in the intra day markets have been rapidly in-
creasing over the last few years, as shares of intermittent RES-E electricity increase. This
trend can be expected to continue, as the shares RES-E increase even further. Since the
subsidy costs are currently only tied to the spot market (day ahead) prices, the analy-
sis in the dissertation holds. However, as the intra-day and balancing markets will play
a greater role, it is imperative that their impact on revenues for RES-E technologies are
evaluated while setting RES-E targets in the future. Such an evaluation would be in-
7 tricately tied with the role such markets play in attracting options for flexibility in the
system as well.
Investor behaviour in EMLab is represented by incorporating bounded rationality -
a distinctive feature of the model. The advantages of such a modelling feature is that
it allows us to analyse and represent observed phenomenon such as investment cycles.
Through the model-building process, after many rounds of trial and error, it was evident
that the magnitude and period of the cycles were largely dependent on the "future-time-
period" of the investing agent, i.e., the future time step for which an agent evaluates
an investment decision vis-a-vis the current time period. The longer into the future an
agent looks to make an investment, the more mistakes he is bound to make, as there
are more unknowns, thus increasing the amplitude and period of the cycles. If different
technologies are given different permit times and construction times, this impacts the
features of cycles correspondingly.

7.2.3. R EFLECTION ON APPROACH , MODEL

The framework introduced in Chapter 2 brings together two separate strands of litera-
ture: the institutional analysis framework and design theory. The combination is shown
to also be coherent with and complementary to the modelling paradigm of agent based
modelling and simulation. The framework does not provide causal relationships which
provide insight on why a certain actor behaves in a certain manner under certain condi-
tions. Nonetheless, it makes for a strong fundamental basis to simulate a test-bench on
7.3. F UTURE WORK 123

which the real world can be represented, and theories or "speculations" tested. Although
this framework was applied and demonstrated using renewable support schemes in Eu-
rope, its fundamental nature makes it eminently employable to other domains. The
claim is supported by the fact that each of the foundational concepts that make up the
framework have been applied in thousands of publications and are being actively used
across sectors, geographies, communities.
One aspect of the modelling framework presented that needs further attention how-
ever is the fact that choosing a limited number of design elements for analysis is, to say
the least, challenging. How does a researcher tell whether a certain element is important
or not? While this dissertation sheds little light on the topic, others such as Taeihagh et al.
(2014) have proposed and demonstrated some interesting ways to reduce the policy de-
sign element set. One such method uses a combination of interviews with stakeholders,
combinatorics, and network theory, to design policies and aid decision making.
A major advantage of the model is that it is open source. Most models that are used
to inform the European Commission’s policy documents are black box simulations. The
assumptions, data, systems used to set policy are not publicly available. Making models
open source allows for replicability, transparency, and for debating their validity, which
are all basic tenets of the scientific method.
The model has provided interesting insights on impacts of design elements of RES-E
support and on the nature of cross border interactions. The results pave the way towards
new questions that need to be resolved as the shares of RES-E increase in the elecricity
system. For instance, they raise pressing questions about the need to further develop a
methodology for sharing the paying the investment costs behind the subsidies between
countries. These are explained in greater detail in the following section on future work.

7
7.3. F UTURE WORK
The framework presented in this work can be extended in several ways. Its reliability and
usefulness could be tested by applying it to other socio-technical systems. It could be ex-
tended methodologically by incorporating endogenous policy emergence and evolution
using policy process theories.
For modellers of long term energy systems, it is paramount that models employ high
temporal resolution at high shares of intermittent RES-E generation. Creative techniques
to accomplish this while having endogenous investment in the model need to be in-
vented. One way of accomplishing this, without compromising on the constraints of
computational resources, is to run high resolution optimization algorithms to compute
hourly market clearing at intervals of 3 or 5 years (ticks), given a certain generation mix.
In Chapter 5 it was evident that the costs of RES-E electricity will rise due to high
spillage, if adequate temporal flexibility is not provided. An interesting point of depar-
ture is to compute the tradeoff between marginal storage and the marginal unit of RES-E
generation: at what point does an extra unit of storage cost lesser than an extra unit of
intermittent RES-E without storage? This would be interesting to policy makers as it sig-
nals whether, and at what point, storage technologies might need regulatory support to
be able to increase the share of RES-E in the system.
Given that electricity markets are intricately connected across national borders, ques-
tions regarding burden sharing of subsidy costs are complex, as merit order effects spread
 124 7. C ONCLUSIONS AND REFLECTIONS

across border. They should also depend on the neighbouring countries’ targets, genera-
tion mixes, and flexibilty sources. These questions need to be explored in greater detail,
and possible mechanisms to cope with the complexities should be evaluated.

