Departamento de Ingeniería de la Organización,

Administración de Empresas y Estadísticas

UNIVERSIDAD POLITÉCNICA DE MADRID

Escuela Técnica Superior de Ingenieros Industriales

A multi-level analysis of

innovation diffusion:

The case of solar photovoltaic systems in Germany

and

the disillusion of grid parity

PhD Thesis

Emrah Karakaya

Supervisors:

Antonio Hidalgo Nuchera

Universidad Politécnica de Madrid, Spain

Cali Nuur

KTH Royal Institute of Technology, Sweden

2015

iii

Tribunal nombrado por el Magfco. y Excmo. Sr. Rector de la Universidad Politécnica de

Madrid, el dia … de …………... de 2015.

Presidente:

Secretario:

Vocal:

Vocal:

Vocal:

Suplente:

Suplente:

Realizado el acto de lectura y defensa de la Tesis el día … de …….……. de 2015 en la

Escuela Técnica Superior de Ingenieros Industriales.

El Presidente El Secretario

Los Vocales

v

Acknowledgements

When I started my journey from Germany to Spain (and to Sweden) in 2011 September, I

knew this opportunity might turn into a great place to discover the knowledge about the

society and, consequently, myself. I knew that the degree of Doctor of Philosophy (PhD)

would not help or change me, but my day-to-day interaction with people would do so. I

have extremely enjoyed having the opportunity to learn from my peers, friends and

colleagues from all over the world, particularly those I met in Spain, Sweden and

Germany. These people have inspired me in every aspect, both academically and socially,

and have definitely influenced who I am today and how this PhD Thesis has been

developed.

First of all, I would like to thank all the people I met at Universidad Politecnica de

Madrid (UPM) during 2 and half years. I thank Antonio, my PhD supervisor, for his great

support and calm guidance during the entire process. I really enjoyed my time in UPM,

also thanks to my awesome colleagues; Marcia, Marin, Frano, Shoaib, Fadi, Sarah, Raul,

Jing, Mitra, Inez, Vicente, Gustavo and Simon. If I know something about the world of

science, politics, academics and society now, it is thanks to these great people. In addition,

I thank Ignacio Romero for his valuable feedback and help during the last year of my PhD.

I also had an amazing 1 year at KTH Royal Institute of Technology in Sweden. My PhD

supervisor, Cali, deserves special thanks. He has always supported me, believed in me and

motivated me in every way possible. Moreover, I thank Johan Hoffmann for his valuable

comments and help during the third year of my PhD. In addition, I was surrounded by

amazing people at KTH, Monia, Maxim, Jonatan, Pranpreya, Niklas, Yasmine M., Mats,

Richard, Sudipa, Nidal, Michael, Annika and Vicky, and I was amazed by the friendly and

productive atmosphere at KTH.

I would like to thank all of those who were involved in making the European Doctorate

of Industrial Management (EDIM) program possible: among others, Martina, Linda,

Felipe, Paolo, Mats, Antonio, Cali, Isabel and Kristin. During the EDIM workshops, the

feedback of Mats, Massimo and Niklas on my work have been very influential and

motivational for me. These workshops have not only improved my academic skills but

also let me get to know great people: Anna, Andres, Hakan, Guido, Isaac, Milan, Seyoum,

Keivan, Yasmine H., and Vikash. Especially, I cannot thank enough Anna, Monia and

Andres, without whom, I would never ever have enjoyed these 4 years so much.

I also express my sincere gratitude to Thomas Hartmann (Hartmann Energietechnik

GmbH) and his colleagues for giving me the opportunity to learn and to obtain the

empirical data in Germany. I especially thank all of the respondents who were

interviewed, including the directors from Solar Partner Association and the residents of

Rottenburg am Neckar. Hubert Wellhausser deserves special thanks as well. He has not

only helped me to understand the German context but also inspired me to be a calmer,

happier, more conifedent and more optimistic person. I also would like to thank Barbara

vi

Breithschopf of the Fraunhofer Institute for Systems and Innovation Research ISI in

Germany for her kind guidance through my 3 month stay in Karlsruhe.

In addition, I have enjoyed being an active member of Erasmus Mundus Student and

Alumni Association (EMA). The amazing people I have met through EMA have been

very influential for developing my interpersonal and intercultural skills. I would like to

express my sincere gratitude to, among others, Pelicano, Anna, Peter, Georgiana, Tahseen

and Leasa, for inspiring me in many personal and professional aspects.

Last but not least, I thank my amazing friends and family. Although my parents and

brother have been 3000 km away, I have enjoyed their encouragement and support all the

time during this work. Sizi çok çok çok seviyorum!

Madrid, 16.06.2015

Emrah Karakaya

vii

This research was conducted within the framework of the “European Doctorate in

Industrial Management” - EDIM - which was funded by The Education, Audiovisual and

Culture Executive Agency (EACEA) of European Commission under Erasmus Mundus

Action 1 programmes.

ix

Abstract

Will renewable energy sources supply all of the world energy needs one day? Some argue yes,

while others say no. However, in some regions of the world, the electricity production through

renewable energy sources is already at a promising stage of development at which their electricity

generation costs compete with conventional electricity sources’, i.e., grid parity. This achievement

has been underpinned by the increase of technology efficiency, reduction of production costs and,

above all, years of policy interventions of providing financial support. The diffusion of solar

photovoltaic (PV) systems in Germany is an important frontrunner case in point. Germany is not

only the top country in terms of installed PV systems’ capacity worldwide but also one of the

pioneer countries where the grid parity has recently been achieved. However, there might be a

cloud on the horizon. The diffusion rate has started to decline in many regions. In addition, local

solar firms – which are known to be important drivers of diffusion – have started to face

difficulties to run their businesses. These developments raise some important questions: Is this a

temporary decline on diffusion? Will adopters continue to install PV systems? What about the

business models of the local solar firms? Based on the case of PV systems in Germany through a

multi-level analysis and two complementary literature reviews, this PhD Dissertation extends the

debate by providing multiple wealth of empirical details in a context-limited knowledge. The first

analysis is based on the adopter perspective, which explores the “micro” level and the social

process underlying the adoption of PV systems. The second one is a firm-level perspective, which

explores the business models of firms and their driving roles in diffusion of PV systems. The third

one is a regional perspective, which explores the “meso” level, i.e., the social process underlying

the adoption of PV systems and its modeling techniques. The results include implications for both

scholars and policymakers, not only about renewable energy innovations at grid parity, but also in

an inductive manner, about policy-driven environmental innovations that achieve the cost

competiveness.

Keywords: Solar; deployment; business model; demand; policy; adoption, eco-innovation

xi

Resumen

¿Suministrarán las fuentes de energía renovables toda la energía que el mundo necesita algún día?

Algunos argumentan que sí, mientras que otros dicen que no. Sin embargo, en algunas regiones del

mundo, la producción de electricidad a través de fuentes de energía renovables ya está en una

etapa prometedora de desarrollo en la que su costo de generación de electricidad compite con

fuentes de electricidad convencionales, como por ejemplo la paridad de red. Este logro ha sido

respaldado por el aumento de la eficiencia de la tecnología, la reducción de los costos de

producción y, sobre todo, los años de intervenciones políticas de apoyo financiero. La difusión de

los sistemas solares fotovoltaicos (PV) en Alemania es un ejemplo relevante. Alemania no sólo es

el país líder en términos de capacidad instalada de sistemas fotovoltaicos (PV) en todo el mundo,

sino también uno de los países pioneros donde la paridad de red se ha logrado recientemente. No

obstante, podría haber una nube en el horizonte. La tasa de difusión ha comenzado a declinar en

muchas regiones. Además, las empresas solares locales – que se sabe son importantes impulsores

de la difusión – han comenzado a enfrentar dificultades para manejar sus negocios. Estos

acontecimientos plantean algunas preguntas importantes: ¿Es ésta una disminución temporal en la

difusión? ¿Los adoptantes continuarán instalando sistemas fotovoltaicos? ¿Qué pasa con los

modelos de negocio de las empresas solares locales? Con base en el caso de los sistemas

fotovoltaicos en Alemania a través de un análisis multinivel y dos revisiones literarias

complementarias, esta tesis doctoral extiende el debate proporcionando riqueza múltiple de datos

empíricos en un conocimiento de contexto limitado. El primer análisis se basa en la perspectiva del

adoptante, que explora el nivel "micro" y el proceso social que subyace a la adopción de los

sistemas fotovoltaicos. El segundo análisis es una perspectiva a nivel de empresa, que explora los

modelos de negocio de las empresas y sus roles impulsores en la difusión de los sistemas

fotovoltaicos. El tercero análisis es una perspectiva regional, la cual explora el nivel "meso", el

proceso social que subyace a la adopción de sistemas fotovoltaicos y sus técnicas de modelado.

Los resultados incluyen implicaciones tanto para académicos como políticos, no sólo sobre las

innovaciones en energía renovable relativas a la paridad de red, sino también, de manera inductiva,

sobre las innovaciones ambientales impulsadas por las políticas que logren la competitividad de

costes.

xiii

Disclaimer

Some parts of this PhD Thesis have already been published as:

Karakaya, E., Nuur, C., & Hidalgo, A. (2016). Business model challenge: Lessons

from a local solar company. Renewable Energy, 85, 1026–1035.

Karakaya, E., & Sriwannawit, P. (2015). Barriers to the adoption of photovoltaic

systems: The state of the art. Renewable and Sustainable Energy Reviews, 49, 60–

66.

Karakaya, E., Hidalgo, A., & Nuur, C. (2015). Motivators for adoption of

photovoltaic systems at grid parity: A case study from Southern Germany.

Renewable and Sustainable Energy Reviews, 43, 1090-1098.

Karakaya, E., Hidalgo, A., & Nuur, C. (2014). Diffusion of eco-innovations: A

review. Renewable and Sustainable Energy Reviews, 33, 392-399.

xv

Index

1. Introduction ................................................................................................................... 1

1.1. Setting the scene ..................................................................................................... 1

1.2. The literature gaps, aim and research questions ..................................................... 3

1.3. Structure and outline of the Thesis ......................................................................... 8

2. Solar PV systems ......................................................................................................... 13

2.1. Technology ........................................................................................................... 13

2.2. Diffusion............................................................................................................... 16

2.3. Grid parity ............................................................................................................ 19

3. Diffusion of innovations .............................................................................................. 23

3.1. Theoretical approach ............................................................................................ 23

3.2. State-of the art ...................................................................................................... 26

3.3. The multi-level perspective .................................................................................. 46

4. Methodology ................................................................................................................ 59

4.1. Case study approach ............................................................................................. 59

4.2. Indicator-based approach ..................................................................................... 64

4.3. Finite element method .......................................................................................... 70

5. Results and discussions ............................................................................................... 77

5.1. Micro level analysis ............................................................................................. 77

5.2. The business models............................................................................................. 83

5.3. Meso level analysis .............................................................................................. 91

5.4. Mathematical model ............................................................................................. 96

5.5. A meta-analysis .................................................................................................. 100

6. Conclusions, contributions and limitations ............................................................... 105

6.1. Theoretical implications and contributions ........................................................ 105

6.2. Methodological implications and contributions ................................................. 107

6.3. Policy implications ............................................................................................. 108

xvi

6.4. Limitations and future research .......................................................................... 109

6.5. Conclusions ........................................................................................................ 111

7. Appendix .................................................................................................................... 113

7.1. List of publications ............................................................................................. 113

8. References .................................................................................................................. 115

xvii

List of figures

Figure 1. The structure of the PhD Thesis .......................................................................... 11

Figure 2. Some historical steps for PV systems .................................................................. 14

Figure 3. A solar PV system on a roof-top ......................................................................... 14

Figure 4. Best research cell efficiencies ............................................................................. 16

Figure 5. Renewable energy share of global electricity production ................................... 17

Figure 6. Annual installations of PV systems ..................................................................... 17

Figure 7. Numbers of yearly PV installations in Germany ................................................. 18

Figure 8. Number of photovoltaic installations in three districts........................................ 19

Figure 9. Comparison of electricity generation costs ......................................................... 21

Figure 10. Monthly installations vs. feed-in tariff in Germany .......................................... 21

Figure 11. The comparison of two literature reviews ......................................................... 27

Figure 12. The evolution of research articles related to "diffusion of eco-innovations" .... 32

Figure 13. Penetration rate in percent ................................................................................. 54

Figure 14- The location of HET.......................................................................................... 60

Figure 15- The HET and its neighborhood ......................................................................... 61

Figure 16. Geographic zip code areas and states in Germany ............................................ 65

Figure 17- Typical implementation steps for Finite Element Model .................................. 71

Figure 18- The illustration of flow in innovations in space ................................................ 73

Figure 19- The modeled geography from Southern Germany ............................................ 74

Figure 20- The computational mesh on approximated geographic .................................... 75

Figure 21- Different PV solutions for various needs and concepts .................................... 84

Figure 22- The value network of a local solar company .................................................... 87

Figure 23- The comparison between diffusions of PV systems at different levels ............ 88

Figure 24- Revenue of HET from PV installations and other services and products ......... 89

Figure 25- The declining average cost of PV installations ................................................. 90

Figure 26. Box plots of indicators for each state ................................................................ 91

Figure 27. The spatial diffusion pattern of PV in 10 zip code areas of Germany .............. 93

Figure 28. The growth rate for cumulative number of PV installations in Germany ......... 94

Figure 29- Diffusion rate after 13 years .............................................................................. 97

Figure 30- The patterns of s-curves at 5 locations .............................................................. 97

Figure 31- Innovation flux after 3 months, 4 years, 7 years and 10 years .......................... 99

Figure 32- Diffusion ratio after 3 months, 4 years, 7 years and 10 years ........................... 99

xviii

Figure 33. Discussions diagram ........................................................................................ 100

xix

List of tables

Table 1. The link between research questions and methods ................................................. 9

Table 2. Keywords used for compiling the database from Google Scholar. ...................... 29

Table 3. Overview of reviewed publications by country .................................................... 31

Table 4. Business model components ................................................................................. 50

Table 5. A system of indicators that proxy lead market attributes for PV.......................... 67

Table 6. The values for innovation diffusivity.................................................................... 74

Table 7. The spatial visualization of the indicators for lead market attributes ................... 92

1

1. INTRODUCTION

1.1. SETTING THE SCENE

Concerns about scarcity of natural resources, climate change and geopolitics have

prompted governments to support the diffusion of sustainable modes of energy

production. As a recent special report from the Economist (2015, p. 3) argues, “modern

life is based on the ubiquitous use of fossil fuels, all of which have big disadvantages.

Coal, the cheapest and most abundant, has been the dirtiest, contributing to rising

emissions. Oil supplies have been vulnerable to geopolitical shocks and price collusion by

producers. Natural gas has mostly come by pipeline—and often with serious political

baggage, as in the case of Europe’s dependence on Russia. Nuclear power is beset by

political troubles, heightened by public alarm after the accident at Japan’s Fukushima

power station in 2011”. For a long time, increasing the share of renewable energy share,

mainly for replacing the fossil fuels and nuclear power, has been high on the policy

agenda of several countries. For instance, the European Union (EU) has set a target of

20% share of renewable energy in the overall energy consumption by 2020 (EP, 2009).

This is a part of the process of “democratizing” the energy market (da Graça Carvalho et

al., 2011). The democratization of the energy market is often linked to the decentralized

structures of electricity production through renewable energy innovations (see e.g. S.

Wirth, 2014, p. 236). In such markets, households do not only have the right for choosing

between the renewable energy sources (e.g., solar, wind and biomass) and non-renewable

ones (fossil and nuclear fuels) but also produce energy and distribute it.

There are several sources of renewable energy which are expected to pave the way for this

transition towards sustainable modes of energy production. One of this is solar

photovoltaic (PV) technology, which generates electricity from solar radiation. The

ambition is that the PV adopters can generate electricity, consume the part they need and

then; supply the rest of the electricity to the grid (if they cannot store). Although the

electricity generated by PV systems account only for 0,7% of the global electricity

production and 2.6% of the Europe’s (EPIA, 2013; Ren21, 2014), the International Energy

Agency (IEA) argues that the sun could be the world’s largest source of electricity by

2

2050, ahead of fossil fuels, nuclear power and other sources of renewable energy such as

wind and hydropower (IEA, 2014). Through the reduction of production costs and the

increase of technology efficiency, solar PV systems could be argued to be becoming

competitive when compared with conventional electricity sources in terms of the levelized

cost of electricity generation. This cost competitiveness is often referred to the “grid

parity” which is a time and space specific stage. When the solar grid parity is achieved in

one region, it means that the cost of generating PV electricity is cheaper than the retail

price of electricity. According to a recent literature review by Munoz et al., (2014), several

scholars believe that achieving the grid parity is an important milestone for wider

diffusion of solar PV systems. The hope is that grid parity might make the diffusion of PV

systems independent of policy support. But the question is whether this is happening.

Recently, in some regions of the world, grid parity of solar PV systems has been achieved.

According to PV parity project, co-financed by the Intelligent Energy Europe programme

of the European Commission, Italy, Spain and Germany did achieve the grid parity in

2012, while Austria, Belgium, Czech Republic, France, Greece, Portugal, the Netherlands

and the United Kingdom had not yet (PvParity, 2013). What makes Germany is one of the

pioneer countries of grid parity is a result of its competitive installations costs for PV

systems, low interest rates on financial loans and high electricity retail costs (compared to

many other countries). Beyond the grid parity, Germany also represents the largest market

in the world, having 27% share of global PV installed capacity (EPIA, 2014a), which

supplies around 5.2% of country´s electricity consumption (Wirth, 2013). Although

having a relatively limited solar radiation potential, the diffusion of solar PV systems in

Germany has outperformed those of other countries, which have larger land area and

higher solar radiation, e.g., the USA, Spain, Turkey, Brazil, Saudi Arabia and Japan.

One of the most important drivers of solar PV diffusion in Germany has been the

implementation of Renewable Energy Act (EEG) in 2000 and its subsequent amendment

in 2004 (Dewald and Truffer, 2011; Jacobsson and Lauber, 2006). EEG is a nationwide

act which applies to all states in Germany. It financially supports PV adopters through

feed-in tariff schemes in all PV market segments. Under a feed-in tariff, adopters are paid

a cost-based price for the PV electricity they supply to the grid. The effective

implementation of EEG had resulted in continuous increase of the number of yearly

installations in Germany lasting until 2011. However, since 2012, the market has been

facing new conditions. First of all, solar PV energy in Germany has already achieved grid

3

parity. Secondly, the policy support, i.e., the feed-in tariff of EEG, probably as a

consequence of the grid parity, has been scaled down. Thirdly, the diffusion rate has

started to decline for the very first time in many regions of Germany. These new

conditions raise some interesting questions. What happens after achieving the grid parity?

Will diffusion rate incline or decline? What about the solar firms, will they be affected by

the change?

1.2. THE LITERATURE GAPS, AIM AND RESEARCH QUESTIONS

For the purpose of this Thesis, PV systems are conceptualized as environmental

innovations. However, this raises the questions of why and how this could be possible. In

the policy discourse, the notion of innovation is used rather loosely. It appears that

everything is about innovations. Influenced by policy discourse or not, the diffusion of a

variety of innovations has been the subject of study for management and economics

scholars for several decades. Innovation, defined as an idea, practice, or object that is

perceived as new by an individual or other unit of adoption (Rogers, 2003), is the core of

the diffusion process. Innovation is a very broad term and indeed can be anything as long

as it is perceived new by adopters. The variety of innovations that are analyzed in the

literature goes back to the early scholars of diffusion. For example, Griliches (1957)

analyzed the diffusion of hybrid seed corn, which is a product of controlled crossing of

specially selected parental strains, among American farmers. Mansfield (1961) studied the

diffusions of twelve type of innovations, such as, among others, trackless mobile loader

and continuous mining machine, among firms in particular industries. The examples of

innovations in the literature are many. Installing solar PV system is just another kind of

innovation for adopters. However, from a conceptual viewpoint, one can raise a question

worth pondering. Why can electricity generation from the solar PV system be perceived as

an innovation? There are two possible arguments for conceptualizing the electricity

generation from PV system as an innovation. Firstly, none of the new PV adopters had

adopted the PV systems before. Therefore, the solar PV installation is (will be) perceived

as new. Secondly, even in the case of re-adoption, solar PV system installation would still

be perceived as new. This is because solar PV systems have been characterized by

continuous incremental changes on their efficiency and appearance, and therefore, the new

installation will be perceived as an innovation by re-adopters as well. Moreover, if PV

systems can be conceptualized as innovations, why also as environmental ones? Beise and

Rennings (2005) argues that environmental innovations are the innovations that avoid or

4

reduce environmental harms. Is it really the case for the PV systems? This can be an

important debate to discuss in detail - which will probably go beyond the scope of this

Thesis. In a nutshell, I am aware of the limitations of the conceptualization PV systems as

environmental innovations. For example, manufacturing of solar PV systems can lead

some negative environmental impacts in terms of waste disposal (Kannan et al., 2006, p.

562) and greenhouse gas intensive production in some particular regions (Hsu et al., 2012,

p. 132). In addition, the regional diffusion of solar PV systems may result in global

rebound effects, i.e., the reduction of greenhouse gas emissions through PV systems in a

particular region may lead an increase of emissions in another region (see Van den Bergh,

2013a, 2013b). In spite of all these, for the purpose of this Dissertation, it is assumed that

adoption of PV systems reduce the usage of non-renewable energy sources such as fossil

fuels. It is because, in many cases of PV adoption, the electricity generation from PV

systems replace the electricity generation from fossil fuels and nuclear energy. As widely

known, the combustion of fossil fuels produces the greenhouse effect and air pollutants.

This means that each adoption of PV systems is likely to reduce or avoid the production of

greenhouse effect and air pollutants. This is also in line with the research on life cycle

analysis that compares the environmental impact of solar PV systems and fossil fuels (e.g.,

Fthenakis et al., 2008).

In the literature, there has been a broad agreement that the diffusion of innovations does

follow a spatiotemporal pattern (Brown, 1975; Hägerstrand, 1967), so is the case of the

diffusion of solar PV systems. However, it does not necessarily mean that diffusion

always follows a local to regional and regional to national pathway in time and space.

Some areas or cities could have better interconnectivity even though they are located on

the other sides of the world, e.g., the cities on the spice trade route in the medieval period

(5th

-15th

century) or the cities that host the airline hubs in the 21st century. In spite of all

these complexity of interconnectivities, one can still observe a typical s-curve when the

diffusion rate is plotted over a length of time (Bass, 1969; Rogers, 1962) or wave-like

propagations when it is plotted over a length of time and space (Haynes et al., 1977).

These diffusion paths seem to be easily observable, but this is not always the case for all

kinds of innovations. Sometimes, policymakers, through interventions, may induce or

hinder any stage of the diffusion, as it recently happens in the cases of environmental and

5

renewable energy innovations1. Addressing this, the literature on diffusion of

environmental innovations such as renewable energy technologies has also focused on

policy measures (e.g. Beise and Rennings, 2005; Foxon and Pearson, 2008; Tsoutsos and

Stamboulis, 2005). In theory, policy makers, by introducing different measures, e.g.,

subsidies or feed-in tariffs, aim to compensate the high upfront costs of environmental and

renewable energy technologies faced by potential adopters. However, high costs are not

anymore a barrier for some particular technologies in some regions. Solar PV systems

provide an important illustration for this (see Karakaya and Sriwannawit, 2015). Although

solar grid parity have been achieved in a few regions, e.g., southern Germany, and will

soon be achieved in the many others2, understanding diffusion dynamics at grid parity

remains as an understudied topic in the literature.

With this in mind, this Dissertation aims to provide a multi-level analysis on the diffusion

of solar photovoltaic systems at grid parity. In the literature, for multi-level analysis

approach, the definitions of micro, meso and macro levels are diverse (see e.g., Geels,

2010; Hannah and Lester, 2009; MacVaugh and Schiavone, 2010; Waarts and van

Everdingen, 2005). For the purpose of this Dissertation, these levels are defined as same

as the level of observation on diffusion. Micro level is the individual level; meso level is

the regional level and the macro level is the global level. Although, the term of “region” is

often used for sub-national territories (see e.g. Cooke and Morgan, 1998), “region” in this

Thesis refers to any particular geographic or political area of the world. This means, a

region can be a country, a sub-area of a country or a set of countries.

So what? Why the multilevel analysis approach? The diffusion of innovations is often

considered as a complex phenomenon (see Garcia and Jager, 2011; Kiesling et al., 2011).

This means that if an innovation is adopted by some individuals, it might (or not) give a

rise to diffusion on macro scales. As Rogers et al., (2005) argues micro and meso level

behaviors influence each other. Micro-level is vital to be understood, as it leads to the

emergence behavior at the meso level. The meso level is also important to be observed, as

it influences micro behaviors as a feedback. Diffusion of innovations in the energy field,

particularly those that have environmental innovation characteristics such as renewable

1 The detailed discussions on the concept of environmental and renewable energy innovations are given in the chapter 3. 2 According to PV Parity Project (2011-2013), which was co-financed by the Intelligent Energy Europe programme of

the European Commission, majority of the EU member states will expected to achieve the PV grid parity for residential

sector by 2020 (see PvParity, 2013)

6

energy innovations, constitutes even more complex patterns. One the one hand, policy

makers or incumbent actors might intervene the process at meso and macro levels (see e.g.

Darmani et al., 2014; Jacobsson and Lauber, 2006). On the other hand, the influence of

local actors could drive the diffusion at micro and meso levels (see e.g. Fabrizio and

Hawn, 2013; Graziano and Gillingham, 2014). Therefore, I believe, in order to understand

the diffusion of solar PV systems at grid parity, we need to consider the multi-levels of

phenomenon.

Interplaying between micro and meso level of diffusion, this Thesis provides three

different types of analysis and a proposal of modeling method. The first one is the micro

level analysis of the diffusion of solar PV systems at grid parity. The analytical

underpinnings of this analysis is the theory of diffusion of innovations (Rogers, 2003).

Moreover, this analysis aims to contribute on the literature of renewable energy. To date,

renewable energy research has extensively studied the factors that influence the diffusion

rate of solar PV systems. These factors are identified as policy measures, including feed-in

tariffs (Hoppmann et al., 2013; Jacobsson and Lauber, 2006; Jäger, 2006; Zhang et al.,

2011), peer effects and adopters’ socioeconomic background (e.g. Bollinger and

Gillingham, 2012; McEachern and Hanson, 2008; Peter et al., 2002; Zhang et al., 2011)

and the role of local solar companies (e.g. Dewald and Truffer, 2012; Fabrizio and Hawn,

2013). Although some recent studies have conducted economic analysis for the grid parity

of solar PV systems (e.g. Haas et al., 2013; Spertino et al., 2014), less has been said about

the micro-level social processes underlying the diffusion at grid parity. Therefore, I raise

the following research question for the micro level. RQ1. Why and how do the PV

systems at grid parity get adopted at micro level?

The second analysis explores the link between the firms’ business models and the

diffusion of PV systems at micro and meso levels. This analysis is anchored in the

assumption that the diffusion of PV systems is affected not only by the technology, policy

and adopters but also by the firms which supply and install the solar PV systems. During

diffusion of innovations, the firms can act as change agents at micro level, while forming

the necessary market structure at meso level. The business models of firms have been

studied for several contexts such as transportation (e.g., Kley et al., 2011; Tongur and

Engwall, 2014) and energy (e.g., Nair and Paulose, 2014; Okkonen and Suhonen, 2010).

Although the business models in PV sector have also been discussed (e.g. Huijben and

Verbong, 2013; Loock, 2012; Richter, 2013a), the empirical focus has only been limited to

7

the large sized electric utilities, solar cells production firms and households. However,

little attention has been paid to the business models of local solar firms, which are known

to be important drivers of diffusion. Thus, the following research question is formulated.

RQ2. How do business models of local solar firms and the diffusion of solar PV systems

at grid parity affect each other?

The third analysis is at meso level. The theoretical underpinnings of this analysis is based

on the lead market hypothesis (Beise, 2001). This analysis aims to contribute on both

literature of diffusion of innovations and the research on sustainability transitions, a

research field that focuses on systematic shifts for the adoption of environmental

innovations in a variety of industrial areas including energy. Although the importance of

space and sub-national regions has recently received attention in the literature on

sustainability transitions, there are not many studies on diffusion phase. Some studies

(Coenen et al., 2012; Truffer and Coenen, 2012) reveal that the majority of previous

literature had failed to reflect about the space-specific contexts and extend an invitation

for research that considers the spatial heterogeneity of regions. Consequently, a variety of

studies (e.g. Binz et al., 2014; Smith et al., 2014) have studied the space on sustainable

transitions and shed light on the role of actors, networks and institutions. Yet, little

systematic attention has been paid to the demand side of sub-national dynamics for such

transitions. Understanding the patterns of diffusion at meso level is particularly critical for

innovations with both private and public good characteristics such as PV systems (see e.g.

Fabrizio and Hawn, 2013; Graziano and Gillingham, 2014). Therefore, I raise the

following research question. RQ3. How do the PV systems at grid parity diffuse at meso

level?

In addition, this Thesis proposes a modeling method at meso level. Modeling the diffusion

patterns of innovations at meso level has been a great interest of a variety of scholars for

several decades, through models that have been based on one dimension, i.e., time (e.g.

Bain, 1963; Harvey, 1984; Meade & Islam, 1998; Sharif & Islam, 1980), and two

dimensions, i.e., time and space (Hägerstrand, 1967; Haynes et al., 1977; Morrill, 1970).

In the literature of solar PV systems, both one and two dimensional models have been

studied (e.g. Guidolin and Mortarino, 2010; Higgins et al., 2014; Kwan, 2012; Luque,

2001; Masini and Frankl, 2002; Mesak and Coleman, 1992). Through these models, the

ambition is to forecast the diffusion of solar PV systems based on different scenarios. If

this can be achieved, policy makers can develop more efficient measures to drive the

8

diffusion process of solar PV systems in the future. However, the majority of current

modeling methods have failed to address the impact of spatial heterogeneity and,

therefore, not have been able to forecast the diffusion in space domain. Thus, the

following research question is formulated. RQ4. How can the diffusion of solar PV

systems be better forecasted at meso level?

1.3. STRUCTURE AND OUTLINE OF THE THESIS

In order to understand the micro and meso level mechanisms behind such energy-related

phenomena, interdisciplinary and multi-methodological approaches are needed (Sovacool,

2014a). Addressing this need, this Dissertation has been developed through a variety of

methods (see Table 1), some results of which have been already published (Karakaya and

Sriwannawit, 2015; Karakaya et al., 2016, 2015, 2014a). The first method has been

systematic literature review approach. The other methods are case study method,

indicator-based approach and finite element method, through which the four overarching

and interrelated RQs of this Thesis have been specifically addressed. The two literature

reviews synthesize the relevant studies in two research domains: diffusion of eco-

innovations and diffusion of solar PV systems. The contribution of these systematic

literature reviews is twofold. Firstly, they help on developing the theoretical frameworks

to be used for answering the RQs. Secondly, the results of them are used for interpreting

the findings and implications of this Thesis. For example, on the one hand, the lead market

model, which is identified to be a comprehensive concept for studying diffusion of

environmental innovations (according to the systematic literature review on environmental

innovations), is used in developing the theoretical framework for meso level analysis. On

the other hand, the barriers identified in the review on solar PV systems are used for

interpreting the findings of micro level analysis.

9

TABLE 1. THE LINK BETWEEN RESEARCH QUESTIONS AND METHODS3

Level Method

Innovation Focus Relevance to Research

Questions

Inn

ovati

on

in

Gen

eral

En

vir

on

men

tal

Inn

ovati

on

s

Sola

r P

V s

yst

ems

Q1 Q2 Q3 Q4

Mix Literature

Reviews

*** * *

*** * *

Micro Case Study

*** *** *

Firms *** ***

Meso

Indicator-based * * * * *** *

Finite Element * * * ***

The overarching spatiotemporal focus of the Thesis is Germany in 2012 and the

technology in focus is solar PV system. However, each analysis has a slightly different

combination of spatiotemporal and technological scope. The case study is based on a case

of solar PV systems with less than 40kWp capacity in Tübingen in Germany in 2012,

while the indicator-based approach focuses on the case solar PV systems with less than

10kWp capacity of whole Germany from 2000 to 2012. In early 2000s, the installations

that have a less than 10 kWp capacity were mostly households’ adoption. However, in

early 2010s, the installations that have a less than 40 kWp capacity were mostly

households. Therefore, in this Thesis, the adopters are sometimes referred as households.