7
R EFERENCES 125

R EFERENCES
G. N. Gilbert and K. G. Troitzsch, Simulation for the social scientist, 2nd ed. (Open Uni-
versity Press, Maidenhead, England ; New York, NY, 2005).

W. S. Parker, Franklin, Holmes, and the Epistemology of Computer Simulation, Interna-

tional Studies in the Philosophy of Science 22, 165 (2008).

E. Winsberg, Sanctioning Models: The Epistemology of Simulation, Science in Context 12

(1999), 10.1017/S0269889700003422.

E. Winsberg, Simulations, Models, and Theories: Complex Physical Systems and Their
Representations, Philosophy of Science 68, S442 (2001).

J. M. Epstein, Agent-based computational models and generative social science, Complex-

ity 4, 41 (1999).

E. Winsberg, Computer simulations in science, in The Stanford Encyclopedia of Philoso-

phy, edited by E. N. Zalta (Metaphysics Research Lab, Stanford University, 2015) sum-
mer 2015 ed.

E. B. Winsberg, Science in the age of computer simulation (University of Chicago Press,

Chicago, 2010) oCLC: 676697492.

D. Dowling, Experimenting on Theories, Science in Context 12 (1999),

10.1017/S0269889700003410.

I. Hacking, Representing and intervening: introductory topics in the philosophy of natural

science (Cambridge University Press, Cambridge [Cambridgeshire] ; New York, 1983). 7
A. Taeihagh, R. Bañares-Alcántara, and M. Givoni, A virtual environment for the formu-
lation of policy packages, Transportation Research Part A: Policy and Practice 60, 53
(2014).
 A
D ATA AND A SSUMPTIONS

A.1. D ATA

A.1.1. D EMAND AND F UEL P RICE T RENDS

Table A.1: Demand and Fuel Price Trends

Growth Rate
Start value Mode Min Max
Electricity demand growth rate 1 1.1 0.99 1.03
Gas price - Basecase 4 1 1 1
Gas price - high 4 1.02 1.04 1
Gas price - low 4 0.98 0.96 1

127
 128 A. D ATA AND A SSUMPTIONS

Figure A.1: Stochastic, annual demand growth factor trends for both Netherlands and Germany.

A.1.2. R ENEWABLE T ECHNOLOGY C OST C URVE A SSUMPTIONS

The assumptions for cost trends are based on data from IRENA reports and are in 2014
prices. The Solar PV investment costs are based on IRENA (2016), while cost trends for
wind offshore and onshore are based on IRENA and IEA-ETSAP (2016).
A.1. D ATA 129

Figure A.2: Assumed Investment cost trends per technology across the time period of the simulation

Figure A.3: Assumed operation and maintenance cost trends across the time period of the simulation

A.1.3. TARGET AND P OTENTIAL C URVES

The targets for renewable energy generation have been set by extrapolating the targets
mentioned in the National Renewable Energy Action Plan of the Netherlands; of Eco-
nomic Affairs Agriculture and Innovation (2010).The trends in csv format are attached in
the zipped folder.
Data points for ’realistic potentials’ at different years have been used to linearly ex-
 130 A. D ATA AND A SSUMPTIONS

trapolate trends for the whole time scope of the model. The data points and their sources
A are mentioned in the table below.

Table A.2: Realistic Technology Potentials

Technology Year Country Potential (in GWh) Source

Wind Onshore (2010, 2040) NL (2151.62,9032) Lako (2010)
Wind Offshore (2010, 2040) NL (837.27,58756) Lako (2010)
Photovoltaic (2013, 2020) NL (1065.19,10839.8) Ragwitz et al. (2003)
Wind Onshore (2014,2050) DE (49290.1596,123000) Scholz (2012)
Wind Offshore (2014,2050) DE (2053.17112,310000) Scholz (2012)
Photovoltaic (2014,2050) DE (34916,107000) Scholz (2012)
Renewable Targets and Potentials in DE in MW

P_Ante P_Post P_AnteTS P_PostTS

6e+08

4e+08

2e+08

Q_Ante Q_Post Q_AnteTS Q_PostTS

6e+08

4e+08

2e+08

0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40
Tick[Year]

variable RE_Target RE_Potential

Figure A.4: Assumed targets and potentials of renewable electricity technologies (all) in Germany in MWh.
A.1. D ATA 131

Renewable Targets and Potentials in NL in MW

P_Ante P_Post P_AnteTS P_PostTS
A
1e+08

5e+07

Q_Ante Q_Post Q_AnteTS Q_PostTS

1e+08

5e+07

0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40
Tick[Year]

variable RE_Target RE_Potential

Figure A.5: Assumed targets and potentials of renewable electricity technologies (all) in Netherlands in MWh.

A.1.4. A SSUMPTIONS : T ECHNOLOGY C HARACTERISTICS

Table A.3: Assumptions regarding fixed costs of conventional technologies

ts
sts

s
Co
Co

M
y

nt
log

y
O&
me

nc
no

cie
est

ed
ch

Effi
Inv

Fix
Te

ur]

ur]
ate

ate

ate
nE

nE
hr

hr

hr
t [i

t [i
t

t
ow

ow

ow
t
r

r
Sta

Sta

Sta
Gr

Gr

Gr

Coal Pulverised SC 1365530 0 40970 0 0.44 0.00327

Lignite 1700000 0 41545 0 0.37 0.00500
Nuclear 2874800 0 71870 0 0.33 0.00001
CCGT 646830 -0.008 29470 0 0.59 0.00207
 132 R EFERENCES

A A.2. C URRENT TECHNOLOGY COST DATA TRENDS

Figure A.6: Global weighted average utility-scale installed solar pv system costs and breakdown, 2009-2025,
Source: IRENA (2016)

R EFERENCES
IRENA, Solar PV in Africa: Costs and Markets, Tech. Rep. (2016).

IRENA and IEA-ETSAP, Wind Power Technology Brief , Tech. Rep. E07 (2016).

M. of Economic Affairs Agriculture and Innovation, National Renewable Energy Action

Plan of the Netherlands: Directive 2009/28/EC, Submitted to European Commission
(2010).

P. Lako, Technical and economic features of renewable electricity technologies, Tech. Rep.
ECN-E–10-034 (ECN, 2010).

M. Ragwitz, G. Resch, C. Huber, and S. White, Dynamic Cost Resource Curves, Tech.
Rep. ENG2-CT-2002-00607 (Energy Economics Group, Vienna University of Technol-
ogy, 2003).

Y. Scholz, Renewable energy based electricity supply at low costs - development of the
REMix model and application for Europe, Ph.D. thesis, German Aerospace Institute,
Stuttgart (2012).
 B
R ESULTS

B.1. C HAPTER 4: R ESULTS AND S ENSITIVITY A NALYSIS

B.1.1. F IGURES
Subsidy Cost in Eur/MWh

Subsidy Cost in Eur/MWh

90 90

60 60

30 30

0 0
P_Ante

P_Post

P_AnteTS

P_PostTS

Q_Ante

Q_Post

Q_AnteTS

Q_PostTS

P_Ante

P_Post

P_AnteTS

P_PostTS

Q_Ante

Q_Post

Q_AnteTS

Q_PostTS

Policy Scenario Policy Scenario

(a) Subsidy costs in Gas Low scenario (b) Subsidy costs in Gas High scenario

Figure B.1: Subsidy costs in scenarios with increasing or decreasing gas price trends

133
 134 B. R ESULTS

Subsidy Cost in Eur/MWh

Subsidy Cost in Eur/MWh

90 90

60 60

B 30 30

0 0
P_Ante

P_Post

P_AnteTS

P_PostTS

Q_Ante

Q_Post

Q_AnteTS

Q_PostTS

P_Ante

P_Post

P_AnteTS

P_PostTS

Q_Ante

Q_Post

Q_AnteTS

Q_PostTS
Policy Scenario Policy Scenario

(a) Subsidy costs in scenario set with (b) Policy cost effectiveness in scenario
the same risk aversion set with constant RES-E target

Figure B.2: Subsidy costs of scenarios addressing each effect on price setting individually

B.1.2. TABLES

Table B.1: Distributional Implications in million Eur.