When it comes to literature reviews, on the one hand, the first literature review is based on

the research papers on eco-innovations published from 1973 to 2012. On the other hand,

the second literature review covers the literature, which was published from 2011 to 2013,

on diffusion of solar PV systems in a variety of countries, including both low and high

income economies.

3 Asterisks represent the proportion of relevance; three asterisks are for the highest relevance, while zero is for the

lowest.

10

Figure 1 presents the flow and interlinks of the chapters and sub-chapters of the PhD

Thesis. In this token, the rest of the Thesis is structured as follows. Chapter 2 outlines the

background information about solar PV systems. Chapters 3 explain the theoretical

constructs and analytical framework that are used at interpreting the data. In this chapter,

sub-chapter 3.1 outlines the theoretical approach and key constructs. Sub-chapter 3.2

presents the core results of systematic literature reviews (along with their methodological

aspects), which are based on solar PV systems in particular and environmental innovations

in general. Sub-chapter 3.3 outlines the analytical framework in respect to four

perspectives: micro-level, business models, meso-level and modeling. These sub-chapters

also present the state of the art for corresponding research streams. Furthermore, chapter 4

is dedicated to methodological aspects, e.g., research design, case selection, unit of

analysis, data collection and data analysis. This chapter is divided into three sub-chapters,

which explains the methods used in this Dissertation. The first one is case study

methodology (sub-chapter 4.1), which is used for micro-level and business model

analyses. The second one is indicator-based approach, which is used for meso level

analysis (sub-chapter 4.2), and the third one is finite element method (sub-chapter 4.3).

Anchored in the analyses at different levels, chapter 5 presents the discussions. The

discussions are divided into sub-chapters presenting the possible answers for the RQs

(sub-chapter 5.1-4) and acombined analysis (sub-chapter 5.5). Finally, chapter 6 presents

the overarching implications, conclusions, contributions and limitations of the PhD Thesis.

This last chapter also extends a detailed invitation for future research on diffusion of, in

particular, renewable energy innovations at grid parity and, in general, policy-driven

environmental innovations at cost competitiveness.

11

FIGURE 1. THE STRUCTURE OF THE PHD THESIS

13

2. SOLAR PV SYSTEMS

Generally, there are three different ways of generating energy from the sun: passive heat

energy that we receive naturally, solar thermal energy which provides us with hot water

(e.g., the case of solar thermal heating systems) and photovoltaic (PV) energy that we use

to generate electricity to run appliances and lighting (EPIA, 2010). Solar PV system is the

field of technology and research related to the devices which directly convert sunlight into

electricity; and the solar cell is the elementary building block of the photovoltaic

technology (EC, 2007). Solar cells can be made from different semiconductor materials

such as silicon which is the most widely used material for this purpose. There is no

limitation to silicon availability as a raw material as silicon is the second most abundant

material in the world (EPIA, 2010). Compared with fossil-fuel- based electricity

generation technologies, solar PV systems produce far less life-cycle air emissions.

According to Fthenakis et al. (2008), if electricity from photovoltaics replaces electricity

from the average grids in the US and Europe, it will lead to at least 89% reductions of

greenhouse gas emissions, pollutants, heavy metals, and radioactive species. These

pollution reductions are expected be even greater if the PV systems are decentralized

systems, such as the ones on the roof-tops of households, instead of centralized ones.

2.1. TECHNOLOGY

The history of PV systems goes back to the discovery of photovoltaic effect and the

discovery of the fact that sun's energy creates a flow of electricity in selenium in the 19th

Century (Petrova-Koch et al., 2009, p. 2). However the turning point was in the early

1950s, when the accidental discovery of Pearson et al. (1954), the scientists from Bell

Labs, took place (Kazmerski, 2006, p. 105). It changed the history of photovoltaic

systems. The discovery was using silicon instead of selenium. In a short period, the

scientists invented more efficient PV cells than selenium cells at generating electricity.

After this invention, the technology of PV has started to diffuse in different areas of use,

such as, satellites, cars and aircrafts (see Figure 2).

14

FIGURE 2. SOME HISTORICAL STEPS FOR PV SYSTEMS4

Nowadays, the majority of the electricity generated by solar PV systems are located on

residential and commercial lands. For example, a PV system on a residential roof-top,

consists of several parts, such as the solar PV panels (made of solar cells), mechanical and

electronic connections, and inverters (and sometimes batteries) (see Figure 3). The PV

systems are usually rated in peak kilowatts (kWp), which is the amount of electrical power

that a system is supposed to deliver under maximum solar exposure (Parida et al., 2011, p.

1626).

FIGURE 3. A SOLAR PV SYSTEM ON A ROOF-TOP5

4 Sources, from left to right.

1) http://www.corp.att.com/attlabs/reputation/timeline/54solar.html

2) http://nssdc.gsfc.nasa.gov/nmc/masterCatalog.do?sc=1958-002B

3) Schoolgen (2012) 4) SEIA (2010) 5 Source: http://www.sundialsolarnh.com/Residential-solar-services.htm

15

The main manufacturing processes (as well as the technology) of photovoltaic can be

currently categorized into mainly two groups: crystalline silicon technology (so called

wafer based) and thin film technology. There are also many different cell types developed

recently which are starting to be commercialized or still at the research level (EPIA,

2010). An efficiency progress graphic of all these PV technologies is shown in Figure 4.

The efficiency rates demonstrated in this chart are the best values reached by various

research & development institutions worldwide. They are laboratory prototype cells and a

couple of years may be needed to be commercialized. For example, the laboratory

prototype Crystalline Si Cells has already reached the level of 25% efficiency, but the

most of the Crystalline Si PV cells in the market have the efficiency rate under 20%. For

many new PV technologies reaching high level of efficiency rates, the main challenge is

the high costs needed for implementation at the commercial level.

16

FIGURE 4. BEST RESEARCH CELL EFFICIENCIES6

2.2. DIFFUSION

The solar PV systems count for less than 1% in the global share for electricity production

(see Figure 5). A recent report from European Photovoltaic Industry Association (EPIA,

2014a) shows the diverse PV diffusion patterns among different regions of the world (see

Figure 6). While some regions have a slowdown in their diffusion, the others are on the

way to achieving a rapid growth. Breaking the long-term domination of Germany, China

becomes the top market with its annual installations in 2013, followed by Japan and the

USA. Decline in annual installation in European countries was compensated with the

growth in China and Asia-pacific countries in the world. Yet, in terms of cumulative

installation capacities by 2013, Germany still tops the market with its 27% share on global

installations, doubling the followers China & Taiwan (13%) , Italy (12%) and Japan

(%10).

6 (NREL, 2014)

17

FIGURE 5. RENEWABLE ENERGY SHARE OF GLOBAL ELECTRICITY PRODUCTION7

FIGURE 6. ANNUAL INSTALLATIONS OF PV SYSTEMS8

Germany is usually viewed as the frontrunner country on the diffusion of PV systems. In

2012, solar PV systems generated almost 35% of current consumption on sunny days

(Wirth, 2013). Between 2004 and 2010, solar PV diffusion had performed the fastest

growth in Germany, far front of other renewable energy innovations. Figure 7 presents the

evolution of diffusion of solar PV systems in three market segments: small (less than 10

KWp capacity), medium (between 10 and 40 KWp capacity) and large (more than 40

KWp). In this figure, the times of implementation and amendment of German Renewable

7 Source: Ren 21 (2014, p. 25). This estimation is based on the end of 2013 as compiled from several sources as

explained in p.145 of Ren21 (2014). 8 Source: EPIA (2014a, p. 18). MEA: Middle East. APAC: Asia Pacific Countries. RoW: Rest of the world. The data for

RoW is not available for 2011 and 2012.

18

Energy Act (EEG) along with estimated time of the grid parity are also plotted. In terms of

number of installations, the small size PV systems have the largest diffusion in Germany.

FIGURE 7. NUMBERS OF YEARLY PV INSTALLATIONS IN GERMANY9

Although the PV systems achieved grid party in Germany, there has been a continuous

decrease in number of PV installations since 2011, which directly influenced the local

firms. Turnovers of many local PV firms have been decreasing, not only because of the

declined number of installations, but also because of the decreased price of PV modules.

However, the diffusion paths of solar PV installations are not the same all over the

country. Depending on the location, different diffusion trajectories can be observed. For

example, Figure 8 shows a district analysis from southern Germany. In this comparison,

Südliche Weinstrasse (a population of 108,875) has a stable growth in the total number of

installations during 2010-2012. However, Tübingen (a population of 212.800) had a

relatively rapid decrease in this period. In addition, Karlsruhe (a population of 424.510)

presented a rather slightly different pattern: sharp decrease in 2011 and stable growth in

2012. This illustrates that there might be different regional patterns of PV diffusion in

Germany.

9 The data is compiled from the Information Platform of four German Transmission Network Operators for the

Renewable Energy Sources Act (EEG) and the Combined Heat and Power Act (KWK-G)

19

FIGURE 8. NUMBER OF PHOTOVOLTAIC INSTALLATIONS IN THREE DISTRICTS10

2.3. GRID PARITY

The efficiency of the different types of solar cells is constantly improving (NREL, 2014,

2012). Moreover, from an economic perspective, the PV production cost has continuously

been decreasing. As a result of these developments, the relative economic advantage of

PV systems (as perceived by potential adopters) has been improved. Particularly in

Germany, PV systems are assumed to be at grid parity, i.e., the price of solar PV

electricity can compete with the price of conventional electricity sources (Lettner and

Auer, 2012a; Pérez et al., 2013; Spertino et al., 2014). How can the cost of solar PV

electricity be compared to the cost of conventional electricity sources? One way of doing

this comparison is based on the calculation of the levelized cost of electricity (LCOE),

€/kW. It is a calculation of the cost of electricity generation that is based on different

variables, such as the initial capital, solar radiation, costs of continuous operation, service

life time and costs of maintenance. When the LCOE of solar PV electricity is below the

price of purchasing electricity from the grid, it means that solar grid party has been

achieved in the corresponding region. The comparison of the LCOE of a PV system and

10 The data is compiled from the Information Platform of four German Transmission Network Operators for the

Renewable Energy Sources Act (EEG) and the Combined Heat and Power Act (KWK-G). The Germany map is plotted

via the QGIS 2.6.0 Brighton Software.

20

the average electricity price in Germany indicates that since the beginning of 2012 the

LCOE of a PV system has been lower than electricity retail price (Figure 9). By the end of

2012, the LCOE of a typical PV system in Germany was between 0.12 and 0.21 €/kWh,

whereas the electricity retail price was approximately 0.26 €/kWh. This fact represents a

rapid improvement in solar PV systems in comparison to May 2010. As a consequence of

this decrease of the PV LCOE and the increase of the electricity prices in Germany, the

government has been gradually reducing the feed-in tariff which was firstly introduced

through the German Renewable Act (Erneuerbare-Energien-Gesetz, EEG) in 2000. The

idea behind EEG was to accelerate cost reductions on renewable energy technologies

through economies of scale over time. It has been applied to all states in Germany,

supporting PV adopters with 10-year contracts and pays them with a cost-based price for

the kW. As a result of the rapid decrease in feed-in tariff in Germany, the adoption rate

has recently experienced some boom and bust cycles, as presented in Figure 10. The

number of installations notably increased just before the reductions of the feed-in tariff as

seen in the months of December 2009, July 2010, December 2010 or July 2011. In

contrast, since April 2012, the feed-in tariff has been diminishing gradually month-by-

month, which prevents boom and bust cycles but reveals seasonal effects (e.g., high

installations between June and October due to adequate weather conditions in the region).

21

FIGURE 9. COMPARISON OF ELECTRICITY GENERATION COSTS11

FIGURE 10. MONTHLY INSTALLATIONS VS. FEED-IN TARIFF IN GERMANY12

11 The data are compiled from different sources (Kost and Schlegl, 2010; Kost et al., 2012) and Bundesagentur. (as cited

in Karakaya et al., 2015) 12 The date are compiled from Bundesagentur and the Information Platform of four German Transmission Network

Operators for the EEG and KWK-G

23

3. DIFFUSION OF INNOVATIONS

3.1. THEORETICAL APPROACH

The theory in science answers questions as to why certain phenomenon might occur

(Berthon et al., 2002, p. 421). In social science research, an analytical theory provides

several elements, such as criteria about which are important to consider, gaps in the

existing knowledge and a source of cross fertilization of related fields (Parsons, 1938).

This chapter, therefore, aims to outline the important perspectives of related fields that are

known to be critical for understanding the dynamics of diffusion of renewable energy

innovations. Subsequently, this chapter serves as an analytical basis, i.e., so called

“theoretical lenses”, to interpret the empirical data used in this Dissertation.

Innovation was traditionally defined as setting up of a new production function, which

covers the cases of a new commodity for the opening up of new markets and new forms of

organizations (Schumpeter, 1939). The innovations are the core of technological change13

in societies. The Schumpeterian approach has been developed by several researchers,

furthering our understanding of innovation with various taxonomies. Starting from the

1960s, several contributions in innovation studies have emerged from different disciplines.

Seminal contributions have been made in the fields of economic sciences (Nelson, 1959;

Schmookler, 1966), management sciences (Burns, T., Stalker, 1961) and sociology

(Rogers, 1962), while Nelson & Winter, (1982) have been accepted as the most important

contributors in innovations studies with their book "An evolutionary Theory of Economic

Change", which combines Schumpeterian and evolutionary perspectives with implications

to innovation (see Fagerberg et al., 2012). In the 1990s new concepts and frameworks (as

well as synthetic overviews) have been introduced into the field. For example, the concept

of Porter (1990) about fostering innovation and growth has rapidly been popularized

especially by analysts and policy makers discussing the importance of location. Kline &

Rosenberg (1986) proposed a chain-linked model of innovation that triggered the

13 According to the Schumpeter, the technological change has three stages: invention, innovation and diffusion (Jaffe et

al., 2002)

24

systematic approaches and their implications that have gained more attention in the last

decade. The concept of "National Systems of Innovation" has been developed on different

perspectives and geographical levels such as sectoral innovation systems (Carlsson,

Jacobsson, Holmen, & Rickne, 2002), technological innovation systems (Malerba, 2002)

and regional innovation systems (Asheim and Coenen, 2005).

Diffusion of innovations theory (Rogers, 1962), in contrast to other conceptualisations of

innovation, is written from a sociological perspective and focuses on the process at which

innovations are diffused and adopted by individuals within a social system. It has been

studied and applied in various academic disciplines (Murray, 2009; Sriwannawit and

Sandström, 2015) including marketing (e.g., for projections of markets) and public health

(e.g., for understanding how new standards can be translated into widespread practice).

However, Rogers' approach is relatively independent in compare with other top innovation

studies' contributors (who are Nelson, Freeman, Rosenberg, Schumpeter, Porter, Griliches,

Von Hippel, Lundvall, Pavitt and Chandler according to Fagerberg et al., 2012). This

different approach can be explained with three possible factors. First, Rogers is a rural

sociologist but many other contributors in innovation studies are mainly from economics

or management sciences. Second, he analyses the adoption of innovations by demand side.

Other scholars usually focus more on the supply side mechanisms and industrial

dynamics, for instance institutional changes, market formation or entry of new firms.

Third, Rogers' theory is empirically deduced from many case studies.

Environmental innovations are the innovations that avoid or reduce environmental harms

(Beise and Rennings, 2005), i.e., the innovations that reduce the use of natural resources

and decrease the release of harmful substances (EIO, 2010). The environmental innovation

concept has been studied by a variety of scholars (e.g. Jänicke, 2008; Oltra and Saint Jean,

2009; Pujari, 2006; Rennings, 2000; Van den Bergh, 2013a). The concept covers many

technologies (e.g., wind energy systems, algae biodiesel), organizational practices (e.g.,

pollution prevention schemes) and services (e.g., car sharing). Environmental innovation

activities are usually driven forward or hampered by various factors both internal and

external to company, among others: economic and financial factors (e.g., pricing, market

position, access to capital, demand) and policy framework (including environmental and

innovation polices, taxes, standards and norms) (EIO, 2011). Yet, policy makers also have

paid much attention to environmental innovations. For example in European Union (EU),

environmental innovations are one of the most crucial aspects of Europe 2020 strategy

25

(EIO, 2010). With environmental innovations, the EU aims to achieve two important

goals: a resource-efficient Europe and an industrial policy based on green growth. It is

assumed that the emergence of environmental innovations will ensure future employment,

contribute to economic growth in Europe and respond to today’s major societal and

environmental challenges (ETAP, 2010).

Renewable energy innovations are typical examples of environmental innovations (Huber,

2008, p. 361). Renewable energy innovation is any renewable energy related idea, object

or practice that is perceived as new by the adoption unit. The term “renewable energy

innovation” is often used by scholars studying the renewable energy technologies, such as

solar PV systems, wind power, biomass and solar thermal (e.g. de Araújo and de Freitas,

2008; Huijben and Verbong, 2013; Mallett, 2007; Wüstenhagen et al., 2007). Renewable

energy innovations, and their diffusion, bring some particular aspects to the debate.

Wüstenhagen et al. (2007) analyze these aspects in three dimensions. Firstly, for the

adoption of renewable energy innovations, more than one decision-maker approval might

need to be taken. For example, this is a typical case of multi-family residential buildings

willing to adopt solar PV systems. Secondly, the adoption of renewable energy

innovations have a visual impact on both the adopters and non-adopters. This is mainly

because the energy production usually happens where the energy users live. However, for

conventional electricity sources such as nuclear or coal, the production units are not easily

visible to energy users. Thirdly, renewable energy innovations have high short-term costs.

The term “diffusion” lies on the core of this Dissertation, and correspondingly in all

analyses. Diffusion is the process by which an innovation is disseminated amongst

potential adopters (Teece, 1980, p. 464). It has been popularized since the seminal work of

Rogers (1962), who is accepted as one of the most well-known scholars of innovation

studies (see Fagerberg et al., 2012). In general, the development and diffusion of

innovations cannot be planned. It is complex, uncertain and subject to changes of many

sorts (Kline and Rosenberg, 1986). Some innovations require a lengthy period to the time

they are adopted even though the obvious advantages (Rogers, 2003). History is full of

examples in which innovations were not been embraced despite their clear benefits or the

adoption period took many years (e.g., electrical cars). Even when it comes to

environmental innovations, which the EU promotes, there is a lot to be desired for their

rapid diffusion. These kinds of adoption problems, which have occurred or may occur in

future, could de-motivate potential innovative companies, mainly the small and medium

26

ones (SME). The results of Eurobarometer survey (EC, 2010) confirms that European

companies think that two of the most significant barriers to environmental innovations are

directly related to adoption rate of their products: uncertain demand from the market and

uncertain return on investment.

3.2. STATE-OF THE ART

This sub-chapter includes two research state-of the art literature reviews. These reviews,

which are about eco-innovations in general and solar PV systems in particular, are

conducted through systematic literature review approach (see Figure 11). Systematic

literature reviews are scientific methodologies that are able to limit bias to systemic

assembly, critical appraisal, and the synthesis of all relevant studies on a specific topic

(Cook et al., 1995, p. 167). According to Cook et al. (1997), systematic literature reviews

distinguish from traditional narrative reviews by adopting a replicable and detailed

methodology (as cited in Tranfield et al., 2003, p. 209). Recently, systematic literature

reviews has become a fundamental scientific activity, for example, as a common part of

PhD dissertations or, even sometimes, the whole dissertation itself (Daigneault et al.,

2012).

27

FIGURE 11. THE COMPARISON OF TWO LITERATURE REVIEWS

The first systematic literature review focuses on the emerging field of diffusion of

environmental innovations. It maps the scientific body, identifies the core contributors and

discusses the main research streams. It is based on a review of all kind of publications

(from 1990 to 2012) of Google Scholar Database, which is a freely accessible scholarly

web search engine that includes full-text journal articles, technical reports, preprints,

theses, books, and other scholarly documents (Rita Vine, 2006). Although this database

has recently been criticized by many scholars because of its shortcomings on bibliometric

purposes (Aguillo, 2011; Jacsó, 2011), it is used for bibliometric studies mainly because

of its broader coverage. Due to its comprehensive citation coverage, the citation numbers

of Google Scholar are usually around four times higher than ISI Web of Knowledge. For

example, the study of Jaffe and Palmer (1997) had a total citation of 142 in ISI Web of

Knowledge and 589 in Google Scholar; the study of Nill and Kemp (2009) had a total

citation of 36 in ISI Web of Knowledge and 113 in Google Scholar.

The identification of keywords was a challenging step. There are many synonyms that

refer to the notion “diffusion of eco-innovations”. To address this, the synonyms of both

“diffusion” and “eco-innovations” are separately identified to create multiple genitive

constructions. Firstly, “adoption” is identified as the only synonym of diffusion due to the

fact that these two terms are traditionally studied together (Halila, 2007; Rogers, 2003;

28

Surinach et al., 2009). The only difference is that adoption refers to a process at the

individual level whereas the process of diffusion occurs in society. Secondly,

“environmental innovations”, “ecological innovations”, “green innovations” and

“sustainable innovations” are taken as synonyms to “eco-innovations” as suggested by

Schiederig, Tietze, & Herstatt (2012). According to these premises, the genitive

constructions of the synonyms of the diffusion and eco-innovations were created as 40

variations (see Table 2). This approach limits the database and excludes the studies that do

not include any of these 40 keywords. Therefore, the studies that about only one type of

eco-innovations (e.g. solar panels, green textile or photovoltaic technology) would not

appear in the database unless they have one of these 40 keywords in their text. In this

review, the extracted publications were analysed through two-level analysis. The first-

level analysis is based on the information of the authors, journal names, publication years

and citation frequency. The second-level analysis focuses on the identification of the core

disciplines and research streams based on the mostly cited articles. The second-level

results of this first review are presented in sub-chapter 3.2.1.14

14 The results of first-level analysis are briefly presented in this Thesis. For further details on this, please see Karakaya et

al. (2014).

29

TABLE 2. KEYWORDS USED FOR COMPILING THE DATABASE FROM GOOGLE SCHOLAR.

Diffusion Category Adoption Category

Eco-Innovation

notion

Diffusion of Eco-innovation Adoption of Eco-innovation

Diffusion of Eco-innovations Adoption of Eco-innovations

Eco-innovation Diffusion Eco-innovation Adoption

Eco-innovations Diffusion Eco-innovations Adoption

Ecological

Innovation

notion

Diffusion of Ecological Innovation Adoption of Ecological Innovation

Diffusion of Ecological Innovations Adoption of Ecological Innovations

Ecological Innovation Diffusion Ecological Innovation Adoption

Ecological Innovations Diffusion Ecological Innovations Adoption

Green

Innovation

notion

Diffusion of Green Innovation Adoption of Green Innovation

Diffusion of Green Innovations Adoption of Green Innovations

Green Innovation Diffusion Green Innovation Adoption

Green Innovations Diffusion Green Innovations Adoption

Sustainable

Innovation

notion

Diffusion of Sustainable Innovation Adoption of Sustainable Innovation

Diffusion of Sustainable Innovations Adoption of Sustainable Innovations

Sustainable Innovation Diffusion Sustainable Innovation Adoption

Sustainable Innovations Diffusion Sustainable Innovations Adoption

Environmental

Innovation

notion

Diffusion of Environmental Innovation Adoption of Environmental Innovation

Diffusion of Environmental Innovations Adoption of Environmental Innovations

Environmental Innovation Diffusion Environmental Innovation Adoption

Environmental Innovations Diffusion Environmental Innovations Adoption

The second systematic literature review focuses on the empirical context of the diffusion

of PV systems (see Table 3). It systematically reviews the articles that were published

during 2011-2013 in the Social Sciences Citation Index of Web of Science Core

Collection. While the first systematic literature review is an important guideline to build

the theoretical framework of this Thesis, both systematic reviews serve as information

source for generating the theoretical and empirical implications of the PhD Thesis. For the

keywords of the second review, we identified a combination of keywords as follows:

(diffus* OR adopt*) AND (photovolt* OR PV OR SHS* OR "solar home system*"). This

set of keywords was searched in the abstracts, titles, and keywords of the publications in

the SSCI from 2011 to 2013, yielding 103 publications. We have also included the term

solar home system and its acronym SHS because PV systems are sometimes referred to as

30

solar home systems, especially in rural contexts. The core results of the second review are

presented in sub-chapter 3.2.2. In addition, the rest of the literature which focuses on the

case of solar PV systems in Germany, but not mentioned in systematic literature reviews,

is outlined in sub-chapter 3.2.3.

31

TABLE 3. OVERVIEW OF REVIEWED PUBLICATIONS BY COUNTRY15

Category Country Studied by

Low-income

economies

Bangladesh (Komatsu et al., 2011; Palit, 2013; Pode, 2013)

Cambodia (Pode, 2013)

Ethiopia (Müggenburg et al., 2012; Pode, 2013)

Kenya (Bawakyillenuo, 2012; Lay et al., 2013; Ondraczek,

2013; Pode, 2013)

Nepal (Mainali and Silveira, 2011; Palit, 2013)

Tanzania (Ondraczek, 2013; Pode, 2013)

Zimbabwe (Bawakyillenuo, 2012)

Lower-middle-

income

economies

Bolivia (Pansera, 2012)

Ghana (Bawakyillenuo, 2012; Pode, 2013)

India (Palit, 2013; Pode, 2013)

Indonesia (Blum et al., 2013; Pode, 2013)

Lao PDR (Pode, 2013)

Nicaragua (Rebane and Barham, 2011)

Senegal (Thiam, 2011)

Sri Lanka (Palit, 2013)

Upper-middle-

income

economies

China (D’Agostino et al., 2011; Yuan et al., 2011)

High-income

economies

Austria (Brudermann et al., 2013; Koinegg et al., 2013)

Canada (Islam and Meade, 2013)

Germany (Huenteler et al., 2012)

Greece (Karteris and Papadopoulos, 2013)

Hong Kong (Zhang et al., 2012)

Italy (Ameli and Kammen, 2012)

Japan (Huenteler et al., 2012; Vasseur et al., 2013; Zhang et

al., 2011)

South Korea (Jeong, 2013)

Spain (Movilla et al., 2013)

The

Netherlands (Vasseur et al., 2013)

The UK (Gooding et al., 2013)

The US

(Drury et al., 2012; Gaul and Carley, 2012; Kwan,

2012; Rai and Robinson, 2013; Sarzynski et al., 2012;

Zhai and Williams, 2012; Zhai, 2013)

15 In 2013, the World Bank categorized countries by GNI per capita. This income categorization is as follows: low

income ($1,035 or less), lower middle income ($1,036 to $4,085) upper middle income ($4,086 to $12,615) and high

income ($12,616 or more) (WB, 2014).

32

3.2.1. ECO-INNOVATIONS

“Diffusion of eco-innovations” has recently become a common term in different scientific

communities. In the period 1990-2012, the number of publications that include the

synonyms of “diffusion of eco-innovations” reached a total of 1024 publications. Figure

12 shows that very little research on diffusion of eco-innovations was conducted in the last

decade of the last century. However, since 2000 the term has been used widely. Starting

from 2006 the relevant articles have strongly grown, as evidenced by 36% of these

publications were published in the period of 2008-2011. Furthermore, yearly publications

in the literature represent a growth of 64% from 2010 to 2011. However, and surprisingly,

only 18% of these publications cited Rogers (183 of 1024 publications). This trend was in

a similar manner in the last years: 24%, 14% and 22% of yearly publications in 2009,

2010 and 2011 cited Rogers respectively.

FIGURE 12. THE EVOLUTION OF RESEARCH ARTICLES RELATED TO "DIFFUSION OF ECO-INNOVATIONS"

The studies on diffusion of eco-innovations have particular research focuses on different

topics, such as the impact of environmental regulations on firm’s performance (Jaffe and

Palmer, 1997), performance measurement of environmental supply chain management

(Hervani et al., 2005) or impact of corporate social performance on firm financial

performance (Hull and Rothenberg, 2008). That said, related to diffusion of eco-

innovations, it is possible to identify some interdisciplinary research streams within three

disciplines (economics, sociology and management) and two traditional research fields

(marketing and agent based modeling).

0

20

40

60

80

100

120

140

160

180

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

2010

2012

Years

Nu

mb

er

of

Pu

bli

cati

on

s

Total Publications

that use the term

"Diffusion of eco-

innovations" and its

synonymsThose of that refer to

works of Everet

Rogers

33

Economics

The role of innovation in economic and social change is one the main focuses of

innovation studies, and it is especially analyzed by scholars with a background in

economics (Fagerberg et al., 2012). However, from an economic perspective, theoretical

and methodological approaches to understand eco-innovations are diverse: e.g., innovation

economics, environmental economics, ecological modernization, sustainable transitions

and lead markets. Innovation economics focus on entrepreneurship and innovations. This

research stream often studies the positive spillovers of both firms’ R&D activities and

diffusion phase of innovation (Rennings, 2000, p. 325). The environmental economics is

concerned with environmental and economical issues that are related to governmental

policies. The approach of ecological modernization (Jänicke, 2008) focuses on innovation

oriented environmental policy (regulations, incentives, policies, actors, etc.) and the

supply-side driving forces of diffusion of eco-innovations with the underpinnings of

sociological theory. The research stream of sustainability transitions focuses on long-term,

multi-dimensional and fundamental transformation processes through which established

socio-technical systems (sectors like energy supply, water supply, or transportation) shift

to more sustainable modes of production and consumption (Markard et al., 2012).

Sustainability transitions are about interactions between not only technology, policy and

economics but also culture and public opinion (Geels, 2011); therefore sustainability

transition can provide insights about the diffusion process of eco-innovations. Finally, the

lead market model focuses on the countries that are first in adopting an internationally

successful innovations and it can be applied for understanding the link between strict

regulation and creation of lead markets (Beise and Rennings, 2005). It potentially gives

insights about the dynamics of international diffusion of eco-innovations.

Sociology

The theory of diffusion of innovations (Rogers, 2003) might shed light on understanding

the social norms of adopters and the social values of societies affecting the diffusion of

eco-innovations. That said, this theory is not necessarily used by many scholars. There are

several studies on eco-innovations which are based on a variety of sociological

approaches. For example, the study of Spaargaren (2003) introduces an interesting

approach from environmental social sciences. This study brings the issues of sustainable

consumption and lifestyles by using the social practices model of sociology. The proposed

34

model indicates that diffusion of eco-innovations through citizen-consumers can be

increased when daily routines (clothing, food, shelter, travel, sport, and leisure) are taken

as a starting point for policymaking. On the other hand, the study of Ozaki (2011)

combines the theories of diffusion of innovations, cognitive and normative behaviour, and

consumption in order to understand the motivation factors of eco-innovation consumers by

focusing on the five factors: perceived benefits, social influence, perceived compatibility,

controllability and perceived uncertainty.

Management

The diffusion term in management of innovation is usually used for diffusion of

environmental innovations within the organizations, especially for environment friendly

practices. The adoption of Environmental Management Systems (EMS), a total quality

approach on management of organization's environmental programs, has taken much

attention of scholars. On the one hand, Florida and Davison (2001) discuss the factors that

make firms to adopt EMS, and Frondel et al. (2008a) research on the role of EMS for the

adoption of technical eco-innovations in firms. On the other hand, Wagner (2008) focuses

on the impact of EMS and environmental managerial activities on the probability of firms

to carry out eco-innovations.

The field of research in entrepreneurship, relevant for the management discipline, also

contributes the understanding for eco-innovations. In this field, the study of Hockerts and

Wüstenhagen (2010) reveals the relation between firm size and diffusion of eco-

innovations, theorizing the interplay between incumbents and new entrants.

Marketing

After the seminal work of Bass (1969), that presented the first purchase growth of a new

durable product in the market, the field of marketing has embraced a common research

tradition for diffusion of innovations. However, the study of Ottman et al., (2006) is one of

the few studies that focus on the concept of diffusion of eco-innovations. These authors

discuss on improved environmental quality and customer satisfaction in the concept of

green marketing. They indicate that three green marketing principles are important to

make green products desirable for consumers. These are consumer value positioning,

calibration of consumer knowledge, and credibility of product claims.