Scenario ∆ Con- ∆ Pro- ∆ Govt ∆ Social

sumer ducer Surpl. Surpl.
Surpl. Surpl.
P_Ante 46.91 -61.86 -74.66 -89.61
P_AnteTS 18.12 -3.39 -10.73 4.00
P_Post 72.68 -47.84 -65.79 -40.96
P_PostTS 71.06 -13.15 -30.27 27.64
Q_Ante 65.24 -33.09 -58.71 -26.56
Q_AnteTS 65.45 -2.23 -26.49 36.73
Q_Post 65.34 -36.61 -61.91 -33.19
Q_PostTS 65.32 -7.71 -31.73 25.89

Table B.2: Comparison of Average Subsidy Between Base Case Set and Same Risk Scenario Set

Base Case Scenario Same Risk Scenario

Scenario Name Avg Subsidy/Unit Avg Subsidy/Unit Difference
(Eur/MWh) (Eur/MWh)
P_Ante 79.72 78.34 1.38
P_Post 93.39 93.11 0.28
P_AnteTS 27.88 27.24 0.64
P_PostTS 35.95 35.96 -0.01
Q_Ante 74.08 69.32 4.76
Q_Post 78.67 76.09 2.58
Q_AnteTS 28.92 27.48 1.44
Q_PostTS 36.28 36.30 -0.02
B.2. C HAPTER 5: R ESULTS 135

B.2. C HAPTER 5: R ESULTS

Subsidy and Revenue Per Unit for Photovoltaic in NL [Eur/MWh]

P_Ante P_AnteTS P_Post P_PostTS

250

200

150
B
100

50

Q_Ante Q_AnteTS Q_Post Q_PostTS

250

200

150

100

50

0
0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40
Time [a]

Variable Subsidy per MWh Elec. market revenue per MWh Fixed cost per MWh

Interconnector Capacity 0 MW 3950 MW 10000 MW

Figure B.3: Fixed costs, revenue from subsidy, revenue from spot market shown per MWh electricity generated,
across time.
 136 B. R ESULTS

Total Cost of Electricity Total Cost of Subsidy

600

Netherlands
400

200
Cost in Billion Euros

600

Germany
400

200

0
m h m h
Lo
w diu Hig Lo
w diu Hig
Int Me Int Int Me Int
Int Int

Interconnection Level

P_Ante P_Post Q_Ante Q_Post

Policy Scenario
P_AnteTS P_PostTS Q_AnteTS Q_PostTS

Figure B.4: Electricity and subsidy costs summed across 40 ticks, per country, and per RES-E policy
B.2. C HAPTER 5: R ESULTS 137

B
Interconnection 0 MW Interconnection 3950 MW Interconnection 10000 MW
500
400

P_Ante
300
200
100
0
500
400

P_AnteTS
300
200
Spillage of electricity generation from RES−E in Germany [TWh]

100
0
500
400

P_Post
300
200
100 technology
0
500 Photovoltaic
400

P_PostTS
300 Wind
200
100 WindOffshore
0
500
400
Q_Ante

300
200
type of production
100
Actual Production
0
500
Expected Production
400
Q_AnteTS

300
200
100
0
500
400
Q_Post

300
200
100
0
500
400
Q_PostTS

300
200
100
0
0 10 20 30 40 0 10 20 30 40 0 10 20 30 40
Time [in years]

Figure B.5: Expected versus actual production of RES-E due to spillage across time in Germany in TWh
 138 B. R ESULTS

B
Interconnection 0 MW Interconnection 3950 MW Interconnection 10000 MW
125
100

P_Ante
75
50
25
0
125
100

P_AnteTS
75
Spillage of electricity generation from RES−E in Netherlands [TWh]

50
25
0
125
100

P_Post
75
50
25 technology
0
125 Photovoltaic
100

P_PostTS
75 Wind
50
25 WindOffshore
0
125
100

Q_Ante
75
50
type of production
25
Actual Production
0
125
Expected Production
100
Q_AnteTS

75
50
25
0
125
100
Q_Post

75
50
25
0
125
100
Q_PostTS

75
50
25
0
0 10 20 30 40 0 10 20 30 40 0 10 20 30 40
Time [in years]

Figure B.6: Expected versus actual production of RES-E due to spillage across time in the Netherlands in TWh
 L IST OF P UBLICATIONS

P EER- REVIEWED JOURNAL ARTICLES

• K. Iychettira, R. Hakvoort, P. Linares, Towards a comprehensive policy for electricity from
renewable energy: An approach for policy design, Energy Policy 106, 169-182 (2017).

• K. Iychettira, R. Hakvoort, R. de Jeu, P. Linares, Towards a comprehensive policy for electricity

from renewable energy: Designing for Social Welfare., Applied Energy 187, 228-242 (2017).