35

Agent Based Modeling

Agent based models refer to the models which consist of a set of agents that encapsulate

the behaviours of the various individuals that make up the system (Parunak and Savit,

1998). This research tradition, within the growing computational power, has been

increasingly applied to social and economic problems which were previously modeled

with nonlinear differential equations (Rahmandad and Sterman, 2008). Evolutionary

economics’ principles can also be formalized in the framework of agent based modeling as

well (Faber and Frenken, 2009).

There are many studies on agent based modeling of diffusion of eco-innovations. For

example, the study of Schwarz and Ernst (2009) is an empirical approach on agent-based

modeling of diffusion of environmental innovations. The implementation of this model is

based on differentiating individuals, communication, innovation characteristics and

decision algorithms. The study of Barthel et al. (2008) makes use of a multi-actor

simulation framework by simulating decision-making process of the relevant actors, while

Cantono and Silverberg (2009) develop a model of innovation diffusion that combines

contagion among consumers with heterogeneity of agents.

3.2.2. BARRIERS TO DIFFUSION OF SOLAR PV SYSTEMS

The barriers to the diffusion of solar PV systems could be categorized in four dimensions:

sociotechnical, management, economic and policy. As is widely known in diffusion

literature (e.g., Rogers, 2003), the adoption of innovations, such as renewable energy

technologies, is a complex process with several connected factors. Therefore, the

dimensions of these barriers are also interrelated and our categorization does not aim at a

rigid taxonomy.

Sociotechnical barriers

The quality of PV systems is of vital importance for adoption. It can be influenced by not

only the local conditions of the user’s environment (Müggenburg et al., 2012) but also the

political and financial arrangements (Palit, 2013) that may change from country to

country. A concrete example of such a phenomenon is given by Palit (2013). He mentions

that Bangladesh and India have better quality standards for PV systems, especially for

battery performance, than Sri Lanka. In China, there is a high level of dissatisfaction with

36

the low performance of SHS. Although such performance may be caused by not only

technical functionality but also improper usage, the dissatisfaction works against other

potential adopters purchasing PV systems. In addition, damaged PV systems were

common, which is partly due to having repair locations too far from the adopters

(D’Agostino et al., 2011). In Ethiopia, there is growing skepticism toward products

manufactured in Asia, especially those from China. Customers mistrust the goods and do

not want to purchase them. Instead, they prefer other products, even with a higher price

(Müggenburg et al., 2012).

The lack of adequate knowledge among both adopters and non-adopters is a crucial

barrier. Such lack of knowledge by adopters may result in improper usage and inability to

maintain the systems, as shown in China. This may create a negative perception and

prevent potential customers making a decision to adopt the systems (D’Agostino et al.,

2011). In the case of Nicaragua, rural inhabitants know too little about PV systems, which

can be one of the barriers for adoption (Rebane and Barham, 2011). It has also been shown

in previous research that the diffusion of new technologies in rural communities requires

more effort in tailoring the information and educating people (Sriwannawit, 2012). In

another case in the US, there is a lack of trust in the information that is widely and

publicly available among some adopters. They feel uncertain about technology

performance and lack information that is relevant to their individual cases (Rai and

Robinson, 2013). A study in Austria (Koinegg et al., 2013) also revealed that lack of

knowledge is not only a matter of adopter-side but also the supply-side. They argue that

architects and planners often have insufficient knowledge about the advantages of building

integrated PV systems. This is the reason why they do not offer such systems to the

potential adopters when planning a new building.

The perception of adopters has a large impact on their decision whether to adopt a new

technology or not. This is because the way that adopters perceive the complexity of

technology can impact their decision on adoption (Davis, 1989; Venkatesh, 2000). Several

studies investigated the diffusion process from the end-user’s point of view. A study in the

US showed that the residents were reluctant to adopt PV systems because of technology

risk and complexity. These barriers are, however, reduced when third-party ownership is

implemented (Drury et al., 2012). For areas that are already connected to the grid, the

perceived maintenance requirement can also be a barrier to adoption. In this case, the

customers have a choice to use electricity from the grid, which would not impose on them

37

any responsibility to operate and maintain the technology, unlike having their own PV

panels (Zhai and Williams, 2012). A survey study in China (Yuan et al., 2011) indicated

that the respondents perceive PV systems as having a low level of application. This can be

related to the lack of technology awareness. Some users are also concerned about the

complexity of the system, technological maturity, durability, efficiency, safety, and

stability. In addition, there are some attributes of PV systems that are perceived negatively

by the adopters. These are, for example, low battery storage and low power capacity.

Solar exposure is a basic requirement for functioning PV systems. Because the solar

exposure depends on the geographic location, so does the competitiveness of PV systems.

For example, the US energy portfolio varies across the country, making solar PV systems

more technically feasible in some particular areas. This is due to large variation in climatic

condition because the US covers a huge geographic area (Sarzynski et al., 2012). In the

areas without other sources of electricity in Ethiopia, adopter’s fear of non-functional

systems during the rainy season is also a barrier preventing their purchase (Müggenburg et

al., 2012). In some areas, PV systems face challenges from other competing technologies

that are more suitable to their geography. A study in China mentioned that the increasing

access to hydropower, as a means to provide electricity access in rural areas, has a

negative influence on the adoption of SHS (D’Agostino et al., 2011).

Society specific factors can also have an impact on the adoption. A comparative study

between Kenya and Tanzania by Ondraczek (2013) is an example of such a phenomenon.

The lack of awareness among users and other stakeholders in Tanzania is one of the

barriers to adoption. This lack of awareness is partly because of the geographical distance

from big cities to rural villages, which makes firms less willing or unable to reach the

market. Moreover, compared with the neighboring Kenyans, Tanzanians tend to be more

reluctant and skeptical toward new technologies. The author further explained that

education is another contributing factor, hindering the adoption of new technologies.

Because of its socialist history, Tanzania is assumed to lack an entrepreneurial culture.

There is also high complexity to doing business in Tanzania. In China (Yuan et al., 2011),

the lack of public awareness is apparent as a barrier to adoption. The Chinese are more

aware of solar water-heater technology than solar PV technology. This affects their

decision making on the adoption of PV systems in a negative way.

38

The architectural dimension of the areas is also an important factor that can become a

barrier to adoption. For urban areas like Hong Kong, a key barrier is an inadequate

installation space. PV panels need to be angled toward the right direction to maximize

solar exposure. For rooftop integration in Hong Kong, the surface is simply too limited in

old high-rise buildings. New buildings can, however, be designed to integrate PV systems

in their structure to maximize the installation space (Zhang et al., 2012). The physical

aspect of the city in terms of local building stock also has an impact on how much

electricity can be generated from rooftop PV systems. In several Scottish cities, tenements

are the preferred type of housing. The limited roof space on tenements results in less

electricity output compared with the output from cities that have large properties. This is a

limitation that is difficult to address by a policy mechanism (Gooding et al., 2013).

In the case of Bolivia as studied by Pansera (2012), the institutional dimension has been

one of the key barriers for rural applications. Bolivia relies substantially on foreign

financial aid. The country cannot easily design and execute renewable energy policy

without international cooperation. However, despite the strong involvement, the

international actors pose limitations. First of all, they lack the methodology to identify

local needs. They also lack technical, human, and financial resources to tackle big projects

in rural areas, which means that they are unable to achieve high impact. On the other hand,

the private sector has little interest in this market. Missing a systematic vision of the local

situation, these foreign actors want to implement one solution in all contexts, which is not

possible. In several cases, the international counterparts also reveal political vested

interests, e.g., establishing their products in a new market or acquiring privileged access to

local natural resources. Moreover, local education, i.e., university research, is neglected,

which prevents local capability building. In the case of Senegal studied by Thiam (2011),

the institutional dimension has also been identified as a barrier. Institutional reliability can

encourage or discourage private sector involvement to create dynamics for them to utilize

skills and knowledge in this market. The institutions also help increasing collaboration

among related actors to work together to achieve larger diffusion. In Senegal, this is not

yet the case.

Another barrier can arise from the applicability of PV technology to individual cases. One

study conducted in Ghana showed that the capacities of the system are too high for the

affordability of target adopters. Only high wattage systems are available in the market and

they are sold only as a whole package. This is too costly for low-income adopters

39

(Bawakyillenuo, 2012). In the lower-income markets, it has been shown that repackaging

the products in small portions can facilitate the sale because it matches with their

affordability (Prahalad, 2010).

Without demand, diffusion will be impeded. In some remote areas, lack of demand for

electricity is a barrier to the adoption of PV systems. Here, we distinguish the reasons for

the lack of demand into two types. First, it is due to the lack of electricity-related

activities. In the low-income markets, some people have recently started to use household

electrical appliances. For example, in the case of Tanzania, TV signals came rather late

(Ondraczek, 2013). In Ethiopia, there are not many additional activities during night time

that its inhabitants can do with the help of electricity access because most people are

illiterate (Müggenburg et al., 2012). Second, it is due to the lack of customer base. In

Tanzania, the market lacks customers from the rural middle and upper class who have

purchasing power for PV systems and at the same time need electricity (Ondraczek, 2013).

Without demand, the market cannot grow. In addition, lack of demand may arise from

different demographic factors. Some studies showed that low education levels, which can

be correlated with income level (Lay et al., 2013), and low purchasing power, which can

be correlated with age groups (Kwan, 2012), negatively influence the adoption of PV

systems.

Management barriers

Insufficient and inappropriate management is one of the main barriers in the diffusion of

new technology, not least for PV systems, especially when they are used in rural contexts.

One of the main management barriers is the inappropriate company business portfolio for

the target market. When a PV system is utilized to supply electricity access in rural areas

in low-income economies, different business strategies should be implemented (see e.g.,

Palit, 2013; Pode, 2013) compared with the high-income economies where it is often used

as an alternative power supply. In addition, implementing similar business models as for

urban usage is not applicable in rural settings because of several conditions that

differentiate this market from the higher income adopters (Prahalad, 2010). Appropriate

financial schemes are required for the low-income markets. These are, for example, fee-

for-service and microcredit (Pode, 2013). Financial schemes are important, as discussed

previosly, and the companies should consider such schemes when entering the PV market.

In addition, the study by Koinegg et al. (2013) also showed that effective and appropriate

40

business strategies are vital to prompt diffusion in particular markets. Based on building

integrated PV market in Austria, they argue that lack of close collaboration between the

building industry and the PV industry hindered take off in the market. This means that,

unlike the conventional type of businesses in the PV industry, the system suppliers in

building integrated PV sector need to develop a different type of business model that

collaborates with more actors.

Weak and neglected after-sales service has also been shown to be a barrier to the diffusion

of PV systems in rural areas. Because of the remoteness of the areas where PV systems are

being used for rural electrification, the adopters lack access to information, knowledge,

communication channels, technical assistance, and other infrastructure. This emphasizes

the need for a functioning service to continue monitoring and maintenance, even after the

customers have already bought the systems (Palit, 2013; Pode, 2013). The need for this

service, however, poses a challenge on cost and manpower to manage and guarantee the

sustainability of the system (D’Agostino et al., 2011; Pansera, 2012). In the case of

Bolivia, rural inhabitants are very remote and relatively poor. When the systems are

broken, they rather let them go instead of getting them repaired because this requires a lot

of effort and money (Pansera, 2012). After-sales service models need further development

to deal with the remoteness of the areas and the low-income level of rural adopters

(D’Agostino et al., 2011). The problems with service can impair the quality of PV

systems, which is another barrier to adoption as already discussed in the previous section.

Some other barriers are related to ineffective marketing approaches and education

campaigns. For example, Ghanaian firms have not attempted to target the right customer

group, i.e., affluent rural dwellers. While possessing sufficient affordability, these people

do not have access to electricity. This strategy has been proven to be successful in Kenya

and Zimbabwe. Establishing an early adopter customer base creates visibility that

functions as a catalyst for diffusion in rural markets (Bawakyillenuo, 2012). In Canada,

Islam and Meade (2013) argue that there should be education campaigns to communicate

more information regarding the investment in PV systems, its environmental effect, and

the feed-in-tariff. Otherwise, the barrier of low technology awareness cannot be overcome.

A national mechanism has an impact on the diffusion of technology. Existing national

technical capacity and infrastructure can affect the diffusion of PV systems. This

phenomenon is more evident among developing nations. In order to promote wide

41

adoption, countries need to build technical capacity. In comparison with Kenya and

Zimbabwe, Ghana has fewer PV companies and technicians. There are also fewer

subsidiary institutions to support the PV industry, which impedes adoption of the

technology (Bawakyillenuo, 2012). In Tanzania, a national supply chain is difficult to

undertake because of the inadequate transport infrastructure. This leaves remote areas

underserved and the PV market concentrates around big cities (Ondraczek, 2013). A study

comparing Japan and the Netherlands by Vasseur et al. (2013) showed that Japan has a

better technological innovation system to support the PV industry. In the Dutch case, there

is a lack of collaboration and knowledge exchange between researchers and policy makers

with adopters. The Dutch case also lacks human and financial resources to stimulate the

PV market. Lastly, Thiam (2011) argues that in the electricity market, fossil fuels are

much more mature than PV technology. This prevents PV systems from being an

attractive choice for customers. However, renewable energy tariffs that are set by the

government can help in overcoming this immature market barrier. This national level of

the barrier is highly associated with policy aspects that we discuss further under in the

sub-section of policy barriers.

Economic barriers

The adoption of PV systems faces several economic barriers. These barriers can be in

different forms, based on both time and location. As is known in the literature on

innovations, the cost of an innovation usually decreases with time (e.g., Spence, 1981) and

can vary depending on the location (Karakaya et al., 2014b).

Economic barriers are usually related to the high cost of solar PV modules. The diffusion

of PV systems is also affected by the cost of other energy sources in the region (Sarzynski

et al., 2012) because the potential adopters might have to choose between PV systems and

conventional sources of energy. If the costs of competing sources are low, these can

constitute a barrier to PV adoption. The lower the installation cost of PV systems, the

more likely that people will adopt them. As a result, high costs for investment in PV

systems are often perceived as a barrier to adoption, e.g., in China (Yuan et al., 2011) and

Japan (Zhang et al., 2011). The high costs of PV installation are mentioned in many

studies that are based in several countries, such as South Korea, Greece, and the US (Gaul

and Carley, 2012; Jeong, 2013; Karteris and Papadopoulos, 2013; Sarzynski et al., 2012).

Examples are spread over both off-grid, e.g., in Ethiopia (Müggenburg et al., 2012) and

42

Senegal (Thiam, 2011) and on-grid applications, e.g., in Italy (Ameli and Kammen, 2012)

and the US (Zhai and Williams, 2012). These discussions also include third-party owners

(Drury et al., 2012). In addition, the perception on the high cost of PV systems can vary.

Koinegg et al. (2013) argue that the perceived cost of PV systems can be a barrier to

adoption. For building integrated PV systems, the cost is not necessarily as high as

commonly perceived.

The study by Karteris and Papadopoulos (2013) shows how the barriers can be bound to

the particular local situation. Based on the Greek case under economic crisis, the authors

mention the unwillingness of banks to fund medium- or long-term investments, including

investments for PV systems. In addition, the shrinking economy results in reducing

electricity demand and reducing interest in PV systems. A study in Ghana

(Bawakyillenuo, 2012) also revealed the importance of the local context on understanding

the barriers. The political instability in Ghana, which lasted over a period of several

decades, resulted in turbulent economic growth. In 1966, the military government expelled

most foreign nationals who worked in the retail business. This imposed insecurity for

foreigners. The situation did not create an attractive environment for foreign direct

investments in PV technology, preventing the fostering of PV systems in the Ghanaian

market (Bawakyillenuo, 2012). A study by Pansera (2012) also emphasized the

importance of the local situation. In rural areas of Bolivia, people cannot financially afford

to fix the components of PV systems if needed. This means that the economic cost of

maintenance of the PV technology can also be a barrier to the sustainability of adoption,

especially because of poverty.

Based on the cases of off-grid systems in Kenya and Tanzania, Ondraczek (2013)

analyzed the residential SHS applications with a minor reference to other types such as

small-scale commercial applications (including kiosk or mobile phone) and social

institutions. In such a context, he conceptualized low purchasing power as a barrier to

adoption. Based on several developing countries in Asia and Sub-Saharan Africa, Pode

(2013) emphasized economic affordability as a barrier to the diffusion of SHS in rural

areas. The author incorporated this barrier in terms of high total cost, high up-front price,

and payment inflexibility. In line with Pode (2013), Lay et al. (2013) identified the low

income level of households to be a barrier for the adoption of SHS in Kenya. Moreover, a

study by Komatsu et al. (2011) indicated that income alone cannot explain the SHS

43

adoption in Bangladesh. In this study, they emphasize that non-income factors can also be

crucial for the households’ decision making.

The study by Bawakyillenuo (2012) focused on the diffusion of PV systems in Kenya and

Zimbabwe and aimed to deduce policy recommendations for the Ghanaian case. As a

synopsis of such a comparison, the lack of international donor funding for PV systems

seems to be a barrier for diffusion to take off. In this manner, Palit (2013) incorporated the

lack of suitable financing mechanism as the most significant barrier, based on a

comparative study of off-grid systems in several South Asian countries. In the case of

Hong Kong for both on- and off-grid systems, Zhang et al. (2012) also mentioned the high

initial costs, high repair costs, and long payback period as barriers to adoption.

D’Agostino et al. (2011) discussed China’s Renewable Energy Development Project for

SHS. They mention the inadequate government subsidy and the mismatch between

consumer demand and certification program. In the case of off-grid systems in Nepal,

Mainali and Silveira (2011) discussed the large financial gap between the electrification

cost and the level that people can afford. Such a mismatch can be perceived as a barrier.

Brudermann et al. (2013) discussed the on-grid applications of PV systems in Austrian

agriculture. They identified two kinds of financial problems: the high investment needed

for PV installation and uncertainties in the funding process.

Policy barriers

In conjunction with their high price, PV systems are usually not profitable without policy

support in many countries. In this manner, policy measures are of vital importance for

rapid diffusion of environmentally friendly innovations including PV systems (e.g.,

Jacobsson and Lauber, 2006; Rennings, 2000).

At the moment, most countries have a variety of policy measures to support renewable

energies. However, in some cases, existing policy support can be removed, causing a

shock effect in the market. The case of Spain is one example of such a phenomenon.

Movilla et al. (2013) incorporated such a barrier when studying the Spanish case of

diffusion of PV systems. In a similar manner for Austria, Brudermann et al. (2013) also

mentioned the negative impact of the reduction of subsidized feed-in tariffs on adoption.

Insufficient and ineffective policy support is often mentioned as a barrier in several studies

in different contexts, e.g., in the Netherlands (Vasseur et al., 2013) and the US (Kwan,

44

2012; Zhai, 2013). For example, Vasseur et al. (2013) indicate the influence of

inconsistent national subsidy schemes in the Netherlands, which can discourage

entrepreneurs from investing in PV systems. In the South Korean case of micro-generation

as studied by Jeong (2013), policy support also exists but some potential adopters still do

not want to install PV systems. This is because the benefits of adopting PV systems are

not more than the costs. In this case, the potential adopters prefer direct subsidies instead

of low-interest loans. Such a mismatch between demand and policy measures can also be

observed in the case of some UK cities, where the policy measures do not correspond to

the socioeconomic factors (Gooding et al., 2013). Huenteler et al. (2012) analyzed market

development paths in Japan and Germany and subsequently deduced policy

recommendations for Japan. They argue that, for an effective renewable energy policy, the

government should reduce the impact of industry interest on regulatory structure because

it can impede the diffusion process. In this diverse discussion, they also mention three

important barriers: non-institutionalizing on feed-in tariff revision, challenges on

overcoming bureaucratic boundaries, and difficulties in establishing a well-coordinated

policy mix. Such insufficient feed-in tariff schemes are also mentioned in the study by

Karteris and Papadopoulos (2013), which is based on on-grid systems in Greece. In the

same manner, the type of policy is also crucial. For example, Sarzynski et al. (2012)

demonstrated that those states offering cash incentives as policy support experienced

faster diffusion of grid-tied PV technology than the states without cash incentives. With a

strong focus on the solar renewable electricity certificate market, Gaul and Carley (2012)

analyzed the case of North Carolina in the US and deduced lessons from this case. They

mention the lack of transparency in the certificates, which can be incorporated as a barrier

to the adoption of PV systems. In another study (D’Agostino et al., 2011), the negative

influence of insufficient subsidies was also mentioned, highlighting the mismatch between

consumer demand and certification program for SHS in China.

Three studies (Brudermann et al., 2013; Pansera, 2012; Zhang et al., 2012) pay attention to

the importance of cooperation and participation of the stakeholders in energy policy

development. Based on the case of Hong Kong for both on- and off-grid systems, Zhang et

al. (2012) incorporated the lack of participation of stakeholders/community in energy

policy planning and lack of incentives to support the adoption as important barriers. In the

same manner, Brudermann et al. (2013) emphasized the negative effects of non-

cooperative building experts, issues with building permits, and the decision process in

45

regional governments, which takes too much time for grant notification. In addition,

Pansera (2012) conceptualized the lack of innovation policy strategy as a barrier in Bolivia

and suggested a better coordination between stakeholders at the national level.

Although some policies in low-income economies target off-grid applications, some

potential adopters still prefer to wait for the grid extension rather than adopting off-grid

PV systems. This means that adopter’s future expectations can become a barrier in some

cases. This was observed by Bawakyillenuo (2012), who analyzed the case of Ghana in

comparison with Kenya and Zimbabwe. However, Lay et al. (2013) found that grid

extension does not necessarily prevent the adoption of SHS in Kenya. This is because SHS

can be sometimes considered as a complementary energy source.

In some cases, policy support for other sources of energy production can be a barrier to the

adoption of PV systems. An example of such a phenomenon can be seen in the study by

Blum et al. (2013). In this study, they analyzed the cost competitiveness of renewable

energy technologies in Indonesia, mentioning that fossil fuel subsidies make a PV system

less competitive. As a consequence, there is a lack of private investment in the PV sector.

3.2.3. THE GERMAN CASE

The drivers, barriers and challenges of the diffusion of PV systems in Germany have been

extensively discussed in the literature. Early studies focused on the institutional and

political challenges behind the diffusion that had occurred before 2004, the year the

German National Renewable Energy Act (EEG) was amended. Jacobsson and Johnson

(2000) developed an analytical framework examining the diffusion of renewable energy

technologies, based on illustrative examples from Sweden and Germany. From a

technological system perspective, they pointed out the importance of actors, networks and

institutions on the successful transition to renewable energy technologies. In a later study,

Jacobsson and Bergek (2004) also explained the success and failures in such transitions,

analyzing the diffusion of renewable energy sources in Netherlands, Sweden and

Germany. They identified the three major challenges that policy makers faced: difficulties

in foreseeing the outcome of policy intervention; the long time scale of the diffusion

process; and the political struggle to overcome the opposition from incumbent actors.

Another study, focusing only on the German case (Jacobsson and Lauber, 2006),

described the formation of political support for renewable energies as a “battle over

46

institutions”. In this case, the regulatory frameworks supporting wind energy and PV

systems in Germany faced the opposition from both coal and nuclear interests.

Later studies have discussed some new aspects on the PV market in Germany. On the one

hand, Frondel et al. (2008b) criticized the feed-in-tariff scheme of the EEG, pointing out

the inefficiency of this scheme on greenhouse gas abatement. Rather than supporting the

adopters of PV systems with the feed-in-tariff, they instead recommended policy makers

support the PV panels’ production. On the other hand, Laird and Stefes (2009) discussed

the German EEG as a role model for the USA. They identified the decades-long

substantial political change as a driving factor on diffusion of PV systems, putting much

stress on the positive effect of historical contingencies for the German success. In

addition, Dewald and Truffer (2011) explained the importance of substructures in the

market, identifying the different market segments of PV systems showing various patterns

of diffusion. In addition, Leepa and Unfried (2013) studied the effects of feed-in tariff

adjustments on diffusion of PV systems in Germany. They had two main conclusions.

First, they argued that the feed-in tariff in Germany, which declined from 0.39 €/Kwh in

2009 to 0.22 €/Kwh in the early 2012, still supported the diffusion. Second, in order to

meet governmental targets in future, the authors recommended alternative ways of

adjustments regarding the feed-in tariff.

3.3. THE MULTI-LEVEL PERSPECTIVE

3.3.1. MICRO LEVEL DIFFUSION

Micro level of diffusion happens at adopter level. Adopters are the individuals or other

units of adoption that make the decision to adopt the innovation (Rogers, 2003). Any

individual can be a potential adopter, however, not all potential adopters necessarily

decide to embrace the innovation at any certain point of time. The diffusion rate of

innovations, which is the ratio of current number of adoptions to the total number of

potential adoptions, is mainly based on the attributes of an innovation as perceived by

potential adopters. Rogers (2003) deduces these attributes to relative advantage (the

degree to which an innovation is perceived to be better than the other), compatibility (the

degree to which an innovation is perceived as being consistent with the existing values),

complexity (the degree to which an innovation is perceived as being relatively difficult to

47

understand and use), trialability (the degree with which an innovation may be

experimented on a limited basis), and observability (the degree to which the results of an

innovation are visible to others), which also explain 49-87% of the variance on the

different diffusion rates of different innovations.

The decisions regarding adopting innovations can be categorized as optional (where the

adopting individual has almost complete responsibility for the decision), collective (where

the individual has a say in the decision) and authority (where the adopting individual has

no influence in the decision). (Rogers, 2003) However, the diffusion process may involve

a mix of all of these decision‐making types, depending on the type of technology,

regulations and adopters, as is the case of the renewable energy technologies in different

countries (Bodas-Freitas et al., 2010; Reardon, 2009). Because all of these types of

decisions center on individuals, this categorization has been criticized for not providing

sufficient emphasis on structure, context, or collective action (Twining, 2005).

In addition, the variables determining the rates of adoption are influenced by a social

system, which is a set of interrelated units that are engaged in joint problem solving to

accomplish a goal (Rogers, 2003). The members of a social system may be individuals,

informal groups, organizations and/or subsystems. Potential adopters can be influenced to

adopt an innovation by the pressure of the social system generated via adopters, public

policies, shareholders and organizations (Bass, 1969; Frondel et al., 2008a). For example,

an important element that can motivate the diffusion of environmental innovations, such

as solar PV technology, is government policy. The policy has the capacity to create

expectations for an environmental innovation that is essential to foster the diffusion

because it provides clear signals to potential adopters and industrial actors regarding the

future attractiveness of the innovation. In this context, policy makers can foster specific

regulations to guide suppliers and adopters to choose an innovation or a specific design,

which in return may increase the economical relative advantage of the innovation. Until

2012 in Germany, policy makers provided clear signals regarding the growth potential of

solar PV systems through the implementation of feed-in tariff (Dewald and Truffer, 2012;

Hoppmann et al., 2013; Jacobsson and Lauber, 2006). Such tariffs influence the perceived

economic relative advantage of solar PV and, therefore, the diffusion (Jäger, 2006).

Research on renewable energy innovations also asserts that adopters are often influenced

by the local actors. Wüstenhagen et al. (2007) describe this influence as the role of

48

technology cooperation on the diffusion, at which several actors (technicians, industry

representatives, local companies etc.) take an active role in eliciting the adoption. This

actors, in general, corresponds to what Rogers (2003) conceptualize as the efforts of

change agents. A change agent is an individual (or institution) that influences the

decisions of potential adopters in a desirable direction. In the case of renewable energy

innovation, the studies on the PV systems in Germany (Dewald, 2008) and wood-fuelled

heating systems in Austria (Madlener, 2007) have revealed that there is a vast variety of

change agents, which depends on the space-specific context and particular innovation.

The adoption of renewable energy is also boosted by the adoption of previous adopters,

i.e., peer effects. In general, potential adopters are often influenced by what they see and

hear from their peers. The peer effects has received a lot of attention from scholars

analyzing the academic and health practices (e.g. Trogdon et al., 2008; Zimmerman,

2003). In the case diffusion of PV systems, several scholars (e.g. Bollinger and

Gillingham, 2012; Graziano and Gillingham, 2014; Müller and Rode, 2013) have also

asserted that the peers, who have already adopted the PV systems in the same

neighborhood, increase the diffusion rate among potential adopters. This resonates quiet

well with the recent work of Pentland (2014) on social physics. He emphasizes that

innovations and good ideas spread faster if the peers have regular physical, e.g. face-to-

face, interactions. In a well-connected global world, although the potential PV adopters

might have tight virtual connections with their peers from other regions, the most of

physical interactions takes place in the local neighborhoods. As Graziano and Gillingham

(2014) also argue, such spatial dimension is especially critical for innovations that have

both private and public good characteristics, e.g. solar PV systems.

49

Some empirical studies have emerged and have focused on understanding the adopter-

specific factors that influence the diffusion rate of solar PV systems. These factors could

be endogenous, originating internally from the adopters, or exogenous, originating

externally (e.g. from the climatic and physical conditions), or sometimes a mix of both

types. Several scholars conceptualized such factors through a variety of constructs: desire

to be independent from the electricity supplier, familiarity with the technology, religion,

education, housing investment of per capita, income level, climatic conditions, limited

roof space and environmental problem awareness (Balcombe et al., 2014; Jäger, 2006;

McEachern and Hanson, 2008; Peter et al., 2002; Zhang et al., 2012, 2011).

3.3.2. BUSINESS MODELS

Penrose (2009, p. 12), in her seminal work, defines the primary economic function of an

industrial firm as “making use of productive resources for the purpose of supplying goods

and services to the economy in accordance with plans developed and put into effect within

the firm”. In the renewable energy sector, the activities of firms not only result in

supplying goods and services but also in the creation of new knowledge and the

development of different types of designs (Jacobsson and Bergek, 2004). The firms are

important stakeholders in the development of renewable energy policies, and therefore,

they try to influence the political decisions about the design of financial support systems

and the grid access (Wüstenhagen et al., 2007, p. 2686). For example, in Germany, the

energy utility firms are well- known for their effective lobbying strategies for renewable

energy, such as, regular and personal maintenance of contact to politicians and forming a

policy network through associations (Sühlsen and Hisschemöller, 2014).

For the micro level analysis of diffusion, firms could be conceptualized as change agents,

which influence the opinion of potential adopters. For the meso level analysis, firms could

be conceptualized as actors or complementary inputs. Whether the analysis is at micro or

meso level, local solar firms are often identified as important drivers of the diffusion of

solar PV systems. For example, in the case of PV systems in Germany, Dewald and

Truffer (2012) identified that local solar firms stimulated rapid diffusion and market

formation. They argued that the successful diffusion is not only driven by the strong

policy support and favorable geophysical conditions, but also by the market formation

activities of local solar firms. This is in line with the study of Fabrizio and Hawn (2013)

which analyzed the role of local solar companies in the USA. Conceptualizing the local

50

firms as complementary inputs, they identified that local firms do function as important

drivers of diffusion of PV systems.

In the renewable energy research, there is a growing interest in understanding the business

models of the firms. A business model is the story of how the firm work and conduct

business (Magretta, 2002, p. 4), explaining how a firm creates economic value

(Chesbrough and Rosenbloom, 2002). The model might evolve over time and can be

described though the factors related to offering, market, strategy, internal capability,

competition and investor (Morris et al., 2005). Table 4 presents these factors and the

corresponding questions.

TABLE 4. BUSINESS MODEL COMPONENTS16

Component Questions that underlie a business model

1. Factors related the

offering

How do we create value?

2. Market factors Who do we create value for?

3. Internal capability

factors

What is our source of competence?

4. Competitive strategy

factors

How do we competitively position ourselves?

5. Economic factors How we make money?

6. Personal/investor factors What are our time, scope, and size ambitious?

The factors related to the offering describe the scope of the product or service offered by

the company. This addresses the value proposition and creation. The value can be in

different forms, including economic or social (Zott et al., 2011, p. 1029). Market factors

are related to the scope of the market within whihc the company competes; as well as the

nature and geographic dispersion of the customers (Morris et al., 2005). The business

model of a company focuses on a specified market segment or a group of customers

(Chesbrough and Rosenbloom, 2002). Internal capability factors are the source of

competence, which can come from various aspects such as production, selling, marketing,

supply chain management or creative capability (Morris et al., 2005). Competitive

16 (from Morris et al., 2005)

51

strategy factors formulate how the firm gains advantage over rival companies. For

example, competitive advantage can be gained by managing environmental innovations

(Shrivastava, 1995). Economic factors provide the logic behind earning profits. The

business model estimates the profit potential of the offering, based on value chain

(Chesbrough and Rosenbloom, 2002). Personal/investor factors are based on the

investment model of the company. Morris et al. (Morris et al., 2005) give some examples

of such models, including the subsistence model, aiming to meet basic standards to

survive; the income model, restricted by a stable income stream; the growth model,

targeting re-investment and growth; and finally, the speculative model, with the goal of

demonstrating of venture potential and selling out.

The research on product life cycle (see e.g. Anderson and Zeithaml, 1984; Rink and Swan,

1979) and, more recently, business model innovation (e.g. Chesbrough, 2010; Markides,

2013; Sosna et al., 2010; Trimi and Berbegal-Mirabent, 2012) sheds light on why and how

the firms change their business strategies and models. Underpinned by the dynamics of

diffusion patterns, the research on product life cycle (see Day, 1981; Rink and Swan,

1979) explains the interrelation between temporal stages on the sales revenues of

companies and the change in business strategies. In line with the s-curve of diffusion, the

temporal stages of the life-cycle of a product have often been identified as introduction,

growth, maturity and decline (e.g., Anderson and Zeithaml, 1984). At the declining stage,

companies face declining revenues, and the functional focus of companies moves from

marketing/logistics to finance (Hofer, 1975). However, such decline can sometimes result

in re-incline (Rink and Swan, 1979, p. 222). Both decline and incline often induces

companies to revise their business strategies (Anderson and Zeithaml, 1984).

In the context of renewable energy sector, Wüstenhagen and Menichetti (2012) argue that,

the firms are more likely to be influenced by the unexpected risks and opportunities

created by policy makers. Empirical studies on energy policy have analyzed the business

models in different sectors, e.g. in solar thermal energy (e.g. Okkonen and Suhonen,

2010), electrical cars (e.g. Kley et al., 2011), algae biofuel (e.g. Nair and Paulose, 2014)

and biomass heat and power (e.g. Pantaleo et al., 2014). Some other studies have also

analyzed the case of solar PV systems. For example, Loock (2012) studied the business

model preferences of 249 investment managers for renewable energy innovations,

including solar PV systems. This study argued that service-centered business models are

more favorable among investment managers. Huijben and Verbong (2013) studied the PV

52

sector in Netherlands. This study offered a categorization of the households’ business

models: customer-owned, community shares and third party business models. In another

study based in Germany, Richter (2013a, 2013b) analyzed 18 electric utilities. He asserts

that electric utilities have not been keen on developing small-scale renewable energy

projects. Although small-scale PV systems are the largest market segment in Germany,

literature have not paid much attention to the business models of local solar firms that

traditionally develop small-scale PV systems for adopters.

3.3.3. MESO LEVEL DIFFUSION

Albeit the dynamic complexity of diffusion (see e.g. Brockmann and Helbing, 2013; Delre

et al., 2007), one can still observe the regional differences of diffusion patterns at meso

level. For example, some sub-national regions may be at the frontier at adoption of the

innovations, while the others act as lags. In the vast literature on diffusion of innovations,

there is a myriad of examples that shows some regions have been the frontrunner (or have

potential to be the lead): Montana and Wyoming for home food freezers in the US

(Ormrod, 1990), Baden-Württemberg and Bayern for Solar PV systems in Germany

(Dewald and Truffer, 2012), New York and Massachusetts for policy innovations in the

US (Walker, 1969) and some particular regions (northern Finland, Paris of France,

Lombardy of Italy and Andalucía of Spain) for electrified vehicles in Europe (Zubaryeva

et al., 2012).

At meso level, we can observe the wave-like diffusion curves of innovations in time and

space. Does it also happen for diffusion of PV systems and why? Yes, usually. This can

be associated with the dynamics of a contagion phenomenon can undertake. In this token,

based on a case study of the city of Wiesbaden in Germany, Müller & Rode (2013) show

that the adopters’ decision for adopting PV systems are positively influenced by

previously installed PV systems that are located nearby. This phenomenon is also known

as peer effect (see Bollinger and Gillingham, 2012). The potential adopters are highly

influenced by the previous adopters in the local neighborhood, resulting a chain of effects

(as if a contagion phenomenon occurs) and, therefore, a s-curve type of diffusion.

The factors governing the spatial patterns of diffusion have long been examined by

economists, sociologists and geographers. Since the early scholars of innovation diffusion

(e.g. Griliches, 1957; Mansfield, 1961; Rogers, 1962), many conceptual and empirical

53

studies have examined not only the role of spatiotemporal variables in influencing the

wavelike patterns of diffusion (e.g. Berger, 2001; Graziano and Gillingham, 2014;

Hägerstrand, 1967; Haynes et al., 1977) but also the complex dynamics of diffusion in

multi-scale networks (Brockmann and Helbing, 2013; Delre et al., 2007). Albeit the

dynamic complexity of diffusion (see e.g. Brockmann and Helbing, 2013; Delre et al.,

2007), one can still observe the regional differences of diffusion patterns at the sub-

national level. For example, some sub-national regions may be at the frontier at adoption

of the innovations, while the others act as lags. In the vast literature on diffusion of

innovations, there is a myriad of examples that shows some regions have been the

frontrunner (or have potential to be the lead): Montana and Wyoming for home food

freezers in the US (Ormrod, 1990), Baden-Württemberg and Bayern for Solar PV systems

in Germany (Dewald and Truffer, 2012), New York and Massachusetts for policy

innovations in the US (Walker, 1969) and a few regions (northern Finland, Paris of

France, Lombardy of Italy and Andalucía of Spain) for electrified vehicles in Europe

(Zubaryeva et al., 2012).

Recent literature on sustainable transitions, which analyze the systematic shifts for the

adoption of environmental innovations in a variety of areas including renewable energy,

has paid attention to the importance of sub-national regions and spatial dimension. For

example, Coenen et al. (2012) criticize the majority of previous literature for being unable

to reflect on the space-specific contexts. In the same line, Truffer and Coenen (2012)

argue that there is a lack of regional studies on sustainable transitions. In order to fill this

gap, they extend an invitation for research on cities and regions. Addressing this need,

some recent studies shed light the understanding of space on sustainable transition, with a

special focus on the actors, networks and institutions (e.g. Binz et al., 2014; Smith et al.,

2014). However, little systematic attention has been paid to the demand side of sustainable

transitions, i.e., diffusion dynamics at sub-national level.

Beise’s (2004, 2001) lead market model is one of the most comprehensive concepts,

shedding light on the role of demand-side factors on diffusion of environmental

innovations (see Quitzow et al., 2014). Anchored in the demand advantage of Porter’s

(1990) diamond model, the lead market concept deduces the international diffusion of

innovations to the five interrelated innovation-specific attributes of regions: cost

advantage (factors related to price decreases), demand advantage (conditions related to

anticipation of the benefit of an innovation), transfer advantage (conditions that increase

54

the perceived benefit of innovation by foreigners), export advantage (factors that support

the inclusion of foreign demand preferences) and market structure advantage (conditions

that increase the level of competition) (Beise, 2001, p. 85). Yet, the model has been used

in a lot of conceptual and empirical studies, analyzing the global diffusion of innovations,

traditionally at the country-level (e.g. Beise and Rennings, 2005; Cleff et al., 2009;

Horbach et al., 2014; Jacob et al., 2005; Tiwari and Herstatt, 2012; Walz and Köhler,

2014). However, as Beise (2001, pp. 125–126) already outlines, lead markets can be sub-

national regions. Beise (2004, 2001) argues that if particular regions, that have higher

advantages in these five attributes, adopt an innovation, the diffusion can spread to other

regions rapidly (see Figure 13).

FIGURE 13. PENETRATION RATE IN PERCENT17

The first of these attributes is cost advantage, which is mostly based on economies of

scale through mass production (Beise, 2004). The cost of an innovation can vary in a

country depending on local characteristics such as transportation infrastructure,

availability of suppliers, natural resources, regional policies or availability of

complementary goods. For example in Egypt, southern coastal region has the lowest costs

of electricity delivered by an off-grid PV (Szabó et al., 2011). In the USA, the lowest

levelized cost of wind energy is in the Midwest region (Lazard, 2013). In some cases,

spatial characteristics implicitly influence the relative affordability of an innovation. A

case in point is the adoption rate of residential air conditioning, which depends also on the

17 (adopted from Beise 2004)

lead market

lag markets

lag markets

Pen

etra

tio

n R

ate

in P

erce

nt

t t

lead marketlead market

lag markets lag markets

Pen

etra

tio

n S

pee

d

55

electricity price which can vary from state to state in a country (Ormrod, 1990). The cost

advantage can be one of the most important spatial lead market attributes, as adopters’

decisions can be often based on economic affordability (Rogers, 2003).

The second attribute is demand advantage which relates to local factors that increase the

demand for adopting a particular innovation. Demand advantage could vary among sub-

national regions. A good example is the airline industry, where the demand for air

transport significantly depends on the spatial area (Graham, 1998). Another example is the

demand for higher education; at which decision of prospective students depend on spatial

factors, e.g., distance to university (Sá et al., 2004). Income level is usually a good

predictor for which sub-national regions might have more demand advantage (Beise,

2004). For example, some particular sub-national regions in Europe, e.g., cities of London,

Madrid, Berlin and Rome, have more potential to adopt electrified vehicles because of,

among many others, the high GDP per capita in these cities (Zubaryeva et al., 2012).

The third attribute is transfer advantage, which is described as the ability of a country to

shape the preferences of other countries (Beise, 2004). Transfer advantage could be space-

dependent as diffusion of innovations depends on the intensity of communication between

spatial areas (Hägerstrand, 1967). Transfer advantage is also related to observability, the

degree to which results of an innovation is visible to others (Rogers, 2003) and so called

“demonstration effect” (Kalish et al., 1995). The rate of diffusion of an innovation

increases when the benefits of it are easier to see. Reputable adopters or markets, that

adopt the innovation at initial phase, create credibility for the innovation and reduce the

risk of adoption for laggards and lag markets (Beise and Cleff, 2004; Rogers, 1962). For

example, Facebook has been adopted initially in the Ivy League universities, eight elite

institutions in the Northeast region of the US, before it spreads in the US and then

worldwide.

The fourth attribute is export advantage which is traditionally explained by three factors:

similarities between local market conditions, export experience of local firms and demand

push (Beise and Rennings, 2005; Beise, 2004). First, diffusion is importantly influenced

by cultural similarities (Strang and Meyer, 1993). Innovations are easier to export from

one spatial area to another if these spatial areas are similar. Second, export experience of

companies varies from local area to local area as well. For example in Germany, export

quota vary from 16% to 45% depending on the region, where Saarland region has the

56

highest export quota (SLBW, 2013). Third, local adopters can be more sensitive to global

problems than potential adopters in other spatial regions. This kind of spatial demand push

can trigger the local companies and potential adopters to meet global challenges ahead to

companies and potential adopters in other regions.

The fifth attribute is market structure, which is mainly associated with competition.

Porter (1990) indicates that national competitive advantage arises from competition in

internal market. From a spatial perspective, the role of spatial location is vital for

competitive advantage (Porter, 2000) and market formation dynamics (Dewald and

Truffer, 2012). The spatial forms of technological specialization and comparative

advantages are often originated from sub-national formations such as regions and clusters

(Morgan, 2004). Spatial proximity of actors is important to form cooperation for

innovation (Asheim and Gertler, 2005). The companies located in competitive markets

are under pressure to develop market-oriented innovations. Google for example, one of the

largest companies in global internet industry, is located in Silicon Valley on the south of

San Francisco Bay Area in the US, a leading geographic hub for high-tech innovation.

3.3.4. MODELING

Diffusion models have been used to capture life-cycle dynamics of a new product and they

traditionally have been applied to forecast (to predict) the demand for a new product at

meso level (Mahajan et al., 2000). The wealth of research on modeling diffusion of

innovations has been impressive and it confirms its continuing importance as a research

topic (see e.g. Kiesling et al., 2011; Meade and Islam, 2006, 1998). However after the

enhanced penetration of communication and other technological innovations, research in

diffusion modeling will have to expand its horizons in order to remain timely and abreast

of market trends (Peres et al., 2010).

The available models can be categorized into two main groups as "differential equation

models" and "agent based models" (similar to Rahmandad and Sterman, 2008). The term

"differential equation models" refers to traditional models of innovation diffusion which

provides an empirical generalization which is usually based on an aggregation at the

cumulative market level. Agent Based models refer to the models which consist of a set of

agents that encapsulate the behaviours of the various individuals that make up the system

by emulating these behaviours (Parunak and Savit, 1998). In differential equation models,

57

the path of the cumulative adoption of an innovation between introduction and saturation

is usually modeled by an S-curve based on Rogers' theory. The pioneer s-shaped diffusion

models (what I call them differential equation models) are those of Fourt and Woodlock

(1960), Mansfield (1961) and Bass (1969) (as cited in Meade and Islam, 2006). One of the

most used diffusion models in the literature and by the companies has been the Bass

model. Bass (1969) presented the first purchase growth of a new durable product in the

market. This model suggests that individuals are influenced by a desire to innovate and by

a need to imitate others in the population. After the Bass model, there have been many

others published. Some of them are based on different mathematical equations while some

proposed modifications on Bass model.

Agent based modeling has been increasingly applied to social, and economic problems

previously modeled with nonlinear differential equations (Rahmandad and Sterman,

2008). This method is rooted from complexity theory. It indicates that phenomena at the

macro level can be understood as emerging from interactions between individuals at the

micro level: and macro phenomena affect the behavioral context at the micro level

(Garcia and Jager, 2011).In innovation diffusion field, agent based models describes the

market as a collection of individual elements (so called agents) interacting with each other

through connections (Peres et al., 2010). These models have been increasingly adopted in

diffusion research in recent years in order to overcome the limitations of traditional

aggregate models. Kiesling et al. (2011) explains that agent based models´ ability to model

complex phenomena in a socio-economic system (e.g., the diffusion of innovations) is one

of the most important reasons why agent based models have gained momentum in recent

years. However in the literature, there was a limited use of diffusion models in

environmental innovations, especially in renewable energy technology analysis (Rao and

Kishore, 2010).

The diffusion of PV systems has been modeled from various theoretical perspectives

through a variety of methods. The study of Mesak and Coleman (1992) appears as one of

the first contributions of this phenomenon. They extended the model of Bass (1969) in

order to forecast the PV systems’ diffusion for the residential sector in Kuwait. Their

study addressed the link between the government subsidies and the PV systems’ diffusion.

Assuming that diffusion of PV systems is based on price evolution, Luque (2001)

combined the learning curve approach (Yelle, 1979) with a demand elasticity model. As a

result, he forecasted both the market and price evolution of PV systems. Learning curve

58

approach was also used by Masini and Frankl (2002). They simulated the diffusion of PV

systems in 5 southern Europe countries under different scenarios. Watanabe and Asgari

(2004) attempted to link the dynamic behavior of learning coefficient with that of

innovation diffusion. They developed a mathematical formulation of the logistic growth

function which incorporates the innovation diffusion and its carrying capacity. The model

was applied to the Japanese case of PV systems from 1986 to 2000.

Kobos et al (2006) extended the debate by combining the frameworks of learning by doing

and learning by searching. Based on the estimation of the cost curves and learning

elasticity through regression analysis, they examined experience curve analysis of wind

energy and PV systems under several scenarios. Based on the well-known Bass model

(Bass, 1969) and the extension of it (Bass et al., 1994), Guidolin and Mortarino (2010)

forecasted the diffusion patterns of PV systems in several countries. Their paper showed

the important differences among studied countries, presenting the different stages of the

diffusion process. However, the comparison between the forecasted values of this study

and what happened in reality indicates a big mismatch, e.g., in the case of Germany.

Since 2011, the number of studies addressing the modeling diffusion of PV systems has

rapidly increased. This is probably, among others, due to the increasing available

empirical data. For example, Popp et al. (2011) studied the relationship between global

technology stock influence the diffusion of renewable energy technologies such as PV

systems. They explained the investment per capita in capacity of installed renewable

energy through a statistical fixed effects model. Although many scholars used a

downscaled unit of analysis at regional or local level (Gooding et al., 2013; Higgins et al.,

2014; Kwan, 2012), how spatial differences can affect the two dimensional diffusion paths

have not been addressed yet.

59

4. METHODOLOGY

“Energy studies must become more socially oriented, interdisciplinary and

heterogeneous. Problem-focused research activities that centre on both physical and

social processes, include diverse actors and mix qualitative and quantitative methods,

have a better chance of achieving analytic excellence and social impact (Sovacool,

2014b, p. 530)”

This PhD Dissertation is based on multi-methodological approach. Anchored in a case

study method (sub-chapter 4.1), the micro and firm level analyses are conducted. For the

meso level analysis and modeling perspective, an indicator-based approach (sub-chapter

4.2) and finite element method (sub-chapter 4.3) have been respectively chosen. Beyond

these three methods, there have been also two literature reviews, methodological aspects

of which are previously explained in sub-chapter 3.2.

4.1. CASE STUDY APPROACH

Case study method is traditionally used for various purposes, such as providing

description, testing theory or generating theory (Eisenhardt, 1989), and the method is

common in energy research (see e.g., Cagno et al., 2015; Darmani, 2015). The case studies

should not only address theory, but also provide real world examples in a new way

(Siggelkow, 2007). In this Dissertation, the case study method is used in order to provide

an in-depth description of the diffusion of solar PV systems at grid parity, a rather under-

researched phenomenon. When there is lack of previous research, a case study approach is

appropriate for exploring a contemporary phenomenon, gaining a holistic view of complex

instances through observation and searching for patterns (Yin, 2011). The strength of case

study lies in its approach to develop empirically collected and context dependent

knowledge with “multiple wealth of details” (Flyvbjerg, 2006; Lincoln and Guba, 2009;

Stake, 2000).

60

4.1.1. EMPIRICAL CONTEXT

The case study approach is based on the findings of a study conducted in a town called

Rottenburg am Neckar (a town with 43000 inhabitants) in Tübingen. It focuses on a local

solar firm, which is called Hartmann Energitechnik GmBH (HET). As shown in Figure

14 and Figure 15, it is located in a rural area in Southern Germany, near a village of

Oberndorf in Rottenburg Am Neckar.

FIGURE 14- THE LOCATION OF HET18

18 The map represents the 5-digit zip code areas of Germany. The map is generated via the QGIS Desktop 2.6 software,

based on the data GADM database (www.gadm.org), version 2.0, December 2011.

61

FIGURE 15- THE HET AND ITS NEIGHBORHOOD

The town and the firm were chosen for two main reasons. Firstly, the town is located in

the frontrunner region of PV systems’ market formation (see Dewald and Truffer, 2012)

and diffusion19

. Secondly, the firm is one of the pioneer local solar companies in

Germany. It has been a part of solar initiative movement (Drücke, 2004) and,

consequently, e.g., featured in the national media (e.g. SB, 2007; SWW, 2004). The firm

was founded in 1995 by a local entrepreneur, Thomas Hartmann, a native of the region.

He has been also the co-founder of two solar initiatives: Solar-Partner e.V (a network of

companies, freelance solar consultants and partner companies) and Sonnenhaus-Institut

e.V (an association of architects, engineers and managers of the solar industry, focusing on

solar-heated and solar-electrified buildings).

Rottenburg am Neckar, is a part of the well-known region in Germany, called “Swabia”. It

is distinguished by its cultural, historical and linguistic characteristics. A special dialect of

German, Swabian (Schwabish) is spoken in the region. Although the region struggled with

poverty and scarcity until the beginning of 20th

Century, it had an economical and

industrial uptake in the 21th century. Today, Swabians are famous for hating debt,

avoiding extravagance and getting the best deal (The Economist, 2014). They are also

known as hard-workers, as one old saying expresses "Schaffe schaffe, Häusle baue",

which refers to ”you should work hard”. Rottenburg am Neckar is today surrounded by

several villages and this geographical area is dominantly full of single/double family

19 Rottenburg am Neckar’s normalized cumulative diffusion number, calculated by dividing the number of installations

in a spatial area by its population, is 0.026. This is higher than average values both in southern Germany (0.015 as

calculated through the zip code areas 7 to 9) and the rest of the Germany (0.005 as calculated through zip code zones 1

to 6).

62

houses rather than multi-family apartments. There are many farmers in the surroundings;

living as single families with a relatively large lands.

Hartmann Energietechnik GmbH

Hartmann Energietechnik GmbH (HET) was founded in 1995 in the village of Oberndorf

in Rottenburg am Neckar by Thomas Hartmann, a local entrepreneur (according to the

definition of Michelacci and Silva, 2007). HET is located in the so called “Solar-Center”

and offers solar PV, solar thermal and biomass boilers for the citizens in the

neighborhood, in partnership with two regional associations: Solar-Partner e.V. and

Sonnenhaus-Institut e.V.. In the PV branch, the main activities of HET are focused on the

promotion, consulting, conceptual designing, assembling and installation. HET offers a

wide range of solutions, depending on the needs and preferences of potential PV adopters:

various montage systems and alternative concepts. In 2012, HET´s sales volume was

approximately 3 million Euros. In the same year, the PV solutions offered by HET was

adopted not only in Rottenburg am Neckar (30% of adoptions) but also in other

neighborhoods in a radius of up to 50 km. HET shared around an average of 10% of the

PV market in the neighboring towns with respect to the installation numbers.

HET history goes back to 1993, when T. Hartmann obtained his first solar collector

training for solar thermal systems. In the late 1990s, inspired by an Austrian model, HET

installed many different solar systems, participated in different exhibitions and co-founded

the association Solar-Partner e.V, a network of heating and solar specialist companies,

freelance solar consultants and partner companies in Bavaria and Baden-Württemberg.

The installation of the solar thermal for the sport hall in the village of Oberndorf was one

of milestones of HET, by which the company achieved visibility and recognition in the

neighborhood.

Since the introduction of the feed-in tariff for PV in Germany in 2000 by the federal

government, HET focused more on the PV systems, installing many systems in the

neighborhood (e.g., for the church and the bishop’s house in 2002). In the beginning of

2000s, the efforts of HET received much attention from the government as well. In 2001,

HET was visited by the Mayor of Rottenburg am Neckar and a member of German

parliament, and in 2002, HET was visited by a minister of the regional government to

have solar-walks and become informed about the diffusion of solar systems in the region.

63

In 2004, HET co-founded the association of Sonnenhaus-Institut e.V., an association that

brings architects, engineers and managers of the solar industry together with the goal of

sustainable development and distribution of construction and heating techniques for

largely solar-heated buildings. In 2006, as a pilot project of the association, Solar-Center

was built with PV systems of a capacity of 60-kWp on the roof and 150 square meters of

solar thermal on the façade that provide 80% of the energy by itself.

HET’s relationship with potential adopters is based on the facilities in the “Solar-Center”

and the monthly “Solar-walks”. The Solar-Center is a transparent place where the

potential adopters can see any part of the facilities (assembly hall, design unit, and

exhibition part) and become informed about PV solutions. The Solar-Center also serves as

a meeting place for seminars, construction courses, open-doors day and guided tours about

solar PV, solar thermal and biomass. In addition, interested potential adopters can request

a personal visit by T. Hartmann to their houses to discuss the best option. “Solar-walks”

are regular 3-4 hours exhibition tours in the village of Oberndorf, guided by T. Hartmann,

where PV systems that HET installed are described and discussed. Anyone who is

interested can join these solar-walks and learn about different PV concepts on the field.

Since 2001, solar-walks have been organized every month, and during this time, it has

never failed. The visitors of these solar-walks have not been only limited to the region, as

there have been many amateur and professional guests from all over the world, such as

USA, Japan and Sri Lanka.

4.1.2. DATA COLLECTION AND ANALYSIS

The data, which were collected during the case study from December 2012 to March 2013,

include 18 semi-structured interviews, approximately 600 hours of observation and the

internal data of the company. The interviews, the duration of which varied between ten

min to one hour, were conducted with Thomas Hartmann, the employees in the firm, the

directors of four other partner local solar firms and the adopters of PV systems. Such

diversity in interviewees gave a holistic view to understand the diffusion of PV systems at

grid parity both from adopter’s perspective and firm’s perspective. The interviews were

conducted at various locations: the firm, the houses of adopters and a regional workshop.

Following a semi-structured interview approach, the interviewer had the freedom to add

new questions, based on the flow of the discussions during interviews. Except in one

instance, the language of the interviews was German. All interviews were electronically

64

recorded. The observations took place both in the firm and in the town. The observations

in the firm were conducted through an average of nine hours/weekday presence for three

months. As a participant observer (Atkinson and Hammersley, 1994), all of the formal and

informal gatherings of the company was attended. In addition, several ongoing PV

installations projects were monitored on the site and a 2-day regional workshop (of 150

participants) was actively attended. The internal data include not only various documents

such as internal agreements, technical and feasibility reports of each PV installation and

the documentation of each project, but also the sales database of HET. Data were analyzed

throughout and after the fieldwork. With an insider-outsider team research approach

(Bartunek and Louis, 1996) and data triangulation with multiple source of evidence (Yin,

2003), the subjectivity was aimed to be avoided. To empirically validate the results, the

findings were presented for the firm, both in an internal meeting and, on a later phase, in a

written form.

4.2. INDICATOR-BASED APPROACH

The indicator-based approach consists of three steps: identification of relevant variables

for PV context, assessment of lead market attributes at sub-national level and diffusion

curve analysis.

1) Following the methodological approaches in relevant literature (e.g. Beise and

Cleff, 2004; Beise and Rennings, 2005; Horbach et al., 2014), in the first part, the

results of the case study is triangulated with a literature review. As a result, a set of

indicators are identified.

2) The second step assesses the lead market attributes at sub-national level. In order

to tackle this, several indicators, which were identified in the previous step, were

then retrieved from Federal and Regional Statistical Offices of Germany (SABL,

2013; SLBW, 2013) and the other studies (Berthold et al., 2009; Diekmann et al.,

2012; Drücke, 2004; Kost et al., 2013). Each indicator is categorized according to

the five attributes of the lead market model and mapped spatially via QGIS 2.0.1

Software, which is an open source geographic information systems (GIS) program

that provides spatial data analysis. In the end, the indicators are aggregated using a

simple method (similar to Beise and Cleff, 2004).

65

3) Existence of lead markets can be formally identified by gathering data on

innovation´s diffusion curves (Beise, 2004; p. 1013). Higher than usual penetration

rate in particular spatial areas is a sign that a lead market exist. In our study, the

data for plotting spatial diffusion curves is gathered from the Information Platform

of four German Transmission Network Operators for the Renewable Energy

Sources Act (EEG) and the Combined Heat and Power Act (KWK-G). The data

has two dimensions: time and space (i.e., it is spatiotemporal). In terms of time, the

unit of analysis is year and time span is from 2000 to 2012. For the spatial

dimension, the boundary is Germany and the unit of analysis is zip code areas

defined by Deutsche Post AG, the main postal service provider in Germany (see

Figure 16).

FIGURE 16. GEOGRAPHIC ZIP CODE AREAS AND STATES IN GERMANY

66

The advantage of using zip code zones is to have relevantly uniform distribution of

population and area for each geographic zone, which is a common practice in economic

geography studies. The dependent variable, which is time and space dependent, is the

number of small scale PV installations rather than capacity of installations. The

penetration rate of PV diffusion cannot precisely be calculated as we do not know what is

the highest possible number of installations given that PV can be installed not only on

rooftops but also on façades or gardens. Therefore, we use a normalized cumulative

diffusion number to represent the penetration rate. It is calculated by dividing the

number of installations in a spatial area by its population , and multiplied by

100.

While the results of first step are more related to methodological aspects, the outcome of

the second and third steps develops the answers of RQ3 of this Thesis. Therefore, the

results of the first step are presented in the following sub-chapter 4.2.1 and the results of

other two steps are discussed in sub-chapter 5.3.

4.2.1. IDENTIFICATION OF INDICATORS

The overview of indicators that approximate the lead market attributes for diffusion of

small scale PV in Germany is presented in Table 5. These indicators can be applied not

only at sub-national level but also at the national level (if necessary). Although the five

lead market attributes are innovation-specific (Beise, 2001), the corresponding indicators

for export and transfer advantage are not available at the specific innovation level (see

Walz and Köhler, 2014, p. 36).

67

TABLE 5. A SYSTEM OF INDICATORS THAT PROXY LEAD MARKET ATTRIBUTES FOR PV

Lead Market

Attributes

Available Indicators as Proxy

Cost Advantage [1/LCOE] 1/Levelized cost of electricity from PV (a)

, 2012

(kWp/€)

Demand Advantage [Green] Green Party Support (b)

, 2009 (%)

[Employment] Employment Rate (b)

, 2010 (%)

[Housing] House/apartment Ratio (b)

, 2011 (%)

Transfer Advantage [R&D] R&D Expenditure Index (c)

, 2009

[Patent] Patent applications for renewable energy technologies

per 100 thousand capita (e)

, 2008-2011

[Demonstration] Number of demonstration projects funded by

DBU (normalized with population ratio) (d)

, 2001 (%)

Export Advantage [Foreigners] Ratio of foreigners in population (b)

, 2012 (%)

[Export] Export Quota (f)

, 2012 (%)

Market Structure

Advantage

[Initiatives] Number of Solar Initiatives (normalized with

population ratio) (d)

, 2004 (%)

Sources:

(a) Kost et al., (2013)

(b) Statistische Ämter des Bundes und der Länder (SABL, 2013)

(c) Berthold et al., (2009)

(d) Drücke (2004)

(e) Diekmann et al., (2012)

f) Statistisches Landesamt Baden-Württemberg (SLBW, 2013)

Cost advantage

Cost advantage of PV diffusion is traditionally measured by levelized cost of electricity

(LCOE) (e.g. Branker et al., 2011; Lettner and Auer, 2012). It is an indicator that

expresses the cost of electricity generation over the years. LCOE is based on the initial

capital, solar radiation, continuous operation costs, service life time and costs of

maintenance. As solar radiation and the complementary goods´ costs vary between

geographic zones so do LCOE for PV. Although the calculation of LCOE is disputed in

68

literature (as explained by Munoz et al., 2014), LCOE is sufficient enough to represent

the relative cost competitiveness for PV (e.g. see Reichelstein and Yorston, 2013). Our

case study also confirms that if the space-specific LCOE of PV is relatively cheaper than

the retail electricity prices per kW, there is more likelihood for the adoption. Potential

adopters are usually informed about such financial aspects by the system suppliers before

the adoption takes places. Therefore, we use the inverse of LCOE [1/LCOE]. Increasing

[1/LCOE] results in increasing cost advantage, and therefore, increasing adoption rate.

Demand advantage

The two important factors that may influence the demand advantage are socioeconomic

status (McEachern and Hanson, 2008; Zhang et al., 2011) and environmental awareness

(Jäger, 2006; Zhang et al., 2011). The former has a strong correlation with employment

rate (Bartley and Owen, 1996) and the latter can be associated with green party support (as

observed in the case study). Therefore, we choose employment rate [Employment] and the

support for green party (Alliance '90/the Greens) [Green] as indicators. Increasing

employment rate and green party support are expected to result in increasing demand

advantage.

In addition, our case study reveals that the entity of decision makers is also an important

factor on adoption. Following the Rogers’ (2003) categorization about decisions on

adoption, we expect that optional decisions (e.g., the case of single-family houses where

the adopting individual has almost complete responsibility for the decision) are much

easier to take rather than collective decisions (e.g., the case of multi-family apartments

where the individual has only a say in the decision). Members of multi-family apartments

are less likely to take the decision in favor of the PV adoption, as it requires the approval

of each family in the apartment. Therefore, we also use an additional indicator for demand

advantage, which is house/apartment ratio [Housing]. It expresses the ratio of single and

double family houses to the total number of houses at sub-national level.