• E. J. L. Chappin, L. J. de Vries, J. C. Richstein, P. Bhagwat, K. Iychettira, S. Khan Simulating

climate and energy policy with agent-based modelling: The Energy Modelling Laboratory
(EMLab), Environmental Modelling and Software 96, 421-431 (2017).

• P. C. Bhagwat, K. Iychettira, J. C. Richstein,E. J. L. Chappin, L. J. de Vries, The effectiveness

of capacity markets in the presence of a high portfolio share of renewable energy sources,
Utilities Policy 48, 76-91 (2017).

• P. C. Bhagwat, J. C. Richstein,E. J. L. Chappin, K. Iychettira, L. J. de Vries, Cross-border effects

of capacity mechanisms in interconnected power systems, Utilities Policy 46, 33-47 (2017).

U NDER REVIEW
• K. Iychettira, M. Klein, R. Hakvoort, Can subsidies be phased out under high shares of RES-E?
Evaluating primary influencing factors, Working paper submitted for publication, (2017).

• K. Iychettira, R. Hakvoort, P. Linares, Congruency between national support schemes and an

international electricity market: A case of The Netherlands and Germany, Working paper to
be submitted for publication, (2018).

C ONFERENCE PROCEEDINGS
• K. Iychettira, Towards a comprehensive policy for electricity from renewable energy: A Struc-
tured Design Approach, 3rd International Conference on Public Policy, Singapore, 28-30
June 2017.

• K. Iychettira, R. Hakvoort, R. de Jeu, P. Linares, Modelling long term impacts of renewable

electricity support designs on Social Welfare, 40th Annual IAEE International Conference,
Singapore, 18-21 June 2017.

• K. Iychettira, R. Hakvoort, R. de Jeu, P. Linares, Designing the Right RES-E Support Schemes:
assessing impacts of design elements using an agent-based model., WholeSEM Annual Con-
ference 2016, Cambridge, UK, 4-5 July 2016.

• K. Iychettira, P. Linares, R. Hakvoort, Harmonising RES-E support schemes using design el-
ements, 12th International Conference on the European Energy Market, Lisbon, Portugal,
19-22 May 2015.

139
140 L IST OF P UBLICATIONS

• K. Iychettira, P. C. Bhagwat, J. C. Richstein, L. J. de Vries, Interaction between Capacity Mar-

kets and Investment into Renewable Energy in the Netherlands and Germany, 37th Annual
IAEE International Conference, New York, USA, 15-18 June 2014.
 C URRICULUM V ITÆ
Kaveri Kariappa I YCHETTIRA

Kaveri was born on the 30th of December 1989. She spent her early years in Mysore,
and later moved to Bangalore where she completed her secondary education and bach-
elor degree. Through school and early college, she showed a penchant for writing; she
wrote for and edited several community magazines and newsletters. In 2007 she gradu-
ated from pre-university college, and in 2011, obtained a Bachelor in Engineering, with
a specialization in Electrical and Electronics, from RV College of Engineering in Banga-
lore. During 2010-11 she also interned at the Centre for Science Technology and Policy in
Bangalore, where she worked on analysing the impact of India’s National Solar Mission.
She went to the Netherlands in 2011, to pursue a Masters in Engineering and Policy Anal-
ysis at Delft University of Technology. She was awarded the Justus van Effenfond Schol-
arship, a full funding of her Masters education, granted to only 8 students in 2011 across
the University. During her Masters, she specialized in the disciplines of Modelling and
Economics. She wrote a dissertation analysing the impact of capacity markets in the
electricity sector in The Netherlands and Germany. She was awarded the "Best Graduate
Award" for the Faculty of TPM at the Delft University of Technology for the year 2013.
In October 2013, she began an Erasmus Mundus Joint Doctorate in Sustainable Energy
Technologies and Strategies (SETS) at Delft University of Technology. As part of the SETS
programme, she spent 11 months at Comillas Pontifical University in Madrid, Spain, and
6 months at KTH Royal Institute of Technology in Stockholm, Sweden. During her doc-
toral programme, she published several peer-reviewed journal publications, supervised
4 master thesis projects, and presented at 5 international conferences. She also served
as the student representative for the 32 students across the three universities during
2015-16. She began working as a post-doctoral fellow at the Belfer Centre of the Har-
vard Kennedy School since September 2017.

141