Transfer advantage

Transfer advantage relates to the ability of a region to shape the preferences of other

regions. We use three indicators to proxy the transfer advantage at sub-national scale.

These are research & development expenditure index [R&D], patent applications for

renewable energy technologies per 100 thousand capita [Patent] and number of

69

demonstration projects (normalized with population ratio) funded by DBU, German

Federal Environmental Foundation [Demonstration]. R&D expenditure and patents are

common to be used for approximating the transfer advantages of lead markets (see e.g.

Beise and Rennings, 2005; Horbach et al., 2014; Köhler et al., 2014). Moreover, transfer

advantage builds also on the demonstration effect (Beise, 2004), which can be, for

example, approximated by the church rooftops project of DBU from 1999 to 2001. This

project aimed to support PV installations for churches’ rooftops. The case study in

Rottenburg am Neckar also confirmed that PV installations at public places, such as

schools or churches, have resulted in increasing perceived observability of PV.

Export advantage

We use two indicators to approximate the export advantage of a lead market at sub-

national level. The first one is the export quota [Export]. Higher export quota expresses

the higher export advantage of a region (Beise, 2004; Köhler et al., 2014). The second

indicator is the ratio of foreigners in population at sub-national scale [Foreigners]. Higher

interaction between the adopters in potential lead market and lag markets drives the

diffusion (Beise and Cleff, 2004; Beise, 2001). Yet, regions with high foreign population

enjoy an export advantage since they have more interaction with other regions

Market structure advantage

Local solar initiatives played a vital role for market formation at sub-national scale and a

successful growth of diffusion of PV in Germany (Dewald and Truffer, 2012). The

number of these initiatives were already over 200 by 2004 (Drücke, 2004). As the solar

initiatives were central intermediary actors for market formation, supporting especially

small scale PV diffusion (Dewald and Truffer, 2012), we use the number of solar

initiatives (normalized with population ratio) [Initiatives] as an indicator, approximating

the market structure advantage. One of these solar initiatives was the Solar Partner e.V,

co-founded by Hartman Energietechnik GmbH. Both adopters and members of Solar

Partner e.V indicate that the activities of this initiative have stimulated the diffusion of

PV, serving as a platform to share the good practices and support market formation.

70

4.3. FINITE ELEMENT METHOD

The research, that applied the Finite Element Model in social science related fields, has

flourished in the 1990s20

. These studies can be categorized under three groups. The first

category is about Economic Geography. This stream is dominated by the studies from

transportation research. The second field is related to the Finance, with a particular

emphasis on American and European options pricing. The studies on implementing Finite

Element Model in option pricing appeared in the early 2000s (e.g., Allegretto et al., 2001;

Tomas and Yalamanchili, 2001) and have been increasing recently. Lastly, the third

category consisted studies from various topics, including economic growth models (e.g.,

Cao-Alvira, 2011) and cash management optimization (e.g., Baccarin, 2009). In all three

categories, most of the studies focus on methodological and modeling aspects, such as

optimization, improvement and extension of applied models with Finite Element Method

(e.g. Angermann and Wang, 2007 in option pricing; Ho et al., 2006 in transportation

systems in a city).

Finite Element Models usually start with a continuum problem and a numerical model, an

example of which is illustrated in Figure 17. Then, one part of the model carries out

domain discretization, i.e., meshing, and the other deals with discretization of the equation

in respect to both spatial and temporal dimensions. In the end, the model is solved with a

numerical software tool, which typically contains three parts: preprocessing, the main

processing and post processing. The results are typically tested both for accuracy and

convergence.

20 E.g. the studies have appeared in a variety of journals from different fields. Some examples are the Annals of Regional

Science (Wong and Sun, 2001; Wong et al., 2006), an international journal of urban, regional and environmental studies;

Computational Economics (Cao-Alvira, 2010, 2011; Shinohara and Okuda, 2010), a journal dedicated to the interface of

computer science, economics and management science; Transportation Research Part B- Methodological (Ho et al.,

2006; Jiang et al., 2011; Wong, 1998; Yang et al., 1994), a journal presenting methodological aspects applied in

transportation systems; and Management Science (Ben-Ameur et al., 2002), a journal that publishes on the practice of

management.

71

FIGURE 17- TYPICAL IMPLEMENTATION STEPS FOR FINITE ELEMENT MODEL21

The continuum problem and mathematical model can vary, based on the studied

phenomena. There are different continuum problems considered. For example, Ho and

Wong (2007) study a continuum problem of the macro role for the interacting housing and

travel patterns in both land-use and transportation system; Wong et al., (2006) considers a

continuum traffic equilibrium problem in a city with several competing facilities; and

Shinohara and Okuda (2010) study continuum innovation diffusion via an analogy with

natural physical phenomena. Some studies deal with a minimization problem while others

caries out a maximization problem. In general, the prediction of customer behavior is the

focal point of most of the mathematical models that are applied in these studies.

The domains, which are subject to discretization, are usually chosen as a whole region of a

city (e.g., Wong et al., 2004) or a country (e.g., Shinohara and Okuda, 2010). The selected

domains are typically meshed with a set of three-node linear triangular finite elements.

The studied variables, such as diffusion rate or traffic density, can be then mapped on

these meshed domains. This enables a deeper understanding on the studied phenomena,

such as pinpointing the sources of diffusion of innovations or the congested roads.

After finite element discretization, the solutions of are calculated by a software, which

also offers accuracy and convergence analysis for real-world applications. Accuracy in

21 (adopted from Lewis et al., 2004)

72

Finite Element Model is usually tested by numerical solutions of restricted problems with

known results; whereas convergence can be tested by observing whether the numerical

results are changing with increasing number of elements or not (Walz et al., 1968). In the

literature, the convergence of the method have been validated with different approaches

(e.g., Ho and Wong, 2007; Wong and Sun, 2001).

4.3.1. APPLIED FORMULATION

The Finite Element Model applied in this study considers a continuum geography where

the innovations are diffused as if heat diffusion (following Haynes et al., 1977).

Innovation diffusion rate (D), the proportion of actual number of adoptions to the number

of maximum possible adoptions in a given area, is the object of the model. Diffusion rate

is an important explanatory parameter used in literature (Beise, 2004; e.g., Rogers, 2003),

explaining maturity of the diffusion process. For instance, if innovation diffusion rate in a

given area is 70% at a given time, it means that 70% of the potential adopters have already

adopted the innovation by that time. If it is 100%, it means that the diffusion in the given

area has already saturated and there is no expected adoption left. Innovation diffusion rate,

as defined in this study, can therefore be interpreted as similar to the temperature in heat

diffusion equation. It depends both on time and space. Generated from an innovation

source, the innovations flow in space in all directions. For example, Figure 18 shows an

illustrative case, where innovations, i.e., PV systems, flow from the innovation source,

e.g., a company, to the adopters.

73

FIGURE 18- THE ILLUSTRATION OF FLOW IN INNOVATIONS IN SPACE

The innovation flow, or in other words, innovation flux (q) can be formulated as similar to

the fundamental rule for flow of heat energy.

where c can be called as innovation diffusivity22

. This rule indicates that innovation flux

(q) is proportional to the negative innovation rate across the space. With the

transformation of this equation, the governing equation on the diffusion rate (D) can be

derived as the following.

where t is time. The x and y are the coordinates of the location in spatial dimension. The

derivation of such formulation, which is based on of the model of Haynes et al., (1977),

can also be seen in the study of Shinohara and Okuda (2010). It is also similar to the

formulation of Morrill (1970), that describes the innovation diffusion as if a wave

propagation. Such mathematical models assume that the innovation will diffuse from the

source to the adopters.

For the application of the Finite Element Model, a case of a solar company, called HET, is

considered. The HET´s target market includes mainly 10 zip codes in southern Germany

22 Innovation diffusivity value can also be interpreted as a combination of the coefficient of innovation and the

coefficient of imitation in Bass model (Bass, 1969). In addition, it can be interpreted as a joint representative of demand-

side factors.

74

(see Figure 19). Consequently, it is assumed that the HET is the innovation source in this

selected area. This means that adoption depends on both the distance to the HET and the

innovation diffusivity of a spatial point. Households are assumed to be the potential

adopters that are heterogeneously spread in the geography.

FIGURE 19- THE MODELED GEOGRAPHY FROM SOUTHERN GERMANY

As shown in Table 6, a variety of innovation diffusivity numbers are considered in order

to address a possible spatial heterogeneity problem. These values are chosen on a purpose,

making the zip code area 72149 an extreme case with very low innovation diffusivity

value23

. As an alternative, these values could have been obtained from the history data of

diffusion curves.

TABLE 6. THE VALUES FOR INNOVATION DIFFUSIVITY

Zip

code

71126 71149 71559 72070 72072 72108 72119 72149 72181 72184

c(x,y) 72 0.5 35 58 138 150 23 80 46 112

23 Such low innovation diffusivity could be originated from low innovativeness or low socioecological resources of a

local area. However, the purpose of this study is not to investigate the reasons behind spatial heterogeneity.

75

To solve the mathematical equation, both domain and mathematical equation are meshed

by ABAQUS software (Abaqus, 2013). This computational mesh on approximated

geographic is shown in Figure 20. The HET is located between zip code areas 72108 and

72119.

FIGURE 20- THE COMPUTATIONAL MESH ON APPROXIMATED GEOGRAPHIC

The type of the elements used for computational mesh is 4-node linear quadrilateral. The

total number of elements is 752. The solution of the model is based on a 13-year

calculation. As boundary condition, the location of the HET is assumed to have 100%

diffusion rate during the whole time period. The reason behind this assumption lies on the

fact that HET consists of only one building which has already adopted a PV system.

77

5. RESULTS AND DISCUSSIONS

5.1. MICRO LEVEL ANALYSIS

The first research question of this Thesis was: Why and how do the PV systems at grid

parity get adopted at micro level? In order to answer this research question, we analyze the

case in Rottenburg am Neckar in 2012. This case has some particular characteristics. First,

the case company (Hartmann Energietechnik GmbH, i.e., HET) is a pioneer local solar

company in the region with a strong network with other local solar companies and

adopters. Second, the year 2012 was the time when the feed-in tariff drastically

diminished.

“Hartmann Energietechnik GmbH is a local installation company, well-known in

the region. There are many references in the area, and Mr. Hartmann is a very

well-known guy…” (P.M., PV Adopter, 21.12.2012).

“2012 was catastrophic bad year - Solid cut in the feed-in tariff.” (G.W., Director

of a partner firm in HET network, 07.12.2012).

To understand why and how do the PV systems at grid parity get adopted at micro level,

we separate the analysis into three parts associating with the role of policy measures, local

solar companies and adopters.

5.1.1. POLICY MEASURES

Germany is expected to achieve the National Renewable Energy Action Plan Target for

PV for 2020 at an earlier stage than planned, sometime between 2016 and 2020 (EPIA,

2013). This is the impact of the strong policy support that resulted in the rapid growth of

the PV market in 2000s. During this period, the cost covering feed-in tariff was the main

factor that policy makers implemented to foster market formation and PV diffusion

(Dewald and Truffer, 2012). As expected, this case study reveals that feed-in tariff is no

longer the most important motivation to adopt PV. This change is a consequence of PV

systems achieving grid parity in 2012 and the reduced feed-in tariff, which pays less than

the average electricity retail price. Although solar PV is supposed to be advantageous

78

compared to other electricity sources due to the grid parity, the interviewed managers said

that reduction in feed-in tariff was the reason why adoption rate of PV in Southern

Germany reduced in 2012 in comparison to 201124

.

In 2012, the role of policy became more “indirect” than previous years. There is clear

evidence in the case study that policy makers, together with other actors, created a

negative expectation about the electricity prices in future. This legitimization motivated

some households to adopt solar PV systems. Policy measures determine the rules for

calculating the EEG levy (the portion of the electricity price that must be paid by the end

user to support renewable energy) and other types of taxes, which partly sets the electricity

price in the market (Wirth, 2013). In the case study, respondents from both supply and

demand side report the importance of the negative future expectations on the adoption

decision.

“The people in their 30s are afraid of high electricity costs and how to manage

it…” (G.W., Director of a partner firm in HET network, 07.12.2012).

“Electricity will be more expensive in the coming years… then I said I just do not

want to pay much. I would also have advantages (of it).” (F.Z., PV adopter,

24.01.2013).

“We wanted to invest in PV because of the economical and global situation that

we don’t know which way we are going.” (P.M., PV Adopter, 21.12.2012).

Policy makers can influence which conceptual design will become the dominant design

(Utterback, 1994) or disruptive innovation (Christensen, 1997) of PV systems. The

respondents frequently discussed self-consumption, a concept based on an additional

battery system and self-usage of PV electricity. Whether more people will be motivated to

adopt self-consumption concept, depends on economic relative advantage of it, which is

partially shaped by decreasing the feed-in tariff and increasing electricity prices:

“I believe that the PV (diffusion) will grow even though the policy support and

feed-in tariff decrease (…). Now we have feed-in tariff less than what we pay for

electricity, it is 10 cents less ... This low feed-in tariff means that the self-power

24 In Rottenburg am Neckar the number of PV installations in 2012 was 85% of the installations in 2011.

79

consumption is becoming increasingly important when people are increasingly

afraid and further consider self-power consumption...” (G.W., Director of a

partner firm in HET network, 07.12.2012).

Interestingly, corresponding to this tendency, a program to promote small scale storage

batteries for photovoltaic systems has recently started to be subsidized for those that have

been installed after 31st December 2012. The official information was released in April

2013. According to this new subsidy program, the concept of self-consumption PV

systems became more advantageous than any other concept possible (IMS Research,

2013). We could also argue that this new subsidy scheme for solar batteries (direction of

search) may develop positive externalities by the entry of new firms and the creation of a

sub-market. In the field study, we observed (e.g., in the 2-day long annual meeting) that

the Solar Partner e.V. association collaborated with a new solar battery firm to discuss

how to integrate PV systems with battery systems. However, the other functional PV

concepts still remain niche markets. Two examples of these niche markets are building-

integrated photovoltaic systems (a concept in which PV systems are used instead of

roofing material) and façade systems (using PV systems on facades):

“Building-integrated photovoltaic technology is very expensive, very expensive!

You can see it only in France because building-integrated PV systems are there (in

France) subsidized. Building-integrated PV architecture is beautiful, no question,

much nicer than seal down there, which is set up so artificially. It is just a question

of whether the customer can afford it or not.” (G.W., Director of a partner firm in

HET network, 07.12.2012).

Regarding the social acceptance of PV systems (Jäger, 2006), observations in the field

study indicate that policy discussions on renewable energy system is a hot topic in mass

media channels. These discussions even reach the young generations through different

communication channels. One of the respondent´s children reported that the use of

renewable energy systems is such an important topic that they comment and discuss such

systems in primary schools. Theoretically, this importance is related to two of the

perceived attributes of the innovation: compatibility and complexity. Because renewable

energy systems are designated as the key tool for the future of the German electricity

supply by policy makers, the on-going discussions of PV systems decreases PV´s

perceived complexity and makes them compatible with socio-cultural values.

80

5.1.2. ADOPTERS

In line with recent literature (Balcombe et al., 2014, 2013), there is an evidence in the case

study that the contribution to an improved natural environment and the ability to gain

independence from electricity suppliers are important adopter motivations. In particular,

the desire to be independent from electricity suppliers has increased recently because of

the increasing electricity prices. In addition, respondents indicated that PV is often

perceived as an “investment” alternative to other traditional investment options.

“Electricity prices are getting expensive. If you currently have money in the bank,

you have interest rate of about 1 percent ... With a photovoltaic system you can

financially be satisfied…” (B.U., PV adopter, 15.01.2013).

“In the back of my head, I would prefer to be completely self-sufficient and if we

could (in our current new house) make everything for ourselves and would not

need public electricity.” (C.W., PV adopter, 29.01.2013).

“(My motives were) on the one side to make an investment and on the other side to

be environmental friendly, i.e., how I can protect my environment. Then, I thought,

yes, I make PV system on it (the roof).” (F.E., PV adopter, 01.02.2013).

Adopters frequently recognized the impact of their peers upon the adoption rate of PV at

the local, regional, and global levels. At the local level, in line with the literature on peer

effects (Bollinger and Gillingham, 2012; Rode and Weber, 2012), the field study confirms

that there is a positive influence of previously installed PV systems located nearby on the

adoption rate of PV. Peer effects in neighborhoods decrease the perceived complexity for

potential adopters and increase the perceived compatibility with the social norms:

“Several reasons available (for adopting PV). One reason is that we are living in

Tübingen… There are a lot of buildings which have PV systems on the roof…”

(P.M., PV adopter, 21.12.2012).

“Since solar systems were actually built more and more, I have been thinking

about building a house and equip it with a PV system, (since) 10 years.” (H.R., PV

adopter, 23.01.2013).

Finally, corresponding to the regional and global levels, the respondents emphasized that

there was an indirect impact of the Fukushima disaster in Japan (2011), followed by the

81

protests against nuclear power in Germany (2011-2012) and the Stuttgart 21 Project (an

urban development project in Stuttgart, which was protested based on economic and

environmental concerns on 2009-2011). All of these types of protests had an impact on the

decision by the policy makers (on May 2011) to shut down all nuclear reactors by 2022

and on the adoption of the new Act on Granting Priority to Renewable energy sources

(EEG 2012 on June 2011). This reaction to protests is an interesting illustration that

adopters, as change agents, can influence the policy decision and therefore influence the

adoption rate of a technology in society.

5.1.3. LOCAL SOLAR COMPANIES

Local solar companies offer and install complete PV systems to adopters and act as change

agents, thus making them vital in the adopters’ decision process. Corresponding to earlier

research (Reardon, 2009; Bodas-Freitas et al., 2010), a mix of all decision types were

reported in the field study: “optional decision”, in which the location of PV system

installation belongs to the adopter; “collective decision”, in which multiple adopters live

in a common place (e.g., apartment) and have a say in the final decision; and “authority

decision”, in which the adopter is temporarily occupying the location at which the PV

system is installed (e.g., renter). The adoption decision is unlikely to be positive if more

than one individual has a say in the decision, as in the collective decision types. However,

regardless of the type of decision, local solar companies can offer alternative PV system

solutions based on the different needs, which can affect the final decision, e.g., if a

potential adopter is building a new house, the local solar company can offer a building-

integrated PV system to reduce the amount that would have to be spent for conventional

building materials because a building integrated PV system is directly used in parts of the

building envelope (increase of the relative advantage). Here, offering the best solution that

fits the needs of the potential adopter (decreasing complexity and increasing

compatibility) is not only a typical role of an experienced local company but also a key to

motivate the adoption.

Respondents emphasized that the decision to adopt photovoltaic systems requires some

level of knowledge about the technology, operation and funding. Many studies, e.g.,

Dewald and Truffer (2011, 2012), have highlighted the importance of effective

communication and the sharing of similar background between suppliers and adopters.

This importance of effective communication is related to the variables determining the

82

adoption rate, given the fact that such effective communication minimizes the perceived

complexity via change agent efforts, such as a local solar company.

Another finding of our case study is that local solar companies are also important in

producing peer-effects. Previously, the understanding of peer-effects have been limited to

the interactions between adopters-to-adopters in the literature (Bollinger and Gillingham,

2012). The case study reveals that local solar companies may also influence the adoption

rate via the neighborhood effect on potential adopters due to image motivation. HET has

PV systems on its own roof, which is visible from the neighborhood, and offers to open its

doors of its solar center to anyone interested through its periodic solar-walks and open-

door days. These interactions improve the trialability of PV systems as perceived by

potential adopters.

“(...) we are a local company and we have many relative reference systems here in

the neighborhood, and yes I think that is well known. Plus, the combination of the

brand Thomas Hartmann has built, solar center, solar walk, that talking to

feelings, not just to purely technical side, also showing off the products. There is a

building here. That is always the most important, if one builds a place where

people can see. Not every customer care about the purely technical data, but they

want to experience.” (S.L., Solar Thermal Expert by HET, 31.01.2013).

“(Why have you chosen HET?) Because I work nearby and I always have contact

with HET.” (H.J.R., PV adopter, 29.01.2013).

5.1.4. SUMMARY OF THE FINDINGS

Why and how do the PV systems at grid parity get adopted at micro level? Primarily, the

adopters are usually motivated by financial and environmental reasons. This means that,

by adopting a PV system, the adopter believes to gain financial benefit or contribute to the

environment, and sometimes both. However, as the PV systems require a fundamental

change at adoption place (e.g., installing PV panels at the rooftop and inverter in the

house), financial and environmental motives are not enough the drive the diffusion. This

means, achieving grid parity, cannot easily boom the diffusion of solar PV systems.

Therefore, the change agents and peers, e.g., previous adopters, entrepreneurs and policy

makers, do play an important role on motivating the households for adopting solar PV

systems.

83

5.2. THE BUSINESS MODELS

The second research question of this Thesis was: How do business models of local solar

firms and the diffusion of solar PV systems at grid parity affect each other? In order to

answer this research question, we look at the business model of HET for PV systems,

which is based on developing projects of PV installations. These projects start with the

declaration of interest by potential adopters, i.e., request of quotation. When a potential

adopter, e.g., a household, declares interest, a representative of the company visits the

place on site where the PV system is planned to be installed. Depending on the

characteristics of the place and the preferences of the potential adopter, e.g., the roof

space, the sunshine potential and the budget of the potential adopter, HET drafts a project

and creates the sales quote (estimate). The sales quote shows what costs would be

involved for the prospective PV installation. If the potential adopter agrees on the sales

quote, he/she confirms this to HET. Based on this confirmation, HET then installs the PV

system on the place as planned. The system components needed for the installation are

provided by supplier companies, among others, PV module or inverter producers. In the

next sub-sections, the business model components are explained, following Morris et al.

(2005) and linking with relevant literature.

5.2.1. HOW DOES THE COMPANY CREATE VALUE FROM PV SYSTEMS?

PV system installation is not a simple task. It is bound to both adopters’ preferences and

the physical availability of the different types of places, e.g., the rooftop, garage, façade or

free-land. This means that each household needs a particular solution for installing the PV

system. In this context, HET creates value from developing projects of PV installations

that provides solutions for particular needs. Figure 21 presents some examples that HET

has installed for various needs in the neighborhood. The case of HET is an example of

product/service mix offering. In order to let the potential adopter know more about the

diverse PV applications, HET organizes monthly solar-walks since 2001. Such solar-

walks take place in the neighborhood of HET and can be attended by anybody interested.

In these three-hour walks, the attendees get the chance to visit a number of houses that

have already installed PV systems by HET. This helps the potential adopter to decide

which type of PV application is the best for their individual needs. Such close

84

relationships with potential adopters cannot be easily offered by other types of companies

in energy sector, e.g., utility companies (Richter, 2013b).

A semi-transparent garage roof (left);

A roof with shadows (middle);

A flat roof with standing modules (right).

FIGURE 21- DIFFERENT PV SOLUTIONS FOR VARIOUS NEEDS AND CONCEPTS

5.2.2. FOR WHOM DOES THE COMPANY CREATE VALUE?

The market which HET competes is a business-to-customer (B2C) type. HET’s primary

market segment is the small-scale homeowner PV systems. This market segment is usually

composed of decentralized on-grid systems with a capacity up to approximately 10 kWp

(Dewald and Truffer, 2011). Customers, i.e., potential PV adopters, are local:

approximately spread to 50 km radius, instead of being national or international (see

Morris et al., 2005). This local area is a part of Swabia, which is a major part of zip code

area 7. The customers of HET are usually Swabian households, farmers and local small

sized companies, who have a sufficient space on site to install PV system and who are

financially stable. These households speak a special dialect of German, called Swabian,

which is also spoken in HET as an internal company language. This enables HET to better

communicate with potential adopters in this local area. Thomas Hartman is personally

known by many adopters.

„(Why have you chosen HET?) Because I work here nearby and I always have

had contact with HET.” (Mr. R, 29.01.2013)

5.2.3. WHAT IS THE COMPANY´S SOURCE OF COMPETENCE?

HET’s source of competence is composed of several interrelated factors. The first is the

fact that the company is led by a local entrepreneur, who knows the local traditions and

lifestyle very well. The second is the visibility of the company, which is driven by

85

complementary social activities that potential adopters can have at the company. This

includes having the opportunity to dine at the company´s restaurant where the food is

cooked by using renewable sources; participating in the monthly three-hour solar-walks

(to observe the previous projects of HET); and touring HET on its annual open-door days.

The third factor is the quality of the PV installations, mainly based on German solar

modules and inverters. In a nutshell, the competence of HET is what Morris et al. (2005)

call an intimate-customer type. As an HET engineer noted,

"Probably because we are a local company and we have many good reference

systems here in the neighborhood, and yes I think we are well known. Plus, the

combination of the brand Thomas Hartmann has built (is very important). Solar

center and solar walks talk to the feelings; they are not just about purely technical

side. This shows off the products. There is a building here. That's always the most

important thing, if one builds a place where people can see. Not every customer

care about the purely technical data, but they want to experience” (An Engineer

in HET, 31.01.2013)

5.2.4. HOW DOES THE COMPANY COMPETITIVELY POSITION ITSELF?

“I received two offers; one from Hartmann and one from other guy. I did the

comparison but I choose the one which was expensive. It is a local installation

company, well-known in the region, there are many references in the area, and he

is a very well-known guy. Go for the local1” (PV system adopter, Mr. M,

21.12.2012)

“He (Thomas Hartmann) is trustworthy. I prefer local craftsmen and German PV

modules, no imports, and for this I am willing to spend more money” (PV system

adopter, Mr. R., 23.01.2013)

HET intends to achieve an advantage over its competitors by providing local products

with a good quality service. As HET is pioneer in the region, the company has many

reference projects that instill trust in HET among adopters in terms of the quality of the

installations. Previous research based in other countries (Müggenburg et al., 2012; Palit,

2013) also argues that quality of PV systems are vital for potential adopters’ decisions.

According to one adopter,

86

“We had three companies in the race; we asked them all for an offer. We thought

we choose the company that is already well-known, which always works for

ideological reasons (environmental reasons). Other companies simply appeared

on the market in recent years, they want to sell a lot... All three were regional. We

want to buy from our local neighborhood- as much as possible- because you have

reliable people here. Many years of experience and highly trained staff were the

deciding factors for HET. HET was the most expensive, but not so much” (PV

system adopters, Mrs. W. & Mr. H., 29.01.2013)

Most adopters believe that Thomas Hartman is driven by an ecological idealism and he is

a trustworthy entrepreneur. Such trust towards local solar companies is usually associated

with the green movement, and, in particular, against nuclear power (Dewald and Truffer,

2012, p. 409). This seems to be one of the main reasons of why several potential adopters

choose HET, instead of its competitors that entered the market several years after HET,

when policy makers started to support PV installations through the EEG law in 2000s. If a

solar company was established after 2000, potential adopters often perceive it as profit-

oriented.

5.2.5. HOW DOES THE COMPANY MAKE MONEY?

Figure 22 presents a simple scheme of the value network of HET. In Germany, policy

support is not provided by subsidies but through the EEG surcharge (see H. Wirth, 2014,

p. 22). Anybody who uses electricity has to pay taxes on the EEG surcharge, which funds

the Feed-in-Tariff. As a result, the Feed-in-Tariff pays for adopters, e.g., to whom HET

installed PV system, for the electricity generated with fixed feed-in tariffs for the next 20

years from the time the PV system is connected. Such a guarantee enables potential

adopters to pay the high installations costs, which are usually passed on to local solar

companies, e.g., HET. This mechanism is similar to the end-user owned residential type in

the PV sector in the USA (see Frantzis et al., 2008, p. 10). As Dewald and Truffer (2011,

p. 292) also argues, such mechanism relies heavily on the policy support generated by the

EEG surcharge.

87

FIGURE 22- THE VALUE NETWORK OF A LOCAL SOLAR COMPANY

5.2.6. WHAT ARE THE COMPANY´S TIME, SCOPE, AND SIZE AMBITIONS?

Thomas Hartmann invests to the point that HET is able to generate stable income, known

as income-model according to the Morris et al. (2005). Moreover, the company is local,

targeting a market up to 50 km radius and aims to stay local. This vision of staying local

can be associated with two reasons. The first is the transportation costs of PV systems.

The greater the distance between the adopter and HET is, the greater the cost of

transportation of the PV system is. The second reason concerns the competences of

Thomas Hartmann: a good knowledge about the potential adopters and his reputation in

the neighborhood. The further the potential adopters are from HET, the less HET knows

about them and the less HET is recognized by them. This partially relates to the literature

on local entrepreneurship (Michelacci and Silva, 2007), which indicates that

entrepreneurs, working in the region where they were born, have more chances to exploit

the economic opportunities available in their local region.

88

5.2.7. THE BUSINESS MODEL VS. DIFFUSION

The number of yearly PV system installations decreased in 2011, 2012 and 2013 at the

local, sub-national and national levels in Germany25

. This decrease can be observed in the

sales of HET as well. Figure 23 presents this decline in different observation levels. The

decrease in the number of PV installations is usually interpreted as a result of rapid decline

in the Feed-in-Tariff from 2010 to 2012 (see e.g. EPIA, 2014b). In spite of such

interpretations, it is also assumed that the PV systems in Germany reached grid parity by

2012, making the technology advantageous against other conventional sources (Lettner

and Auer, 2012b; Spertino et al., 2014).

FIGURE 23- THE COMPARISON BETWEEN DIFFUSIONS OF PV SYSTEMS AT DIFFERENT LEVELS26

In any case, the decline in the number of installations resulted in a sharp decrease in

HET´s revenues in 2011 and 2012. This was a challenge because the running costs of the

company, such as the costs for the human resources needed for engineering and

installation, stayed at the same level as before. Figure 24 shows the evolution of HET´s

revenues from 2004 to 2012, revealing the decline in the last two years. In general, such

declines sometimes do not last long and may result in re-incline (Rink and Swan, 1979).

However, whether the decline stage will continue is not easily predictable.

25 If the capacity of installations is considered, the decrease happens only in 2012 and 2013. 26 Sources: The Information Platform of four German Transmission Network Operators for EEG and KWK-G; and HET)

89

FIGURE 24- REVENUE OF HET FROM PV INSTALLATIONS AND OTHER SERVICES AND PRODUCTS27

One additional reason for the sharp decline in the revenue of HET is also decreasing

turnover per PV installation (see Figure 25). For example in 2006, the average price of a

PV system that HET installed was around 60000 Euros. This average price decreased to

around 30000 Euros in 2012, probably as a result of a learning curve effect (Spence, 1981)

on the price of system components such as solar modules (see Nemet, 2006 for other

possible reasons beyond learning curve).

27 The values are approximate.

90

FIGURE 25- THE DECLINING AVERAGE COST OF PV INSTALLATIONS28

Business model conceptualization considers the process of doing business and explains the

challenges in operationalizing the construct (Zott et al., 2011). In this context, we

recognize a critical challenge for HET. Similar to the other actors in the renewable energy

sector (Wüstenhagen and Menichetti, 2012), the challenge of HET has been bound to the

risks and opportunities, created by the policy measures. The challenge lies not only on the

decreasing revenue but also on the possible strategies to tackle this decrease. In general,

such revenue decreases, e.g., which is associated with declining adoption or costs, can be

tackled by different strategies such as launching incremental or disruptive innovations in

the market (see e.g., Christensen, 1997) or expansion into national and international

markets with the existing innovations (see e.g., Beise, 2005; Johanson, 1978)29

. The

former strategy is not easy the case of HET because the existing adopters are bound to the

20-year lasting EEG support. In addition, these adopters usually do not have extra space

left at their houses for new types of installations. The latter strategy can be a possible

option, but it contradicts HET´s desire to stay local. HET does not plan to expand its

market because its source of competence is highly associated with its local market.

28 The values are approximate. 29 Such a challenge might be also tackled by different strategies, e.g., introducing new products or services in the market.

However, we do not aim to give a full picture of all possible options that the company has considered or might consider

in the future. Instead, we exemplify only a few most relevant options.

91

5.3. MESO LEVEL ANALYSIS

The third research question of this Dissertation has been: How do the PV systems at grid

parity diffuse at meso level? In order to shed light on this question, we introduced an

indicator-based approach (see sub-chapter 4.2), the results of which are discussed further

in the following sub-chapters.

5.3.1. ASSESSMENT OF LEAD MARKET ATTRIBUTES AT SUB-NATIONAL LEVEL

Table 7 presents the spatial visualization of each indicator. In addition, we rank the states

according to these indicators. Because the weights of lead market attributes are

unidentified in the literature , we use a simple method to aggregate the rankings of

indicators (similar to Beise & Cleff, 2004). Figure 26 presents the box plots, showing the

median and percentiles. These box plots are generated according to the rankings in each

indicator, with 1 representing the best and 16 the worst. By using box plots, we also

ensure that the results are not heavily influenced by possible correlated indicators. If there

are some collinear indicators, this will not change the upper and lower borders of the plots.

FIGURE 26. BOX PLOTS OF INDICATORS FOR EACH STATE

92

TABLE 7. THE SPATIAL VISUALIZATION OF THE INDICATORS FOR LEAD MARKET ATTRIBUTES

[1/L

CO

E]

(kWp/€) [Gre

en]

(%)

[Em

plo

ym

ent]

(%) [Housi

ng]

(%)

[R&

D]

(%) [Pat

ent]

(%)

[Dem

onst

rati

on]

(%) [Fore

igner

s]

(%)

[Export

] (%

)

(%) [Init

iati

ves

]

(%)

93

Baden-Württemberg and Bayern have relatively better rankings in indicators than the rest

of the states. These two states appear as the best in [1/LCOE], i.e., cost advantage,

[Employment] and [R&D] and no worse than 9th

in any indicator out of 16 states. It means

that the presence of lead markets attributes at sub-national level may be observable. The

city-states Hamburg and Bremen present an interesting example. Although any of these

city-states appear to be top in some indicators, such as [Foreigners], [Patents] or [Export],

they are both ranked worse than 12th

in other indicators, such as [1/LCOE] and [Housing].

5.3.2. SPATIAL DIFFUSION CURVE ANALYSIS

Figure 27 shows the spatial curves of small scale PV diffusion in Germany, which is

similar to the proposed pattern of Beise´s model (Beise, 2004). As predicted by the theory,

the states of Baden-Württemberg and Bayern, mainly zip code zones 7, 8 and 9, do have

not only advantages in most of the indicators, therefore, in lead market attributes, but also

higher adoption rates.

FIGURE 27. THE SPATIAL DIFFUSION PATTERN OF PV IN 10 ZIP CODE AREAS OF GERMANY

The lead market model (Beise, 2004) also assumes that lead markets have a rapid growth

in the initial phase whereas the lag markets start to catch up in the later phases; and that is

what we partially observe in the case of small scale PV market at sub-national scale. As

illustrated in Figure 28, the spatial lead market (zones from seven to nine) grew faster than

94

the spatial lag markets (zones from zero to six) between 2000, the year German

Renewable Energy Act (EEG) was implemented, and 2004, the year the first amendment

EEG was carried out. After 2004, the growth rate of spatial lag markets overtook the

spatial lead markets’.

FIGURE 28. THE GROWTH RATE FOR CUMULATIVE NUMBER OF PV INSTALLATIONS IN GERMANY

Yet, the case of Bremen and Hamburg is, again, an interesting example of which

additional insight lead market model could provide. One can easily assume that Hamburg

and Bremen do not score high in market development because of their city-state

characteristics, e.g., limited space on roofs, or the limited sunshine in the north. However,

what might be overlooked is the fact that the growth rate in lag markets, e.g., Bremen and

Hamburg, overtake the lead ones, e.g., Bayern and Baden-Wuerttemberg.

Having said all that, we are aware of that the diffusion of PV systems has neither been

saturated in the lead markets nor in the lag ones. The convergence in growth rates does not

necessarily mean that lag markets will fully catch up with the leads in the future. This is

especially important to understand for a technology like PV systems that are exposed to

the new ways of policy-intervention (e.g., new German subsidy scheme for solar battery

adopters since 2013) and technological breakthroughs, e.g., advancement of efficient solar

batteries (which are complementary to PV systems), increasing efficiency of PV systems

and the disruptive innovations on PV installation practices (see e.g. The Economist, 2015).

Such developments might revive the spatial lead markets, where the growth rates had

recently been decreasing.

95

5.3.3. NATIONAL REGULATIONS IN AN INTERNATIONAL CONTEXT

The extant literature has acknowledged the necessity of understanding not only the spatial

dimension in advancing sustainable transitions (Binz et al., 2014; Truffer and Coenen,

2012) but also the dynamics of diffusion processes for policy-driven environmental

innovations, especially in the renewable energy sector (Quitzow et al., 2014). Beise and

Rennings (2005) state that strict national policy might result in the creation of lead

markets when complemented by global diffusion of regulation and demand. This resonates

with the study of Quitzow et. al. (2014), which argues that international diffusion of

environmental innovations are bound to either diffusion of successful regulations or the

decrease of the financial costs. In the case of PV systems, the German feed-in tariff had

not diffused to other countries (in its effective form), nor had the costs of electricity

generation from PV systems been competitive until 2012. As the lead market model

implies, we assume that any region outside the lead market is potentially a lag market.

This means that in a new era when the LCOE of PV systems become financially

compatible with other sources of electricity generation in some regions (e.g. see Breyer

and Gerlach, 2013; Spertino et al., 2014) and the updated policy support schemes in

emerging countries, e.g., China (Zhang et al., 2014) , we can also observe the recent take

off, e.g., approx. since 2013, in other regions outside of Germany.

Our study acknowledges that in order to achieve National Renewable Energy Action Plans

(NREAPs), which are heterogeneous among countries (see e.g. Beurskens et al., 2011), it

is important to understand the underlying regional dynamics of diffusion. The PV systems

in Germany is expected to achieve the NREAP for 2020 earlier than planned (EPIA,

2013). The German case is an illustrative example of how national-wide regulations

induce sub-national lead markets which consequently results in national diffusion. As the

German EEG was able to drive the diffusion of photovoltaic diffusion in a particular sub-

national region that has innovation-specific lead market attributes, which are Baden

Württemberg and Bayern, the diffusion in other regions has incrementally followed the

lead ones. This can be important to understand for policymakers as similar developments

might occur for environmental innovations in other countries. When developing national

policies, it is essential to assess the lead market potential of sub-national regions, and,

consequently, develop regulations responding to the demand of the potential lead market.

96

5.4. MATHEMATICAL MODEL

The fourth research question of this thesis has been: How can the diffusion of solar PV

systems be better forecasted at meso level? Keeping this in mind, I propose a finite

element method to forecast the solar PV systems (see chapter 4.3 for method details).

Through this method, the governing equation on the diffusion rate (D) has been solved at

43 incremental time steps through the ABAQUS software (Abaqus, 2013). This

calculation lasted for 35 seconds30

. Based on this solution, it is possible to obtain the

values of diffusion rate (D) for any spatial point in any given time. For example, the

diffusion rate in each spatial location after 13 years of diffusion is plotted as a counter map

in Figure 29. The red color accounts for 100% diffusion rate, while blue color indicates for

around 9% diffusion rate. The HET is located on the red area and that is why this area has

100% diffusion ratio. The zip code area 72149, which has the lowest innovation

diffusivity, has also accounts for the lowest diffusion ratio after 13 years. In line with the

literature (see e.g. Morrill, 1970; Rogers, 2003), not only the distances between potential

adopter and innovation source but also innovation diffusivity seem to be decisive for the

heterogeneous diffusion rate that is spread in the given geography.

In order to see the differences on diffusion patterns on different spatial points, the 5

locations, namely as A, B, C, D and E (as shown in Figure 29), are additionally analyzed.

The resultant curves present the evolutions of diffusion rate (D) on given points as

illustrated in Figure 30. The curves have a typical s-shape as studied by literature (Bass,

1969; Rogers, 1962), showing that the selected locations are at a variety of stages of

diffusion process. If a location is nearer to the HET, it is most likely to have a rapid

diffusion. However, the diffusion is also bound to the innovation diffusivity. For example,

the location B has a slower diffusion than the location A, although the location B is closer

to the HET than the location A is. This can be explained due to the fact that location B has

lower innovation diffusivity value than the location A does.

30 Such calculation is almost impossible to conduct without a computational software

97

FIGURE 29- DIFFUSION RATE AFTER 13 YEARS

FIGURE 30- THE PATTERNS OF S-CURVES AT 5 LOCATIONS

The interrelation among the s-curves (as presented in Figure 30) are similar to those of

what Hägerstrand (1967) and Morrill (1970) proposed. The patterns resemble the wave

propagation. With the words of Morrill (1970), spatial heterogeneity causes the diffusion

to spread gradually in space and time. Such diffusion pattern can be also recognized in the

lead market model of Beise (2004) which argues that some regions may have a rapid

diffusion with a leading role. The differences on diffusion patterns of innovations both in

space and time are commonly associated with the characteristics of the potential adopters,

such as socioeconomic factors, and the environmental context, such as geographical and

political settings (Wejnert, 2002). Therefore, it can be assumed that the innovation

diffusivity c(x,y) contains the combined information of the adopters and the

98

environmental context. This means, for example, the low innovation diffusivity value of

zip code area 71148 may represent the unfavorable conditions of this region, e.g., lack of

households or low income level etc. (see the literature on PV systems’ diffusion for other

possible explantions, e.g., Fabrizio and Hawn, 2013; Zhang et al., 2011) .

The Finite Element Model also gives detailed information on the innovation flux (q),

which is not possible to obtain neither from the one dimensional conventional models

(e.g., Bass et al., 1994; Bass, 1969) nor from those of that study the innovation diffusion at

meso level (e.g., Gooding et al., 2013; Higgins et al., 2014). Figure 31 presents the

evolution of innovation flux for a time period of 10 years. As expected, the innovations

flow from the HET to the other regions. However, the innovation flux is not homogenous

in all directions. For example, the flux is higher into the direction of zip code area 72108

than the one into the zip code are 72119. This is expected due to the fact that zip code area

72108 has higher innovation diffusivity than that of 72119 has. One can also observe that

the maximum value of innovation flux decreases with the time. This can be explained with

the diffusion rate differences among regions. In the beginning, e.g., after 3 months, the

high diffusion rate differences result in higher flux. However, in the later phases, e.g., after

10 years, the diffusion rate differences among regions decrease, resulting in lower

innovation flux. In addition, the result of propagated evolution in innovation flux can be

observable in the evolution of diffusion ratio (see Figure 32). Diffusion ratio increases

heterogeneously on the space as driven by the heterogonous innovation flux.

99

FIGURE 31- INNOVATION FLUX AFTER 3 MONTHS, 4 YEARS, 7 YEARS AND 10 YEARS

FIGURE 32- DIFFUSION RATIO AFTER 3 MONTHS, 4 YEARS, 7 YEARS AND 10 YEARS

The findings of the Finite Element Model on the case of PV systems´ diffusion have

important implications for scholars who seek to study the diffusion of innovations as a

space-time process. The Finite Element Model applied in this Thesis presents a

comprehensive explanation on the spatial patterns of the diffusion. In addition, the model

serves as a powerful method to solve the partial differential equations which have infinite

number of unknowns. This means that the Finite Element Models can be also easily

applicable to the other mathematical formulations of innovation diffusion that are based on

100

partial differential equations (e.g. Bass, 1969; Mahajan and Peterson, 1979; Morrill,

1970).

5.5. A META-ANALYSIS

This sub-chapter offers a meta-analysis31

that contrasts and combines implications from

the analysis conducted at different levels. The overarching discussions are organized

according to nine topics, which lay in the intersections of the findings of diverse analyses

(see Figure 33 for the diagram). These discussions offer additional insights for answering

the RQs, by providing a meta-analysis, which is not necessarily available in the previous

sub-chapters.

FIGURE 33. DISCUSSIONS DIAGRAM32

1) Today, many innovations that are widely adopted are easy to adopt and easy to de-

adopt33

. There is a myriad of examples that point to this: smart phones, electrified

bicycles, virtual smart-phone applications, Facebook, computer software programs and

fashion-driven textile clothes. One can adopt them in any time, and quickly de-adopt them

in any other time. The decision process sometimes does not exceed a couple of minutes.

31 The term meta-analysis is usually used for statistical studies or literature reviews (Card, 2011). In this Dissertation, the

term is used for overarching analysis of the findings. 32 Circles correspond to the findings of analyses; while bold numbers correspond to discussion topics in the intersection

of these findings. 33 By using de-adoption term, I refer to discontinuance which is the decision to reject an innovation after having

previously adopted (Rogers, 2003, p. 182)

101

They do not also have any long-term remaining and visible consequences on the adoption

units. When de-adoption occurs, the main consequences of the adoption quickly disappear.

However, solar PV systems present particular differences from such kinds of innovations

that influence the perception of potential adopters and, consequently, the diffusion

process. One cannot easily adopt solar PV systems; nor in 5 minutes neither in 1 day. In

the high-income economy context, solar PV systems need a relatively high financial

investment by the adopter unit, e.g., twenty-thousand Euros (€) for a small single-family

house in Germany by 2013, and have long-term effects on the adoption units. By adopting

the PV systems and benefiting from the feed-in tariff, the adopter makes a tacit agreement

that binds him/her/them for the next 10 years. Once the adoption takes place, the PV

adopter cannot easily de-adopt it anymore, at least in a short-term time. It is because the

financial benefits of adopting PV systems appear in long-term; and adopters naturally

prefer to wait for years to receive these benefits. In the context of low-income economies,

adopting solar PV systems usually means “electrification”. The potential adopters are

often those who had been used to live without electrification. Adopting a PV system do

change the lifestyle, e.g., they can use mobile phones and charge them with the electricity

generated by the PV systems or read a book under a lamp, electricity of which is generated

by the PV system. In both contexts, adopting a PV system become easily visible by

others, making it a symbol of lifestyle and, sometimes, social status. Adoption has long-

terms benefits for adopters in both low-income and high-income economies. Therefore, in

a nutshell, “solar PV systems are neither easy to adopt nor easy to de-adopt”.

2) In general, it seems that several supply and demand side factors might affect each other.

For example, the production choice of the firms affects what customers buy in the market,

or the vice-versa: what customers buy in the market affects the production choice of the

firms. On the one hand, one can easily observe correlation between events but it does not

necessarily mean causality. On the other hand, one can also observe several versatile

causalities between several sets of events. One way of understanding the relation between

events could be, to some extent, conducting in-depth studies. In the case of Rottenburg am

Neckar, the relation between the activities of the solar firm and the diffusion of PV

systems seem to have influenced each other. The solar firm influences the potential

adopters in favor of adoption of solar PV systems, resulting in a direct effect on the local

diffusion. However, the wide adoption of solar PV systems (driven by many other local

solar companies all over the country) leads to price reductions on solar PV panels and

102

their installations (due to economies of scale). Consequently, the reduced prices result in

decline on the turnovers of solar local companies- giving rise to business model

challenges. Although this has a context limited basis, i.e., Germany, the review on PV

systems (sub-chapter 3.2.2) presents that local firms or other bodies of intermediaries, are

important drivers of the diffusion in other countries as well. Therefore, such challenges

(and versatile interplays) might occur in other contexts, including both low and high

income level economies.

3) Nowadays, many internationally adopted innovations have been built upon the

international diffusion of their standardized complementary innovations. For example, the

international diffusion of high quality bike tires have built upon the wide international

diffusion of standardized bicycles. If an individual does not have one of those bicycles,

there is no way that can lead to adoption of a high quality tire. However, it is not the case

for for small scale PV systems. It is mainly because there is no standardized way of

housing around the world. In some regions, households do have proper spaces for the

installation of solar PV systems, while in some others, they do not. This raises an

important contrast between the firms’ business models located in different parts of the

world. For example, in low-income economies, after-sale services can be an important part

of firms’ businesses- mainly due to poor maintenance standards. However, in high-income

economies, developing building integrated PV solutions can be more suitable to be

offered, as the potential adopters might be more willing to pay more for aesthetical

reasons.

4) Comparing the motives for diffusion of PV systems at adopter level and the attributes

of leading regions, one can observe some similar dynamics. Some individuals –with

certain characteristics- might adopt the PV systems at an earlier phase than others and can

motivate their peers in the neighborhood for the adoption of PV systems, i.e. peer effects

(Bollinger and Gillingham, 2012). In the same token, some regions- if they have the lead

market attributes- might have some leading effects on other regions, i.e., lead market

hypothesis (Beise, 2001). When these different observation levels could be combined, one

can analyze the cascading effects among the levels. Adoption takes place at adopters’

level, but the diffusion occurs at the regional and national level. If the early adopters of an

innovation do affect their peers and the pioneer regions do affect other regions (in favor of

the adoption of an innovation), diffusion can outreach to national and global levels.

103

5) It is a hard task to forecast the behavior of individual adopters or firms. Although two

potential adopters in the same neighborhood seem to be similar, one can adopt the

innovation while the other not. In the same token, two firms in the same sector may seem

to have similar business models. However, one can have an incline in the revenue, while

the other can have a decline. That said, if the observation is conducted at a system level,

e.g., regional level, one can have some analytical basis to model and forecast the systemic

behaviors e.g., diffusion of innovations. The regional trends become observable and

forecasting, based on the historical data, can be possible.

6) The well-known Bass (1969) model has been used by firms to develop marketing

strategies for decades. Similarly, one can use the finite element model. For example, by

using the spatiotemporal historical diffusion data through a finite element model, firms

can forecast the declines and inclines in specific regions and develop strategies to harvest

the maximum out of such diffusion trends. This is specifically important for the cases, at

which the spatial proximity matters. For example, in the case of Hartmann Energietechnik

GmbH, the firm offers installation services for the potential PV adopters. It means that the

firm needs to transport the PV panels to the households when the installation takes places.

The more the potential adopter is far from the firm, the more costly the PV system is to be

transported. This means that, if a firm is able to forecast the spatial trends of diffusion, the

firm can optimize its strategy in a way that it can harvest the increase of diffusion rate in

some far regions but, at the same time, keep the transportation cost as low as possible. One

possible way of doing this could be setting up new branches of the firm at high-growing

regions which might be located far away.

7) Sustainable energy transition is a complex process and designing policies to enforce

this transition is challenging. It has multi-dimensional aspects, e.g., management,

economics, sociology and marketing (see sub-chapter 3.2.1). Yet, one of the key

approaches to design effective policy measures could be lying on the better understanding

of the regional dynamics. The countries lagging in such transitions could design measures

that fit well to the potential regional lead market of the respective countries, instead of

transferring other policy measures from foreign countries. In such way, lead regions of

diffusion can be induced and, subsequently, the national adoption can be driven by the

lead region- as it happened in the case of diffusion of solar PV systems in Germany (as

argued in sub-chapter 5.3).

104

8) Climate change is a global issue. Reducing the emissions in some particular countries

do not necessarily lead to stopping the climate change. For example, in the energy sector,

a wide diffusion of renewable energy technologies in particular countries might

surprisingly give a rise to reduction on the prices of fossil fuels (because of lack of

demand in the countries that do not need any more fossil fuels), which can result in more

use of those fuels in other countries, accelerating the climate change. Therefore, diffusion

of renewable innovations should have a global outreach. This has already been addressed

in the literature on international diffusion of environmental innovations, particularly

trough the lead market concept. That said, if an immediate action is needed to be taken

against the climate change, one can be interested to be able to forecast the diffusion

patterns. In this token, the leading roles of regions (the regions that has the five lead

market attributes as proposed in the indicator-based analysis) on the global diffusion of

innovations could be modeled and forecasted with advanced methods such as finite

element method.

9) The strong similarities between how salt diffuse in water and how innovations or

diseases diffuse in societies attracted many scholars to model the diffusion of innovations

through a variety of approaches inspired by natural sciences. However, the social

processes are complex and it is hard to measure the micro characteristics of societies that

mind hinder or drive the diffusion of innovations. In case of environmental technologies,

not many innovations have been able to diffuse to a larger extent. Yet, many scholars have

recently attempted to model and forecast the diffusion of environmental innovations

through agent based models (see e.g., the findings of literature review, sub-chapter 3.2.1).

Finite element analysis, by using the historical data on the micro level, is just another

alternative method to tackle the modeling issues. The method can successfully model the

diffusion of salt in water in time and space because we can precisely know the micro

characteristics of the water used in the respective experiment. So why not the diffusion of

innovations? The only need might be the big data collected at micro-level.

105

6. CONCLUSIONS, CONTRIBUTIONS AND

LIMITATIONS

6.1. THEORETICAL IMPLICATIONS AND CONTRIBUTIONS

6.1.1. GRID PARITY: A DISILLUSION?

The literature on environmental and renewable energy innovations has assumed that the

high cost has been the main barrier for the wide adoption of such innovations (e.g.

Jacobsson and Johnson, 2000; Lund, 2011; Painuly, 2001). There has been a great hope

that achieving the grid parity for renewable energy innovations, or cost competiveness of

environmental innovations in general, would boost the diffusion. A recent study by Munoz

et al. (2014), based on a discourse analysis of the literature, also shows that the majority of

scholars claims that achieving the grid parity is either one or the key milestone(s) of the

diffusion of solar PV systems. However, a few scholars, e.g., Yang (2010) (as cited in

Munoz et al., 2014), expected that grid parity would not necessarily guarantee the wide

adoption of solar PV systems.

To my knowledge, there has not been any study conducting an in-depth analysis of

diffusion of solar PV systems at grid parity to date. The micro analysis of this

Dissertation, therefore, represents the first in-depth attempt that falls into this highly

debated gap in the literature. This analysis illustrates that achieving grid parity has not

necessarily boosted the diffusion of renewable energy innovations, in particular that of the

small scale solar PV systems. This is also in line with the primary finding of the literature

review on PV systems (sub-chapter 3.2.2), which focuses on the barriers to the adoption of

PV systems. This review shows that cost and price related barriers have been only a little

part of the big picture. Most of the barriers, that hinders the wide adoption of solar PV

systems, have been related to the socio-technical, political and managerial dimensions.

Therefore, I can argue that grid party has not boosted the diffusion and neither will it

easily boost the diffusion in the future. This can be a bad news for those who expected

much from the grid parity.

106

6.1.2. LOCAL FIRMS: DRIVERS OR VICTIMS?

In regard to the firms, Penrose (2009, p. 4) suggests that, “a comprehensive theory should

take account of not only the sequence of changes created by a firm’s own activities but

also the effect of changes that are external to the firm and lie beyond its control”. In this

token, business model analysis of this Thesis illustrates both the firm’s activities and the

effect of external changes. These external changes have been the reduction of turnover per

PV project, the decline in diffusion (dynamics of which is evident in the indicator-based

analysis) and decrease in feed-in tariff. All these external factors consequently resulted in

a rapid decline in the revenue of the solar firm and stimulated a business model challenge.

However, as the case study conducted in this Thesis, in line with other scholars (Dewald

and Truffer, 2012; Fabrizio and Hawn, 2013), argue that, local solar firms are the

important drivers of the diffusion, acting as intermediaries among the technology, policy

makers and the adopters. This means that, if the local solar fails to survive, so does the

diffusion. Nevertheless, I can argue that a local actor, which has driven the diffusion, can

also become a victim of the high diffusion rate.

6.1.3. LEAD AND LAG REGIONS

In line with the findings of other scholars (e.g. Fabrizio and Hawn, 2013), indicator-based

analysis confirms that country-level policies do not influence the diffusion of innovations

in the same manner in all sub-national regions. Extending this argument, this analysis also

makes an important empirical contribution to the literature. It argues that, if a national

policy can drive the diffusion of an innovation in a particular sub-national region that has

the lead market attributes, the diffusion in other regions will follow up subsequently.

Therefore, the findings of indicator-based analysis, which is partially anchored in the

findings of case study method, represents a novel context-only extension of the lead

market concept (Beise, 2004, 2001) by applying it at the sub-national level. This is an

important contribution to the literature, given the fact that all previous studies on lead

market concept had been applied to the diffusion of innovations –only- at the national

level.

107

6.2. METHODOLOGICAL IMPLICATIONS AND CONTRIBUTIONS

6.2.1. PLAYING WITH MATH: HOW TO COUNT THE DIFFUSION RATE?

In the majority of the innovation literature (e.g. Bass, 1969; Griliches, 1957; Hägerstrand,

1967; Rogers, 2003), the diffusion rate is calculated through on a simple rule: if a potential

adopter has adopted the innovation, it counted for 1. However, diffusion research on

renewable energy innovations, in this case it is the solar PV system, traditionally has a

different approach when evaluating the diffusion. For calculating the diffusion rate, the

scholars usually have focused on the capacity of electricity generated through installations,

instead of the number of installations (see e.g. Dewald and Truffer, 2012; Guidolin and

Mortarino, 2010). However, this might lead to a misunderstanding on whether the

diffusion is increasing or decreasing. For example from 2010 to 2011 in Germany, there

was an increase in the total capacity of PV installations, while corresponding a decrease of

number of installations. In the case study of this Dissertation, we intentionally used the

number of installations for diffusion evaluation, and we hope that this will become a

common standard for future diffusion studies on renewable energy. Otherwise, the

continuous increase of capacity of PV systems might create a misleading bubble on the

calculation of diffusion rates. For example, in order to calculate the diffusion rate of

electrical cars, no one calculate the sum of the power capacity of all adopted cars. Instead,

the scholars calculate the number of adoptions. Why not to do the same for solar PV

systems? The adoption number of PV installations does tell about diffusion rate, while the

capacity of installations does tell about something else, i.e., technology efficiency.

6.2.2. FINITE ELEMENT METHOD: IN AN NEW ERA OF BIG DATA

Theories on social patterns of organizations and societies have been rooted several

decades ago (e.g. Durkheim, 1933; Marx, 1954; Weber, 1922). However, neither the

pioneers nor the followers had an access to the big data- the newly ubiquitous digital data

that is available now about all aspects of human experiences (Pentland, 2014). Today, for

example, we can precisely know the location of many individuals through global

positioning system (GPS) during 24 hours per day. In similar, we can have an access the

detailed information of diffusion of a variety of innovations through public and private

databases. Consequently, leveraging the big data with high computational power, social

sciences (including economics and management) have tremendously been advancing (e.g.

108

Brockmann and Helbing, 2013; Mani et al., 2013; Pentland, 2013, 2012). In the same

token, the mathematical model applied in this dissertation, using today´s high

computational power, presents an application of the finite element method, which has

traditionally been used in engineering disciplines. If there is access to the big data, the

model might serve as a powerful method for solving the partial differential equations

describing the social phenomena. For instance, in the case of solar PV systems in

Germany, there is a public database on the information of PV adopters, e.g., revealing the

information on precise adoption time, the precise location of adoption and the capacity of

installations- which indeed motivated the development of finite element model.

6.3. POLICY IMPLICATIONS

6.3.1. IGNORE THE LAG REGIONS

How can policy makers promote the diffusion of environmentally friendly innovations?

This question has been addressed in the literature for a long time. When it comes to the

sub-national dynamics, this Dissertation argues that developing sub-national policies (e.g.

as suggested by Fabrizio and Hawn, 2013; Zhang et al., 2011) is not only the effective

way to drive the diffusion of environmental innovations. The results of indicator-based

analysis, conducted in this Dissertation, argue that policy makers might also choose to

develop a national measure that fits the lead sub-national region and, consequently, drive

diffusion in the national level in a progressive manner. This means that if a national policy

is able to drive the diffusion of an environmental innovation in a particular sub-national

region that has lead market attributes (advantages on demand, transfer, export, market

structure and cost), the diffusion in other regions will follow the lead up. Therefore, by

being respondent to the lead market’s demand (and ignorant to the lag market’s), policy

makers can foster and drive national adoption innovations. Therefore, I argue that

policymakers should not be worried about the lagging regions in the early phases of

diffusion. If the diffusion of an innovation in a region with lead market attributes can be

boosted, the lag regions are likely to catch up.

6.3.2. LOCAL FIRMS: ON THE EDGE OF A CLIFF?

Business model analysis, conducted in this Dissertation, argues that local companies,

which act as important sources to motivate the potential adopters, have a business model

109

challenge due to new conditions in the market. Given that cost competiveness of

renewable energy innovations, i.e., grid parity, does not fully foster the wide diffusion,

policymakers might need to consider new policy measures to support local companies. As

the results of the micro analysis indicate that the policy measures are unlikely to have the

expected impact without local solar companies, survival of such companies is critical for

the wide adoption of PV systems. Therefore, I argue that, if the local firms are on the edge

of a cliff (as seen in business model analysis), it means the diffusion of PV systems is on

the edge of a cliff as well. Given that diffusion of solar PV systems has a limited outreach,

even in the most developed market, i.e., Germany, I recommend policymakers to support

the local drivers of diffusion, i.e., local solar firms, which help potential adopters to

understand the benefits of adopting solar PV systems. Otherwise, potential adopters’

negative perceptions on the PV technology (e.g., related to the complexity, trialabilty,

compatibility and relative advantage) - that might hinder the diffusion- cannot be easily

overcome.

6.4. LIMITATIONS AND FUTURE RESEARCH

Given the multi dimensions of the diffusion of PV systems, there are some limitations to

this study, which could motivate to open new avenues for future research. The case study,

conducted in this Thesis, is based on an illustrative case and has some limitations. For

instance, it is only based on a single-case study. However, as argued by Flyvbjerg (2006)

single-case studies can contribute to the collective process of knowledge accumulation in a

particular field that lies among several fields, e.g., energy research and social science. In

addition, the diffusion of PV systems in Germany has been highly influenced by its

particular conditions, e.g., the German feed-in tariff. Therefore, the results of case study

might not easily be transferable to other national markets. However, similar processes are

likely to occur in lag countries where the diffusion might take off recently. Nevertheless,

as with other renewable energy related case studies that focus on particular regions (e.g.,

Darmani et al., 2015; Eder et al., 2015; Sriwannawit and Laestadius, 2013), I also believe

that the case study of this Dissertation can provide empirical insights that can be used as a

basis for future research.

There have also been some more methodological limitations. In the indicator-based

approach, finding the indicators and aggregating them was a challenge. The data for some

110

indicators were not available to collect. For the finite element method, not all the data used

for the modeling was empirically collected. This could be easily overcome by using the

historical s-curve data in future studies. Moreover, the literature reviews (sub-chapters

3.2.1-2) may lack some important keywords to be used for generating the databases. For

example, the terms “acceptance” and “deployment” miss in the sets of keywords in both

studies. Nevertheless, one can argue that these terms do not necessarily correspond to the

diffusion process. That said, including these terms might be useful to expand the scope of

future studies in similar contexts.

Moreover, several topics, which were beyond the scope of this Thesis, appears to be

promising for future research. Firstly, the influence of the recent German policy support

for solar batteries (implemented in 2013) on the adoption of PV systems should be studied

further. Solar batteries let the households to store the electricity they generated. It means

that they will necessarily need to supply the electricity to the grid. Instead, they can store

during the day (when the sun is shining) and use it during the night. Secondly, the future

studies could also investigate the barriers to the adoption of a variety of PV systems’

concepts, e.g., building integrated systems, particularly in times of grid parity. To do this,

several factors can be important to be considered: socio-technical, managerial, policy and

economic factors (as suggested by the literature review, see sub-chapter 3.2.2). Thirdly,

further research can also further develop the indicators used for lead market attributes of

PV at sub-national level. Moreover, scaling down the unit of observation can provide new

insights. Fourthly, studying the role of cost competiveness for environmental innovations

can be a promising future research area. As happened in the case of PV systems in

Germany, reaching cost competiveness does not boost the diffusion, against the

expectations of many scholars. Fifthly, finite element model is a promising starting point

for more detailed and advanced investigations in the future, including the modeling the

diffusion of environmental innovations through analyzing the big data. Sixthly, towards

energy transitions to decentralized energy solutions, adopters are supposed to, as first step,

discontinue using conventional electricity sources (in high-income economies) or

discontinue living without electrification (in low-income economies) (see sub-chapters

3.2.2, and 5.1.). However in energy research, surprisingly, less attention has been paid to

the discontinuance. This topic could be a promising future research area. Seventhly, the

case study in Rottenburg am Neckar also provided additional perspectives about other

technologies that could have competed with solar PV- which go beyond the scope of this

111

Dissertation. One example is the solar thermal systems. The diffusion competition

between solar PV systems and other alternatives could be an interesting future research

topic. Number of yearly solar thermal installations was around 15k in Germany, whereas it

was 18k for PV in 2012. The case study showed that adopters usually decide between

solar PV systems and solar thermal systems. In some cases, they decide to adopt both, e.g.,

PV for half of the roof and solar thermal for the other half. However in some cases, they

choose either of them. According to the number of yearly installations, the time period that

PV had the highest growth corresponds to the time period that solar thermal had the

sharpest decline. From 2008 to 2010, the number of yearly installations of PV has

increased 120%, whereas the number of yearly installations of solar thermal decreased

45% as the sharpest decline in its history. This contradiction supports the argument that

PV and solar thermal compete with each other, at least to get a place in the roof as the

capacity of a roof is often limited. As Thomas Hartmann already questioned in one of the

interviews: “To whom belongs the roof? To PV Systems or to Solar Thermal Systems?”.

Eighthly, given the Germany’s ambitious target to double its energy production from

renewable energy sources in the next 15 years (BDEW, 2014), further research might

investigate how policy measures could support local solar companies, which are used to be

the local drivers of the diffusion. Therefore, further research on local solar companies’

business models could help them to respond to the challenges of decline in an effective

way. In this context, the literature on business model innovation (e.g. Chesbrough, 2010;

Sosna et al., 2010), including the energy related empirical cases of it (Christensen et al.,

2012; e.g. Richter, 2013a, 2013b), and local entrepreneurship (Julien, 2007; e.g.

Michelacci and Silva, 2007) could be complementary to further understand such

phenomenon. Last but not least, the interlinks among the micro, meso and macro levels of

diffusion could be studied further. For example, the future research can focus on how the

macro level diffusion can influence the micro and meso level of diffusion, and the vice-

versa.

6.5. CONCLUSIONS

Conceptualizing the solar PV systems as innovations, this Dissertation has tried to

contribute on both innovation studies and renewable energy research. This has been done

through five synchronized steps. Firstly, -anchored in two literature reviews- it tried to

shed light on the accumulated knowledge of diffusion of solar PV systems and

112

environmental innovations. Secondly, at micro level, it explored the diffusion dynamics of

a unique local case in southern Germany. Thirdly, at firm level, it illustrated a business

model challenge of a leading local solar company in the same region. Fourthly, at meso

level, it explored how leading regions might affect the lagging ones and how it can be

mathematically modeled. Finally, a meta-analysis of the findings previous steps has been

conducted and, than, implications for theory and practice have been discussed.

Among all the contributions discussed in previous (sub-) chapters, two contributions of

this Thesis can be particularly interesting for a broader audience from economics,

geography and management disciplines. Firstly, the Thesis provides in depth-analysis of a

case of environmental technology which has achieved the cost competitiveness (to some

extent). The results of this case are against the arguments that achieving cost

competitiveness would boost the diffusion of environmental innovations such as

renewable energy technologies. Second, albeit the complexity of diffusion dynamics in the

well-connected world of today, the Thesis presents an empirical evidence for the existence

of regional patterns – that can give a rise to our understanding on how particular sub-

national regions lead the rest. This can be perceived as against the claims that today’s high

mobility of individuals and high-connectivity of distanced-regions will diminish the major

effects of spatial proximity on diffusion of innovations.

113

7. APPENDIX

7.1. LIST OF PUBLICATIONS

7.1.1. INTERNATIONAL PEER-REVIEWED ARTICLES IN JCR

Karakaya, E., Nuur, C., & Hidalgo, A. (2016). Business model challenge: Lessons

from a local solar company. Renewable Energy, 85, 1026–1035.

Karakaya, E., & Sriwannawit, P. (2015). Barriers to the adoption of photovoltaic

systems: The state of the art. Renewable and Sustainable Energy Reviews, 49, 60–

66.

Karakaya, E., Hidalgo, A., & Nuur, C. (2015). Motivators for adoption of

photovoltaic systems at grid parity: A case study from Southern Germany.

Renewable and Sustainable Energy Reviews, 43, 1090-1098.

Karakaya, E., Hidalgo, A., & Nuur, C. (2014). Diffusion of eco-innovations: A

review. Renewable and Sustainable Energy Reviews, 33, 392-399

7.1.2. INTERNATIONAL CONFERENCES

Karakaya, E., Nuur, C., & Hidalgo. A. (2014). Business model challenge at

decline: Learnings from a local solar company. EDIM Conference, Milan, Italy

Karakaya, E., Nuur, C., Breitschopf, B., & Hidalgo, A. (2014). Spatial lead

markets at sub-national level. DRUID Society Conference, Copenhagen, Denmark.

Karakaya, E. (2014).Finite Element Model: An Application to Diffusion of

Photovoltaic Systems. EDIM Conference, Milan, Italy.

114

Karakaya, E., Hidalgo, A., & Nuur, C. (2013). Diffusion of Eco-Innovations:

Exploring the Literature. The 22nd International Conference on Management of

Technology (IAMOT), Porto Alegre, Brazil.

Karakaya, E., Hidalgo, A., & Nuur, C. (2013). Modelling the Diffusion of

Photovoltaic: Concepts and Applications. The 22nd International Conference on

Management of Technology (IAMOT), Porto Alegre, Brazil.

Karakaya, E., Hidalgo, A., & Nuur, C. (2013). Diffusion of Photovoltaic: Policy

vs. Adopters. The Eu-Spri Conference, Madrid, Spain.

115

8. REFERENCES

Abaqus, 2013. Abaqus/CAE Student Edition 6.13.

Allegretto, W., Lin, Y., Yang, H., 2001. A fast and highly accurate numerical method for the

evaluation of American options. Dyn. Contin. Discret. Impuls. Syst. 465.

Ameli, N., Kammen, D.M., 2012. Clean energy deployment: Addressing financing cost. Environ.

Res. Lett. 7, 034008. doi:10.1088/1748-9326/7/3/034008

Anderson, C., Zeithaml, C., 1984. Stage of the product life cycle, business strategy, and business

performance. Acad. Manag. J. 27, 5–24.

Angermann, L., Wang, S., 2007. Convergence of a fitted finite volume method for the penalized

Black–Scholes equation governing European and American Option pricing. Numer. Math.

106, 1–40. doi:10.1007/s00211-006-0057-7

Asheim, B., Gertler, M.S., 2005. The Geography of Innovation: Regional Innovation Systems, in:

Jan Fagerberg, David C. Mowery, R.R.N. (Ed.), The Oxford Handbook of Innovation.

Oxford University Press, pp. 291 – 317.

Asheim, B.T., Coenen, L., 2005. Knowledge bases and regional innovation systems: Comparing

Nordic clusters. Res. Policy 34, 1173–1190. doi:10.1016/j.respol.2005.03.013

Atkinson, P., Hammersley, M., 1994. Ethnography and participant observation, in: Denzin, N.K.,

Lincoln, Y.S. (Eds.), Handbook of Qualitative Research.

Baccarin, S., 2009. Optimal impulse control for a multidimensional cash management system with

generalized cost functions. Eur. J. Oper. Res. 196, 198–206. doi:10.1016/j.ejor.2008.02.040

Bain, A., 1963. The growth of demand for new commodities. J. R. Stat. Soc. 126, 285–299.

Balcombe, P., Rigby, D., Azapagic, A., 2013. Motivations and barriers associated with adopting

microgeneration energy technologies in the UK. Renew. Sustain. Energy Rev. 22, 655–666.

doi:10.1016/j.rser.2013.02.012

Balcombe, P., Rigby, D., Azapagic, A., 2014. Investigating the importance of motivations and

barriers related to microgeneration uptake in the UK. Appl. Energy 130, 403–418.

doi:10.1016/j.apenergy.2014.05.047

Barthel, R., Janisch, S., Schwarz, N., Trifkovic, a, Nickel, D., Schulz, C., Mauser, W., 2008. An

integrated modelling framework for simulating regional-scale actor responses to global

change in the water domain. Environ. Model. Softw. 23, 1095–1121.

doi:10.1016/j.envsoft.2008.02.004

Bartley, M., Owen, C., 1996. Relation between socioeconomic status, employment, and health

during economic change, 1973-93. Br. Med. J. 445–449.

116

Bartunek, J.M., Louis, M.., 1996. Insider/Outsider Team Research. Sage. Thousand Oaks (CA).

Bass, F., Krishnan, T., Jain, D., 1994. Why the Bass model fits without decision variables. Mark.

Sci. 13, 203–223.

Bass, F.M., 1969. A new product growth for model consumer durables. Manage. Sci. 15, 215–227.

Bawakyillenuo, S., 2012. Deconstructing the dichotomies of solar photovoltaic (PV) dissemination

trajectories in Ghana, Kenya and Zimbabwe from the 1960s to 2007. Energy Policy 49, 410–

421. doi:10.1016/j.enpol.2012.06.042

BDEW, 2014. Energie-Info Erneuerbare Energien und das EEG : Zahlen , Fakten , Grafiken,

BDEW Bundesverband der Energie- und Wasserwirtschaft e.V. BDEW Bundesverband der

Energie- und Wasserwirtschaft e.V.

Beise, M., 2001. Lead Markets, Country-Specific Succes Factors of Global Diffuison of

Innovations. ZEW Economic Studies.

Beise, M., 2004. Lead markets: country-specific drivers of the global diffusion of innovations.

Res. Policy 33, 997–1018. doi:10.1016/j.respol.2004.03.003

Beise, M., 2005. Lead markets, innovation differentials and growth. Int. Econ. Econ. Policy 1,

305–328. doi:10.1007/s10368-004-0013-x

Beise, M., Cleff, T., 2004. Assessing the lead market potential of countries for innovation projects.

J. Int. Manag. 10, 453–477. doi:10.1016/j.intman.2004.08.003

Beise, M., Rennings, K., 2005. Lead markets and regulation: a framework for analyzing the

international diffusion of environmental innovations. Ecol. Econ. 52, 5–17.

doi:10.1016/j.ecolecon.2004.06.007

Ben-Ameur, H., Breton, M., L’Ecuyer, P., 2002. A dynamic programming procedure for pricing

American-style Asian options. Manage. Sci. 48, 625–643.

Berger, T., 2001. Agent-based spatial models applied to agriculture: a simulation tool for

technology diffusion, resource use changes and policy analysis. Agric. Econ. 25, 245–260.

doi:10.1111/j.1574-0862.2001.tb00205.x

Berthold, N., Kögel, D., Kullas, M., 2009. Die Bundesländer im Innovationswettbewerb.

Bertelsmann Stift.

Berthon, P., Pitt, L., Ewing, M., Carr, C., 2002. Potential research space in MIS: A framework for

envisioning and evaluating research replication, extension, and generation. Inf. Syst. Res. 13,

416–427.

Beurskens, L., Hekkenberg, M., Vethman, P., 2011. Renewable energy projections as published in

the national renewable energy action plans of the European member states. Eur. Environ.

Agency.

Binz, C., Truffer, B., Coenen, L., 2014. Why space matters in technological innovation systems—

Mapping global knowledge dynamics of membrane bioreactor technology. Res. Policy 43,

138–155. doi:10.1016/j.respol.2013.07.002

117

Blum, N.U., Sryantoro Wakeling, R., Schmidt, T.S., 2013. Rural electrification through village

grids—Assessing the cost competitiveness of isolated renewable energy technologies in

Indonesia. Renew. Sustain. Energy Rev. 22, 482–496. doi:10.1016/j.rser.2013.01.049

Bodas-Freitas, I., Dantas, E., Iizuka, M., 2010. The global institutional frameworks and the

diffusion of renewable energy technologies in the BRICS countries.

Bollinger, B., Gillingham, K., 2012. Peer Effects in the Diffusion of Solar Photovoltaic Panels.

Mark. Sci. 31, 900–912.

Branker, K., Pathak, M.J.M., Pearce, J.M., 2011. A review of solar photovoltaic levelized cost of

electricity. Renew. Sustain. Energy Rev. 15, 4470–4482. doi:10.1016/j.rser.2011.07.104

Breyer, C., Gerlach, A., 2013. Global overview on grid parity. Prog. photovoltaics Res. Appl. 21,

121–136. doi:10.1002/pip

Brockmann, D., Helbing, D., 2013. The hidden geometry of complex, network-driven contagion

phenomena. Science (80-. ). 342, 1337–42. doi:10.1126/science.1245200

Brown, L., 1975. The market and infrastructure context of adoption: a spatial perspective on the

diffusion of innovation. Econ. Geogr. 51.

Brudermann, T., Reinsberger, K., Orthofer, A., Kislinger, M., Posch, A., 2013. Photovoltaics in

agriculture: A case study on decision making of farmers. Energy Policy 61, 96–103.

doi:10.1016/j.enpol.2013.06.081

Burns, T., Stalker, G.M., 1961. The Management of Innovation. Tavistock, London.

Cagno, E., Ramirez-Portilla, A., Trianni, A., 2015. Linking energy efficiency and innovation

practices: Empirical evidence from the foundry sector. Energy Policy 83, 240–256.

doi:10.1016/j.enpol.2015.02.023

Cantono, S., Silverberg, G., 2009. A percolation model of eco-innovation diffusion: The

relationship between diffusion, learning economies and subsidies. Technol. Forecast. Soc.

Change 76, 487–496. doi:10.1016/j.techfore.2008.04.010

Cao-Alvira, J., 2010. Finite elements in the presence of occasionally binding constraints. Comput.

Econ. 355–370. doi:10.1007/s10614-010-9203-5

Cao-Alvira, J.J., 2011. Velocity Volatility Assessment of Monetary Shocks on Cash-in-Advance

Economies. Comput. Econ. 40, 293–311. doi:10.1007/s10614-011-9292-9

Card, N.A., 2011. Applied meta-analysis for social science research. guilford Press.

Carlsson, B., Jacobsson, S., Holmen, M., Annika, R., 2002. Innovation systems: analytical and

methodological issues. Res. Policy 31, 233–245.

Chesbrough, H., 2010. Business Model Innovation: Opportunities and Barriers. Long Range

Plann. 43, 354–363. doi:10.1016/j.lrp.2009.07.010

Chesbrough, H., Rosenbloom, R., 2002. The role of the business model in capturing value from

innovation: evidence from Xerox Corporation’s technology spin-off companies. Ind. Corp.

Chang. 11, 529–555.

118

Christensen, C.M., 1997. The Innovator’s Dilemma: When New Technologies Cause Great Firms

to Fail. Harvard Business School Press, Cambridge, MA.

Christensen, T.B., Wells, P., Cipcigan, L., 2012. Can innovative business models overcome

resistance to electric vehicles? Better Place and battery electric cars in Denmark. Energy

Policy 48, 498–505. doi:10.1016/j.enpol.2012.05.054

Cleff, T., Grimpe, C., Rammer, C., 2009. Demand-oriented innovation strategy in the European

energy production sector. Int. J. Energy Sect. Manag. 3, 108–130.

doi:10.1108/17506220910970542

Coenen, L., Benneworth, P., Truffer, B., 2012. Toward a spatial perspective on sustainability

transitions. Res. Policy 41, 968–979. doi:10.1016/j.respol.2012.02.014

Cook, D.J., Mulrow, C.D., Haynes, R.B., 1997. Systematic reviews: synthesis of best evidence for

clinical decisions. Ann Intern Med 126, 376–380.

Cook, D.J., Sackett, D.I., Spitzer, W.O., 1995. Methodological Guidelines for systematic reviews

of randomized control trials in health care from the postdam consultation on meta-analyisis.

J. Clin. Epidemology 48, 167–171.

Cooke, P., Morgan, K., 1998. The associational economy: firms, regions, and innovation, OUP

Catalogue. Oxford University Press. doi:10.1093/acprof

D’Agostino, A.L., Sovacool, B.K., Bambawale, M.J., 2011. And then what happened? A

retrospective appraisal of China’s Renewable Energy Development Project (REDP). Renew.

Energy 36, 3154–3165. doi:10.1016/j.renene.2011.03.017

Da Graça Carvalho, M., Bonifacio, M., Dechamps, P., 2011. Building a low carbon society.

Energy 36, 1842–1847. doi:10.1016/j.energy.2010.09.030

Daigneault, P.-M., Jacob, S., Ouimet, M., 2012. Using systematic review methods within a Ph.D.

dissertation in political science: challenges and lessons learned from practice. Int. J. Soc. Res.

Methodol. 1–17. doi:10.1080/13645579.2012.730704

Darmani, A., 2015. Renewable electricity generation and investors’ dynamic capabilities: A cross-

sector analysis of investment path choices. Util. Policy In press.

Darmani, A., Arvidsson, N., Hidalgo, A., 2015. Do the strategic decisions of multinational energy

companies differ in divergent market contexts? An exploratory study. Energy Res. Soc. Sci.

In press.

Darmani, A., Arvidsson, N., Hidalgo, A., Albors, J., 2014. What drives the development of

renewable energy technologies? Toward a typology for the systemic drivers. Renew. Sustain.

Energy Rev. 38, 834–847. doi:10.1016/j.rser.2014.07.023

Davis, F., 1989. Perceived usefulness, perceived ease of use, and user acceptance of information

technology. MIS Q. 13, 319–340.

Day, G., 1981. The product life cycle: analysis and applications issues. J. Mark. 45, 60–67.

119

De Araújo, M.S.M., de Freitas, M.A.V., 2008. Acceptance of renewable energy innovation in

Brazil—case study of wind energy. Renew. Sustain. Energy Rev. 12, 584–591.

doi:10.1016/j.rser.2006.03.017

Delre, S. a., Jager, W., Janssen, M. a., 2007. Diffusion dynamics in small-world networks with

heterogeneous consumers. Comput. Math. Organ. Theory 13, 185–202. doi:10.1007/s10588-

006-9007-2

Dewald, U., 2008. Solarinitiativen in Deutschland: Struktur, Aufgabenfelder und gegenwärtige

Herausforderungen. Lehrstuhl für Wirtschaftsgeographie, RWTH Aachen University.

Dewald, U., Truffer, B., 2011. Market Formation in Technological Innovation Systems—

Diffusion of Photovoltaic Applications in Germany. Ind. Innov. 18, 285–300.

doi:10.1080/13662716.2011.561028

Dewald, U., Truffer, B., 2012. The local sources of market formation: explaining regional growth

differentials in German photovoltaic markets. Eur. Plan. Stud. 37–41.

Diekmann, J., Groba, F., Vogel-Sperl, A., Püttner, A., Vohrer, P., Schmidt, J., 2012. Vergleich der

Bundesländer: Analyse der Erfolgsfaktoren für den Ausbau der Erneuerbaren Energien 2012-

Indikatoren und Ranking. Endbericht: Deutsches Institut für Wirtschaftsforschung.

Drücke, O., 2004. RegioSolar : Kurzstudie Solarinitiativen in Deutschland. Bundesverband

Solarindustrie e.V.

Drury, E., Miller, M., Macal, C.M., Graziano, D.J., Heimiller, D., Ozik, J., Perry IV, T.D., 2012.

The transformation of southern California’s residential photovoltaics market through third-

party ownership. Energy Policy 42, 681–690. doi:10.1016/j.enpol.2011.12.047

Durkheim, E., 1933. The division of Labor in Society. Macmillan, New York.

EC, 2007. Photovoltaic solar energy, Development and current research. European Commision

(EC), Belgium. doi:10.2768/38305

Eder, J.M., Mutsaerts, C.F., Sriwannawit, P., 2015. Mini-grids and renewable energy in rural

Africa: How diffusion theory explains adoption of electricity in Uganda. Energy Res. Soc.

Sci. 5, 45–54. doi:10.1016/j.erss.2014.12.014

EIO, 2010. Methodological report. Eco-Innovation Observatory (EIO).

EIO, 2011. The Eco-Innovation Challenge: Pathways to a resource-efficient Europe. Eco-

Innovation Observatory. Eco-Innovation Observatory (EIO), Funded by the European

Commission, DG Environment, Brussels.

Eisenhardt, K., 1989. Building theories from case study research. Acad. Manag. Rev. 14, 532–550.

EP, 2009. Directive 2009/28/EC of the European Parliment and the Council of the 23 April 2009

on the promotion of the use of energy from renewable sources and amending and

subswquently repealing Directives 2001/77/EC and 2003/30/EC. Off. J. Eur. Union 16–62.

EPIA, 2010. Photovoltaic energy, electricity from the sun, Energy. European Photovaltaic Industry

Association, ESTMAC project “Creating Markets for Renewable Energy Sources.”

120

EPIA, 2013. Global Market Outlook For Photovoltaics 2013-2017. European Photovoltaic

Industry Association (EPIA).

EPIA, 2014a. Global Market Outlook- For Photovoltaics 2014-2018. Eur. Photovolt. Ind. Assoc.

EPIA, 2014b. Market report 2013, European Photovoltaic Industry Association. European

Photovoltaic Industry Association (EPIA).

ETAP, 2010. Clean , Clever , Competitive Eco-innovation for a sustainable future, Newsletter

Issue 18. Environmental Technologies Action Plan (ETAP), European Commision, Brussels,

Belgium.

Faber, A., Frenken, K., 2009. Models in evolutionary economics and environmental policy:

Towards an evolutionary environmental economics. Technol. Forecast. Soc. Change 76,

462–470. doi:10.1016/j.techfore.2008.04.009

Fabrizio, K.R., Hawn, O., 2013. Enabling diffusion: How complementary inputs moderate the

response to environmental policy. Res. Policy 42, 1099–1111.

doi:10.1016/j.respol.2013.02.003

Fagerberg, J., Fosaas, M., Sapprasert, K., 2012. Innovation: Exploring the knowledge base. Res.

Policy 41, 1132–1153. doi:10.1016/j.respol.2012.03.008

Florida, R., Davison, D., 2001. Gaining from Green management: Environmental Management

Systems Inside and Outside the Factory. Calif. Manage. Rev.

Flyvbjerg, B., 2006. Five misunderstandings about case-study research. Qual. Inq. 420–434.

Fourt, L.A., Woodlock, J.W., 1960. Early Prediction of Market Success for New Grocery

Products. J. Mark. 25, 31–38.

Foxon, T., Pearson, P., 2008. Overcoming barriers to innovation and diffusion of cleaner

technologies: some features of a sustainable innovation policy regime. J. Clean. Prod. 16,

S148–S161. doi:10.1016/j.jclepro.2007.10.011

Frantzis, L., Graham, S., Katofsky, R., Sawyer, H., 2008. Photovoltaics business models. National

Renewable Energy Laboratory.

Frondel, M., Horbach, J., Rennings, K., 2008a. What triggers environmental management and

innovation? Empirical evidence for Germany. Ecol. Econ. 66, 153–160.

doi:10.1016/j.ecolecon.2007.08.016

Frondel, M., Ritter, N., Schmidt, C.M., 2008b. Germany’s solar cell promotion: Dark clouds on

the horizon. Energy Policy 36, 4198–4204. doi:10.1016/j.enpol.2008.07.026

Fthenakis, V.M., Kim, H.C., Alsema, E., 2008. Emissions from Photovoltaic Life Cycles

Emissions from Photovoltaic Life Cycles 42, 2168–2174. doi:10.1021/es071763q

Garcia, R., Jager, W., 2011. From the Special Issue Editors: Agent-Based Modeling of Innovation

Diffusion. J. Prod. Innov. Manag. 148–151.

121

Gaul, C., Carley, S., 2012. Solar set asides and renewable electricity certificates: Early lessons

from North Carolina’s experience with its renewable portfolio standard. Energy Policy 48,

460–469. doi:10.1016/j.enpol.2012.05.043

Geels, F.W., 2010. Ontologies, socio-technical transitions (to sustainability), and the multi-level

perspective. Res. Policy 39, 495–510. doi:10.1016/j.respol.2010.01.022

Geels, F.W., 2011. The multi-level perspective on sustainability transitions: Responses to seven

criticisms. Environ. Innov. Soc. Transitions 1, 24–40. doi:10.1016/j.eist.2011.02.002

Gooding, J., Edwards, H., Giesekam, J., Crook, R., 2013. Solar City Indicator: A methodology to

predict city level PV installed capacity by combining physical capacity and socio-economic

factors. Sol. Energy 95, 325–335. doi:10.1016/j.solener.2013.06.027

Graham, B., 1998. Liberalization, regional economic development and the geography of demand

for air transport in the European Union. J. Transp. Geogr. 6, 87–104. doi:10.1016/S0966-

6923(98)00003-9

Graziano, M., Gillingham, K., 2014. Spatial Patterns of Solar Photovoltaic System Adoption: The

Influence of Neighbors and the Built Environment ǂ. J. Econ. Geogr. 1–25.

doi:10.1093/jeg/lbu036

Griliches, Z., 1957. Hybrid corn: An exploration in the economics of technological change.

Econom. J. Econom. Soc. 25, 501–522.

Guidolin, M., Mortarino, C., 2010. Cross-country diffusion of photovoltaic systems: Modelling

choices and forecasts for national adoption patterns. Technol. Forecast. Soc. Change 77,

279–296. doi:10.1016/j.techfore.2009.07.003

Haas, R., Lettner, G., Auer, H., Duic, N., 2013. The looming revolution: How photovoltaics will

change electricity markets in Europe fundamentally. Energy 57, 38–43.

doi:10.1016/j.energy.2013.04.034

Hägerstrand, T., 1967. Innovation Diffusion as a Spatial Process. The University of Chicago Press

(Postscript and translation by Allan Pred).

Halila, F., 2007. The adoption and diffusion of environmental innovations, Doctoral Thesis.

Management. Lund University of Technology.

Hannah, S.T., Lester, P.B., 2009. A multilevel approach to building and leading learning

organizations. Leadersh. Q. 20, 34–48. doi:10.1016/j.leaqua.2008.11.003

Harvey, A., 1984. Time series forecasting based on the logistic curve. J. Oper. Res. Soc. 35, 641–

646.

Haynes, K., Mahajan, V., White, G., 1977. Innovation diffusion: A deterministic model of space-

time integration with physical analog. Socioecon. Plann. Sci. 11, 25–29.

Hervani, A. a., Helms, M.M., Sarkis, J., 2005. Performance measurement for green supply chain

management. Benchmarking An Int. J. 12, 330–353. doi:10.1108/14635770510609015

122

Higgins, A., McNamara, C., Foliente, G., 2014. Modelling future uptake of solar photo-voltaics

and water heaters under different government incentives. Technol. Forecast. Soc. Change 83,

142–155. doi:10.1016/j.techfore.2013.07.006

Ho, H., Wong, S., 2007. Housing allocation problem in a continuum transportation system.

Transportmetrica 3, 21–39.

Ho, H.W., Wong, S.C., Loo, B.P.Y., 2006. Combined distribution and assignment model for a

continuum traffic equilibrium problem with multiple user classes. Transp. Res. Part B

Methodol. 40, 633–650. doi:10.1016/j.trb.2005.09.003

Hockerts, K., Wüstenhagen, R., 2010. Greening Goliaths versus emerging Davids — Theorizing

about the role of incumbents and new entrants in sustainable entrepreneurship. J. Bus.

Ventur. 25, 481–492. doi:10.1016/j.jbusvent.2009.07.005

Hofer, C., 1975. Toward a contingency theory of business strategy. Acad. Manag. J. 18, 784–810.

Hoppmann, J., Peters, M., Schneider, M., Hoffmann, V.H., 2013. The two faces of market

support—How deployment policies affect technological exploration and exploitation in the

solar photovoltaic industry. Res. Policy 42, 989–1003. doi:10.1016/j.respol.2013.01.002

Horbach, J., Chen, Q., Rennings, K., Vögele, S., 2014. Do lead markets for clean coal technology

follow market demand? A case study for China, Germany, Japan and the US. Environ. Innov.

Soc. Transitions 10, 42–58. doi:10.1016/j.eist.2013.08.002

Hsu, D.D., O’Donoughue, P., Fthenakis, V., Heath, G. a., Kim, H.C., Sawyer, P., Choi, J.K.,

Turney, D.E., 2012. Life Cycle Greenhouse Gas Emissions of Crystalline Silicon

Photovoltaic Electricity Generation: Systematic Review and Harmonization. J. Ind. Ecol. 16.

doi:10.1111/j.1530-9290.2011.00439.x

Huber, J., 2008. Pioneer countries and the global diffusion of environmental innovations: Theses

from the viewpoint of ecological modernisation theory. Glob. Environ. Chang. 18, 360–367.

doi:10.1016/j.gloenvcha.2008.03.004

Huenteler, J., Schmidt, T.S., Kanie, N., 2012. Japan’s post-Fukushima challenge – implications

from the German experience on renewable energy policy. Energy Policy 45, 6–11.

doi:10.1016/j.enpol.2012.02.041

Huijben, J.C.C.M., Verbong, G.P.J., 2013. Breakthrough without subsidies? PV business model

experiments in the Netherlands. Energy Policy 56, 362–370. doi:10.1016/j.enpol.2012.12.073

Hull, C., Rothenberg, S., 2008. Firm performance: The interactions of corporate social

performance with innovation and industry differentiation. Strateg. Manag. J. 29, 781–789.

doi:10.1002/smj

IEA, 2014. Technology Roadmap, Solar Photovoltaic Energy. Int. Energy Agency.

IMS Research, 2013. PV Storage Market Set to Explode to $19 Billion in 2017; Germany leads

Again (Press Release, 25 April 2013).

Islam, T., Meade, N., 2013. The impact of attribute preferences on adoption timing: The case of

photo-voltaic (PV) solar cells for household electricity generation. Energy Policy 55, 521–

530. doi:10.1016/j.enpol.2012.12.041

123

Jacob, K., Beise, M., Blazejczak, J., Edler, D., Haum, R., Jänicke, M., Löw, T., Petschow, U.,

Rennings, K., 2005. Lead Markets for VOC-Reduced Paints, in: Wolfgang, F. (Ed.), Lead

Markets for Environmental Innovations. Physica-Verlag Heidelberg New York.

Jacobsson, S., Bergek, A., 2004. Transforming the energy sector: the evolution of technological

systems in renewable energy technology. Ind. Corp. Chang. 13, 815–849.

doi:10.1093/icc/dth032

Jacobsson, S., Johnson, A., 2000. The diffusion of renewable energy technology: an analytical

framework and key issues for research. Energy Policy 28, 625–640.

Jacobsson, S., Lauber, V., 2006. The politics and policy of energy system transformation-

explaining the German diffusion of renewable energy technology. Energy Policy 34, 256–

276. doi:10.1016/j.enpol.2004.08.029

Jaffe, A., Newell, R., Stavins, R., 2002. Environmental policy and technological change. Environ.

Resour. … 41–69.

Jaffe, A., Palmer, K., 1997. Environmental regulation and innovation: a panel data study. Rev.

Econ. Stat. 79, 610–619.

Jäger, W., 2006. Stimulating the diffusion of photovoltaic systems: A behavioural perspective.

Energy Policy 34, 1935–1943. doi:10.1016/j.enpol.2004.12.022

Jänicke, M., 2008. Ecological modernisation: new perspectives. J. Clean. Prod. 16, 557–565.

doi:10.1016/j.jclepro.2007.02.011

Jeong, G., 2013. Assessment of government support for the household adoption of micro-

generation systems in Korea. Energy Policy 62, 573–581. doi:10.1016/j.enpol.2013.07.015

Jiang, Y., Wong, S.C., Ho, H.W., Zhang, P., Liu, R., Sumalee, A., 2011. A dynamic traffic

assignment model for a continuum transportation system. Transp. Res. Part B Methodol. 45,

343–363. doi:10.1016/j.trb.2010.07.003

Johanson, J., 1978. The Mechanism of Internationalisation.

Julien, P.-A., 2007. A Theory of Local Entrepreneurship in the Knowledge Economy. Edwaed

Elgar Publishing Limited.

Kalish, S., Mahajan, V., Muller, E., 1995. Waterfall and sprinkler new-product strategies in

competitive global markets. Int. J. Res. Mark. 12, 105–119. doi:10.1016/0167-

8116(94)00008-C

Kannan, R., Leong, K.C., Osman, R., Ho, H.K., Tso, C.P., 2006. Life cycle assessment study of

solar PV systems: An example of a 2.7 kWp distributed solar PV system in Singapore. Sol.

Energy 80, 555–563. doi:10.1016/j.solener.2005.04.008

Karakaya, E., Hidalgo, A., Nuur, C., 2014a. Diffusion of eco-Innovations: A review. Renew.

Sustain. Energy Rev. 33, 392–399.

Karakaya, E., Hidalgo, A., Nuur, C., 2015. Motivators for adoption of photovoltaic systems at grid

parity: A case study from Southern Germany. Renew. Sustain. Energy Rev. 43, 1090–1098.

doi:10.1016/j.rser.2014.11.077

124

Karakaya, E., Nuur, C., Breitschopf, B., Hidalgo, A., 2014b. Spatial Dimension of Lead Markets:

Evidences from Diffusion of Photovoltaic Systems in Germany, in: DRUID Society

Conference 2014, CBS, Copenhagen, June 16-18.

Karakaya, E., Nuur, C., Hidalgo, A., 2016. Business model challenge: Lessons from a local solar

company. Renew. Energy 85, 1026–1035. doi:10.1016/j.renene.2015.07.069

Karakaya, E., Sriwannawit, P., 2015. Barriers to the adoption of photovoltaic systems: The state of

the art. Renew. Sustain. Energy Rev. 49, 60–66.

Karteris, M., Papadopoulos, A.M., 2013. Legislative framework for photovoltaics in Greece: A

review of the sector’s development. Energy Policy 55, 296–304.

doi:10.1016/j.enpol.2012.12.001

Kazmerski, L.L., 2006. Solar photovoltaics R&D at the tipping point: A 2005 technology

overview. J. Electron Spectros. Relat. Phenomena 150, 105–135.

doi:10.1016/j.elspec.2005.09.004

Kiesling, E., Günther, M., Stummer, C., Wakolbinger, L.M., 2011. Agent-based simulation of

innovation diffusion: a review. Cent. Eur. J. Oper. Res. 183–230. doi:10.1007/s10100-011-

0210-y

Kley, F., Lerch, C., Dallinger, D., 2011. New business models for electric cars—A holistic

approach. Energy Policy 39, 3392–3403. doi:10.1016/j.enpol.2011.03.036

Kline, S.J., Rosenberg, N., 1986. An overview of innovation. Posit. sum Strateg. Harnessing

Technol. Econ. growth 275, 305.

Kobos, P.H., Erickson, J.D., Drennen, T.E., 2006. Technological learning and renewable energy

costs: implications for US renewable energy policy. Energy Policy 34, 1645–1658.

doi:10.1016/j.enpol.2004.12.008

Köhler, J., Walz, R., Marscheder-Weidemann, F., Thedieck, B., 2014. Lead markets in 2nd

generation biofuels for aviation: A comparison of Germany, Brazil and the USA. Environ.

Innov. Soc. Transitions 10, 59–76. doi:10.1016/j.eist.2013.10.003

Koinegg, J., Brudermann, T., Posch, A., Mrotzek, M., 2013. “It would be a shame if we did not

take advantage of the spirit of the times …” An analysis of prospects and barriers of building

integrated photovoltaics. Ecol. Perspect. Sci. Soc. 22, 39–45.

Komatsu, S., Kaneko, S., Shrestha, R.M., Ghosh, P.P., 2011. Nonincome factors behind the

purchase decisions of solar home systems in rural Bangladesh. Energy Sustain. Dev. 15,

284–292. doi:10.1016/j.esd.2011.03.003

Kost, C., Mayer, J.N., Thomsen, J., Hartmann, N., Senkpiel, C., Philipps, S., Nold, S., Lude, S.,

Saad, N., Schlegl, T., 2013. Levelized Cost of Electricity, Renewable Energy Technologies.

Fraunhofer ISE.

Kost, C., Schlegl, T., 2010. Studie Stromgestehungskosten Erneuerbare Energien, Fraunhofer, ISE,

Fraunhofer ISE.

Kost, C., Schlegl, T., Thomsen, J., Nold, S., Mayer, J., 2012. Studie Stromgestehungskosten

Erneuerbare Energien, Fraunhofer, ISE, Fraunhofer ISE.

125

Kwan, C.L., 2012. Influence of local environmental, social, economic and political variables on

the spatial distribution of residential solar PV arrays across the United States. Energy Policy

47, 332–344. doi:10.1016/j.enpol.2012.04.074

Laird, F.N., Stefes, C., 2009. The diverging paths of German and United States policies for

renewable energy: Sources of difference. Energy Policy 37, 2619–2629.

doi:10.1016/j.enpol.2009.02.027

Lay, J., Ondraczek, J., Stoever, J., 2013. Renewables in the energy transition: Evidence on solar

home systems and lighting fuel choice in Kenya. Energy Econ. 40, 350–359.

doi:10.1016/j.eneco.2013.07.024

Lazard, 2013. Lazard’s levelized cost of energy analysis - version 7.0.

Leepa, C., Unfried, M., 2013. Effects of a cut-off in feed-in tariffs on photovoltaic capacity:

Evidence fromGermany. Energy Policy 56, 536–542. doi:10.1016/j.enpol.2013.01.018

Lettner, G., Auer, H., 2012a. Realistic roadmap to PV grid parity for all target countries. Vienna

Univ. Technol. - PV Parit. Proj.

Lettner, G., Auer, H., 2012b. Realistic roadmap to PV grid parity in the commercial sector for all

target countries.

Lewis, R.W., Nithiarasu, P., Seetharamu, K.N., 2004. Fundamentals of the Finite Element Method

for Heat and Fluid Flow. John Wiley & Sons.

Lincoln, C.Y.S., Guba, E.G., 2009. The Only Generalization is: There is no Generalization, in:

Gomm, R., Hammersley, M., Foster, P. (Eds.), Case Study Method. Sage Publications, pp.

27–45.

Loock, M., 2012. Going beyond best technology and lowest price: on renewable energy investors’

preference for service-driven business models. Energy Policy 40, 21–27.

doi:10.1016/j.enpol.2010.06.059

Lund, P.D., 2011. Boosting new renewable technologies towards grid parity – Economic and

policy aspects. Renew. Energy 36, 2776–2784. doi:10.1016/j.renene.2011.04.025

Luque, A., 2001. Photovoltaic market and costs forecast based on a demand elasticity model. Prog.

Photovoltaics Res. Appl. 9, 303–312. doi:10.1002/pip.371

MacVaugh, J., Schiavone, F., 2010. Limits to the diffusion of innovation: A literature review and

integrative model. Eur. J. Innov. Manag. doi:10.1108/14601061011040258

Madlener, R., 2007. Innovation diffusion, public policy, and local initiative: The case of wood-

fuelled district heating systems in Austria. Energy Policy 35, 1992–2008.

doi:10.1016/j.enpol.2006.06.010

Magretta, J., 2002. Why Business Models Matter. Harv. Bus. Rev. 80, 86–92.

Mahajan, V., Muller, E., Wind, Y., 2000. New-Product Diffusion Models, in: International Series

in Quantitative Marketing. Kluwer Academic Publishers.

126

Mahajan, V., Peterson, R. a., 1979. Integrating time and space in technological substitution

models. Technol. Forecast. Soc. Change 14, 231–241. doi:10.1016/0040-1625(79)90079-9

Mainali, B., Silveira, S., 2011. Financing off-grid rural electrification: Country case Nepal. Energy

36, 2194–2201. doi:10.1016/j.energy.2010.07.004

Malerba, F., 2002. Sectoral systems of innovation and production. Res. Policy 31, 247–264.

Mallett, A., 2007. Social acceptance of renewable energy innovations: The role of technology

cooperation in urban Mexico. Energy Policy 35, 2790–2798.

doi:10.1016/j.enpol.2006.12.008

Mani, A., Rahwan, I., Pentland, A., 2013. Inducing peer pressure to promote cooperation. Nature,

Sci. Reports 3, 1735. doi:10.1038/srep01735

Mansfield, E., 1961. Technical change and the rate of imitation. Econom. J. Econom. Soc. 29,

741–766.

Markard, J., Raven, R., Truffer, B., 2012. Sustainability transitions: An emerging field of research

and its prospects. Res. Policy 41, 955–967. doi:10.1016/j.respol.2012.02.013

Markides, C., 2013. Business model innovation: What can the ambidexterity literature teach us?

Acad. Manag. Perspect. 27, 313–323.

Marx, K., 1954. Capital: a critique of political economy. Lawrence & Wishart, London.

Masini, A., Frankl, P., 2002. Forecasting the diffusion of photovoltaic systems in southern

Europe : A learning curve approach. Technol. Forecast. Soc. Change 70, 39–65.

McEachern, M., Hanson, S., 2008. Socio-geographic perception in the diffusion of innovation:

Solar energy technology in Sri Lanka. Energy Policy 36, 2578–2590.

doi:10.1016/j.enpol.2008.03.020

Meade, N., Islam, T., 1998. Technological forecasting—Model selection, model stability, and

combining models. Manage. Sci. 44, 1115–1130.

Meade, N., Islam, T., 2006. Modelling and forecasting the diffusion of innovation – A 25-year

review. Int. J. Forecast. 22, 519–545. doi:10.1016/j.ijforecast.2006.01.005

Mesak, H., Coleman, R., 1992. Modeling the effect of subsidized pricing policy on new product

diffusion. Omega 20, 303–312. doi:10.1016/0305-0483(92)90035-6

Michelacci, C., Silva, O., 2007. Why so many local entrepreneurs? Rev. Econ. Stat. 89, 615–633.

Morgan, K., 2004. The exaggerated death of geography: learning, proximity and territorial

innovation systems. J. Econ. Geogr. 4, 3–21. doi:10.1093/jeg/4.1.3

Morrill, R., 1970. The shape of diffusion in space and time. Econ. Geogr. 46, 259–268.

Morris, M., Schindehutte, M., Allen, J., 2005. The entrepreneur’s business model: toward a unified

perspective. J. Bus. Res. 58, 726–735. doi:10.1016/j.jbusres.2003.11.001

127

Movilla, S., Miguel, L.J., Blázquez, L.F., 2013. A system dynamics approach for the photovoltaic

energy market in Spain. Energy Policy 60, 142–154. doi:10.1016/j.enpol.2013.04.072

Müggenburg, H., Tillmans, A., Schweizer-Ries, P., Raabe, T., Adelmann, P., 2012. Social

acceptance of PicoPV systems as a means of rural electrification — A socio-technical case

study in Ethiopia. Energy Sustain. Dev. 16, 90–97. doi:10.1016/j.esd.2011.10.001

Müller, S., Rode, J., 2013. The adoption of photovoltaic systems in Wiesbaden, Germany. Econ.

Innov. New Technol. 22, 519–535. doi:10.1080/10438599.2013.804333

Munoz, L. a H., Huijben, J.C.C.M., Verhees, B., Verbong, G.P.J., 2014. The power of grid parity:

A discursive approach. Technol. Forecast. Soc. Change 87, 179–190.

doi:10.1016/j.techfore.2013.12.012

Murray, C.E., 2009. Diffusion of Innovation Theory : A Bridge for the Research - Practice Gap in

Counseling. J. Couns. Dev. 87, 108–116.

Nair, S., Paulose, H., 2014. Emergence of green business models: The case of algae biofuel for

aviation. Energy Policy 65, 175–184. doi:10.1016/j.enpol.2013.10.034

Nelson, R., 1959. The simple economics of basic scientific research. J. Polit. Econ. 67, 297–306.

Nelson, R., Winter, S., 1982. An evolutionary Theory of Economic Change. Harvard University

Press.

Nemet, G.F., 2006. Beyond the learning curve: factors influencing cost reductions in

photovoltaics. Energy Policy 34, 3218–3232. doi:10.1016/j.enpol.2005.06.020

NREL, 2012. Best research-cell efficiences, National Reneable Energy Laboratory. National

Renewable Energy Laboratory (NREL).

NREL, 2014. Best Research-Cell Efficiences [WWW Document]. Natl. Reneable Energy Lab.

URL http://www.nrel.gov/ncpv/images/efficiency_chart.jpg (accessed 6.10.14).

Okkonen, L., Suhonen, N., 2010. Business models of heat entrepreneurship in Finland. Energy

Policy 38, 3443–3452. doi:10.1016/j.enpol.2010.02.018

Oltra, V., Saint Jean, M., 2009. Sectoral systems of environmental innovation: An application to

the French automotive industry. Technol. Forecast. Soc. Change 76, 567–583.

doi:10.1016/j.techfore.2008.03.025

Ondraczek, J., 2013. The sun rises in the east (of Africa): A comparison of the development and

status of solar energy markets in Kenya and Tanzania. Energy Policy 56, 407–417.

doi:10.1016/j.enpol.2013.01.007

Ormrod, R., 1990. Local context and innovation diffusion in a well-connected world. Econ. Geogr.

66, 109–122.

Ottman, J., Stafford, E., Hartman, C., 2006. Avoiding green marketing myopia: ways to improve

consumer appeal for environmentally preferable products. Environment 48, 24–36.

Ozaki, R., 2011. Adopting sustainable innovation: what makes consumers sign up to green

electricity? Bus. Strateg. Environ. 20, 1–17.

128

Painuly, J.P., 2001. Barriers to renewable energy penetration: A framework for analysis. Renew.

Energy 24, 73–89. doi:10.1016/S0960-1481(00)00186-5

Palit, D., 2013. Solar energy programs for rural electrification: Experiences and lessons from

South Asia. Energy Sustain. Dev. 17, 270–279. doi:10.1016/j.esd.2013.01.002

Pansera, M., 2012. Renewable energy for rural areas of Bolivia. Renew. Sustain. Energy Rev. 16,

6694–6704. doi:10.1016/j.rser.2012.08.015

Pantaleo, A., Candelise, C., Bauen, A., Shah, N., 2014. ESCO business models for biomass

heating and CHP: Profitability of ESCO operations in Italy and key factors assessment.

Renew. Sustain. Energy Rev. 30, 237–253. doi:10.1016/j.rser.2013.10.001

Parida, B., Iniyan, S., Goic, R., 2011. A review of solar photovoltaic technologies. Renew.

Sustain. Energy Rev. 15, 1625–1636. doi:10.1016/j.rser.2010.11.032

Parsons, T., 1938. The role of theory in social research. Am. Sociol. Rev. 3, 13–20.

Parunak, H.V.D., Savit, R., 1998. Agent-based modeling vs. equation-based modeling: A case

study and users’ guide, in: Proceedings of Multi-Agent Systems and Agent-Based Simulation

(MABS’98), Springer, LNAI 1534. pp. 10–25.

Pearson, D., Fuller, G., Chapin, C.;, 1954. A New Silicon p-n Junction Photocell for Converting

Solar Radiation into Electrical Power. J. Appl. Phys. 25, 676–677.

Penrose, E., 2009. The theory of the growth of the firm, 4th Editio. ed. Oxford University Press.

Pentland, A., 2012. The new science of building great teams. Harv. Bus. Rev.

Pentland, A., 2013. Beyond the Echo Chamber. Harv. Bus. Rev.

Pentland, A., 2014. Social Physics, How Good Ideas Spread- The Lessons from a New Science.

The Penguin Press, New York.

Peres, R., Muller, E., Mahajan, V., 2010. Innovation diffusion and new product growth models: A

critical review and research directions. Int. J. Res. Mark. 27, 91–106.

doi:10.1016/j.ijresmar.2009.12.012

Pérez, D., Cervantes, V., Báez, M.J., Domínguez, F.T., 2013. PV Grid Parity Monitor Residential

Sector, Spain eclareon. Spain Eclareon.

Peter, R., Ramaseshan, B., Nayar, C.., 2002. Conceptual model for marketing solar based

technology to developing countries. Renew. Energy 25, 511–524. doi:10.1016/S0960-

1481(01)00080-5

Petrova-Koch, V., Hezel, R., Goetzberger, A., 2009. High-Efficient Low-Cost Photovoltaics -

Recent Developments, Springer-Verlag Berlin Heidelberg.

Pode, R., 2013. Financing LED solar home systems in developing countries. Renew. Sustain.

Energy Rev. 25, 596–629. doi:10.1016/j.rser.2013.04.004

Popp, D., Hascic, I., Medhi, N., 2011. Technology and the diffusion of renewable energy. Energy

Econ. 33, 648–662. doi:10.1016/j.eneco.2010.08.007

129

Porter, M., 1990. The competitive advantage of nations. Free Press, New York.

Porter, M.E., 1990. The Competitive Advantage of Nations. Free Press, New York.

Porter, M.E., 2000. Location, Competition, and Economic Development: Local Clusters in a

Global Economy. Econ. Dev. Q. 14, 15–34. doi:10.1177/089124240001400105

Prahalad, C.K., 2010. The fortune at the bottom of the pyramid: Eradicating poverty through

profits (revised and ppdated 5th anniversary edition), ptgmedia.pearsoncmg.com. Pearson

Education, Inc, Upper Saddle River, New Jersey.

Pujari, D., 2006. Eco-innovation and new product development: understanding the influences on

market performance. Technovation 26, 76–85. doi:10.1016/j.technovation.2004.07.006

PvParity, 2013. PV Parity Project, How to Support Different Photovoltaic Applications in the

Achievement of Competitiveness and Beyond. Final Rep.

Quitzow, R., Walz, R., Köhler, J., Rennings, K., 2014. The concept of “lead markets” revisited:

Contribution to environmental innovation theory. Environ. Innov. Soc. Transitions 10, 4–19.

doi:10.1016/j.eist.2013.11.002

Rahmandad, H., Sterman, J., 2008. Heterogeneity and Network Structure in the Dynamics of

Diffusion: Comparing Agent-Based and Differential Equation Models. Manage. Sci. 54,

998–1014. doi:10.1287/mnsc.1070.0787

Rai, V., Robinson, S.A., 2013. Effective information channels for reducing costs of

environmentally- friendly technologies: evidence from residential PV markets. Environ. Res.

Lett. 8, 014044. doi:10.1088/1748-9326/8/1/014044

Rao, K.U., Kishore, V.V.N., 2010. A review of technology diffusion models with special reference

to renewable energy technologies. Renew. Sustain. Energy Rev. 14, 1070–1078.

doi:10.1016/j.rser.2009.11.007

Reardon, C., 2009. The diffusion of sustainable practices within the Australian housing industry :

implications for future change management.

Rebane, K.L., Barham, B.L., 2011. Knowledge and adoption of solar home systems in rural

Nicaragua. Energy Policy 39, 3064–3075. doi:10.1016/j.enpol.2011.02.005

Reichelstein, S., Yorston, M., 2013. The prospects for cost competitive solar PV power. Energy

Policy 55, 117–127. doi:10.1016/j.enpol.2012.11.003

Ren21, 2014. Renewables 2013: Global Status Report, Renewable Energy Policy Network for the

21st Century. Renewable Energy Policy Network for the 21st Centruy (Ren21).

Rennings, K., 2000. Redefining innovation—eco-innovation research and the contribution from

ecological economics. Ecol. Econ. 32, 319–332.

Richter, M., 2013a. Business model innovation for sustainable energy: German utilities and

renewable energy. Energy Policy 62, 1226–1237. doi:10.1016/j.enpol.2013.05.038

Richter, M., 2013b. German utilities and distributed PV: How to overcome barriers to business

model innovation. Renew. Energy 55, 456–466. doi:10.1016/j.renene.2012.12.052

130

Rink, D.R., Swan, J.E., 1979. Product life cycle research: A literature review. J. Bus. Res. 7, 219–

242. doi:10.1016/0148-2963(79)90030-4

Rode, J., Weber, A., 2012. Does Localized Imitation Drive Technology Adoption? A Case Study

on Solar Cells in Germany.

Rogers, E., Medina, U., Rivera, M., Wiley, C., 2005. Complex adaptive systems and the diffusion

of innovations. Innov. J. Public Sect. Innov. J. 10.

Rogers, E.M., 1962. Diffusion of Innovations. The Free Press, New York.

Rogers, E.M., 2003. Diffusion of innovations, 5th ed. The Free Press, New York.

Sá, C., Florax, R.J.G.M., Rietveld, P., 2004. Determinants of the Regional Demand for Higher

Education in The Netherlands: A Gravity Model Approach. Reg. Stud. 38, 375–392.

doi:10.1080/03434002000213905

SABL, 2013. Statistische Ämter des Bundes und der Länder [WWW Document]. URL

www.regionalstatistik.de

Sarzynski, A., Larrieu, J., Shrimali, G., 2012. The impact of state financial incentives on market

deployment of solar technology. Energy Policy 46, 550–557.

doi:10.1016/j.enpol.2012.04.032

SB, 2007. Ein Zentrum fürs Leben. Sol. Blvd.

Schiederig, T., Tietze, F., Herstatt, C., 2012. Green innovation in technology and innovation

management - an exploratory literature review. R&D Manag. 42, 180–192.

doi:10.1111/j.1467-9310.2011.00672.x

Schmookler, J., 1966. Invention and Economic Growth. Harvard University Press, Massachusetts.

Schoolgen, 2012. Student Fact Sheet: A Short History of Photovoltaic ( PV ) Cells, downloaded.

ed, World. Schoolgen Project, Genesis Energy, New Zealand.

Schumpeter, J.A., 1939. Business Cycles, A Theoretical, Historical and Statistical Analysis of the

Capitalist Process, Édition nu. ed. McGraw-Hill Book Company, Québec, Canada.

doi:10.1016/j.socscimed.2006.11.007

Schwarz, N., Ernst, A., 2009. Agent-based modeling of the diffusion of environmental innovations

— An empirical approach. Technol. Forecast. Soc. Change 76, 497–511.

doi:10.1016/j.techfore.2008.03.024

SEIA, 2010. The History of Solar Energy A Look at Uses of Solar Energy from BC to ITC. Solar

Energy Industries Association (SEIA).

Sharif, M., Islam, M., 1980. The Weibull distribution as a general model for forecasting

technological change. Technol. Forecast. Soc. Change 256, 247–256.

Shinohara, K., Okuda, H., 2010. Dynamic Innovation Diffusion Modelling. Comput. Econ. 35,

51–62. doi:10.1007/s10614-009-9191-5

131

Shrivastava, P., 1995. Environmental technologies and competitive advantage. Strateg. Manag. J.

16, 183–200.

Siggelkow, N., 2007. Persuasion With Case Studies. Acad. Manag. J. 50, 20–24.

doi:10.5465/AMJ.2007.24160882

SLBW, 2013. Statistisches Landesamt Baden-Württemberg [WWW Document]. URL

www.statistik.baden-wuerttemberg.de/?

Smith, A., Kern, F., Raven, R., Verhees, B., 2014. Spaces for sustainable innovation: Solar

photovoltaic electricity in the UK. Technol. Forecast. Soc. Change 81, 115–130.

doi:10.1016/j.techfore.2013.02.001

Sosna, M., Trevinyo-Rodríguez, R.N., Velamuri, S.R., 2010. Business Model Innovation through

Trial-and-Error Learning. Long Range Plann. 43, 383–407. doi:10.1016/j.lrp.2010.02.003

Sovacool, B.K., 2014a. What are we doing here? Analyzing fifteen years of energy scholarship

and proposing a social science research agenda. Energy Res. Soc. Sci. 1, 1–29.

doi:10.1016/j.erss.2014.02.003

Sovacool, B.K., 2014b. Diversity: Energy studies need social science. Nature 511, 529–30.

doi:10.1038/511529a

Spaargaren, G., 2003. Sustainable consumption: a theoretical and environmental policy

perspective. Soc. &Natural Resour. 687–701. doi:10.1080/08941920390217429

Spence, A., 1981. The learning curve and competition. Bell J. Econ. 12, 49–70.

Spertino, F., Leo, P. Di, Cocina, V., 2014. Which are the constraints to the photovoltaic grid-parity

in the main European markets? Sol. Energy 105, 390–400. doi:10.1016/j.solener.2014.03.021

Sriwannawit, P., 2012. Managing the diffusion of electrification in Bangladesh, in: International

Conference for the International Association of Management of Technology (IAMOT)

Taiwan.

Sriwannawit, P., Laestadius, S., 2013. Diffusion of photovoltaic systems for rural electrification in

Thailand. Int. J. Energy Environ. 4, 49–58.

Sriwannawit, P., Sandström, U., 2015. Large-scale bibliometric review of diffusion research.

Scientometrics 102, 1615–1645. doi:10.1007/s11192-014-1448-7

Stake, R., 2000. The case study method in social inquiry, in: Gomm, R., Hammer, M., Foster, P.

(Eds.), Case Study Method. London: Sage [originally published in Educational Researcher (7

February 1978): 5–8], pp. 19–26.

Strang, D., Meyer, J.W., 1993. Institutional conditions for diffusion. Theory Soc. 22, 487–511.

doi:10.1007/BF00993595

Sühlsen, K., Hisschemöller, M., 2014. Lobbying the “Energiewende”. Assessing the effectiveness

of strategies to promote the renewable energy business in Germany. Energy Policy 69, 316–

325. doi:10.1016/j.enpol.2014.02.018

132

Surinach, J., Autant-Bernard, C., Manca, F., 2009. The diffusion/adoption of innovation in the

internal market. Eur. Econ. Econ. Pap. 384 1, 1829–1841. doi:10.2765/40088

SWW, 2004. Flexibilität hört auf dem Dach nicth auf. Sonne, Wind wärme 1–4.

Szabó, S., Bódis, K., Huld, T., Moner-Girona, M., 2011. Energy solutions in rural Africa: mapping

electrification costs of distributed solar and diesel generation versus grid extension. Environ.

Res. Lett. 6, 034002. doi:10.1088/1748-9326/6/3/034002

Teece, D., 1980. The diffusion of an administrative innovation. Manage. Sci. 26, 464–470.

The Economist, 2014. Hail , the Swabian housewife Feb 1st.

The Economist, 2015. Special Report: Energy and technology. Econ. 17 Jan.

Thiam, D.R., 2011. An energy pricing scheme for the diffusion of decentralized renewable

technology investment in developing countries. Energy Policy 39, 4284–4297.

doi:10.1016/j.enpol.2011.04.046

Tiwari, R., Herstatt, C., 2012. Assessing India’s lead market potential for cost-effective

innovations. J. Indian Bus. Res. 4, 97–115. doi:10.1108/17554191211228029

Tomas, M., Yalamanchili, K., 2001. An application of finite elements to option pricing. J. Futur.

Mark. 21, 19–42.

Tongur, S., Engwall, M., 2014. The business model dilemma of technology shifts. Technovation

1–11. doi:10.1016/j.technovation.2014.02.006

Tranfield, D., Denyer, D., Smart, P., 2003. Towards a methodology for developing evidence-

informed management knowledge by means of systematic review *. Br. J. Manag. 14, 207–

222. doi:10.1111/1467-8551.00375

Trimi, S., Berbegal-Mirabent, J., 2012. Business model innovation in entrepreneurship. Int.

Entrep. Manag. J. 8, 449–465. doi:10.1007/s11365-012-0234-3

Trogdon, J.G., Nonnemaker, J., Pais, J., 2008. Peer effects in adolescent overweight. J. Health

Econ. 27, 1388–99. doi:10.1016/j.jhealeco.2008.05.003

Truffer, B., Coenen, L., 2012. Environmental innovation and sustainability transitions in regional

studies. Reg. Stud. 1–21.

Tsoutsos, T.D., Stamboulis, Y. a., 2005. The sustainable diffusion of renewable energy

technologies as an example of an innovation-focused policy. Technovation 25, 753–761.

doi:10.1016/j.technovation.2003.12.003

Twining, W., 2005. Social science and diffusion of law. J. Law Soc. 32, 203–240.

Van den Bergh, J., 2013a. Environmental and climate innovation: Limitations, policies and prices.

Technol. Forecast. Soc. Change 80, 11–23. doi:10.1016/j.techfore.2012.08.004

Van den Bergh, J., 2013b. Policies to enhance economic feasibility of a sustainable energy

transition. Proc. Natl. Acad. Sci. U. S. A. 110, 2436–2437. doi:10.1073/pnas.1221894110

133

Vasseur, V., Kamp, L.M., Negro, S.O., 2013. A comparative analysis of Photovoltaic

Technological Innovation Systems including international dimensions: The cases of Japan

and The Netherlands. J. Clean. Prod. 48, 200–210. doi:10.1016/j.jclepro.2013.01.017

Venkatesh, V., 2000. Determinants of perceived ease of use: Integrating control, intrinsic

motivation, and emotion into the technology acceptance model. Inf. Syst. Res. 1997, 342–

365.

Waarts, E., van Everdingen, Y., 2005. The influence of national culture on the adoption status of

innovations: An empirical study of firms across Europe. Eur. Manag. J. 23, 601–610.

doi:10.1016/j.emj.2005.10.007

Wagner, M., 2008. Empirical influence of environmental management on innovation: Evidence

from Europe. Ecol. Econ. 66, 392–402. doi:10.1016/j.ecolecon.2007.10.001

Walker, J., 1969. The diffusion of innovations among the American states. Am. Polit. Sci. Rev. 63,

880–899.

Walz, J., Fulton, R., Cyrus, N., 1968. Accuracy and convergence of finite element approximations.

Walz, R., Köhler, J., 2014. Using lead market factors to assess the potential for a sustainability

transition. Environ. Innov. Soc. Transitions 10, 20–41. doi:10.1016/j.eist.2013.12.004

Watanabe, C., Asgari, B., 2004. Impacts of functionality development on dynamism between

learning and diffusion of technology. Technovation 24, 651–664. doi:10.1016/S0166-

4972(02)00126-8

WB, 2014. World Bank categorization of countries by income [WWW Document]. World Bank.

URL http://data.worldbank.org/about/country-classifications/country-and-lending-groups

(accessed 5.6.14).

Weber, M., 1922. Economy and society: An outline of interpretive sociology.

Wejnert, B., 2002. Integrating Models of Diffusion of Innovations : A Conceptual Framework.

Annu. Rev. Sociol. 28, 297–326. doi:10.1146/annurev.soc.28.110601.141051

Wirth, H., 2013. Aktuelle Fakten zur Photovoltaik in Deutschland. Fraunhofer ISE.

Wirth, H., 2014. Recent facts about photovoltaics in Germany. Fraunhofer ISE.

Wirth, S., 2014. Communities matter: Institutional preconditions for community renewable

energy. Energy Policy 70, 236–246. doi:10.1016/j.enpol.2014.03.021

Wong, S., 1998. Multi-commodity traffic assignment by continuum approximation of network

flow with variable demand. Transp. Res. Part B Methodol. 32, 567–581.

Wong, S.C., Du, Y.C., Sun, L.J., Loo, B.P.Y., 2006. Sensitivity analysis for a continuum traffic

equilibrium problem. Ann. Reg. Sci. 40, 493–514. doi:10.1007/s00168-006-0071-9

Wong, S.C., Sun, S.H., 2001. A combined distribution and assignment model for continuous

facility location problem. Ann. Reg. Sci. 35, 267–281. doi:10.1007/s001680100042

134

Wong, S.C., Zhou, C., Lo, H.K., Yang, H., 2004. Improved Solution Algorithm for

Multicommodity Continuous Distribution and Assignment Model. J. Urban Plan. Dev. 130,

14–23. doi:10.1061/(ASCE)0733-9488(2004)130:1(14)

Wüstenhagen, R., Menichetti, E., 2012. Strategic choices for renewable energy investment:

Conceptual framework and opportunities for further research. Energy Policy 40, 1–10.

doi:10.1016/j.enpol.2011.06.050

Wüstenhagen, R., Wolsink, M., Bürer, M.J., 2007. Social acceptance of renewable energy

innovation: An introduction to the concept. Energy Policy 35, 2683–2691.

doi:10.1016/j.enpol.2006.12.001

Yang, C.J., 2010. Reconsidering solar grid parity. Energy Policy 38, 3270–3273.

doi:10.1016/j.enpol.2010.03.013

Yang, H., Yagar, S., Iida, Y., 1994. Traffic assignment in a congested discrete/ continuous

transportation system. Transp. Res. Part B Methodol. 28, 161–174. doi:10.1016/0191-

2615(94)90023-X

Yelle, L.E., 1979. the Learning Curve: Historical Review and Comprehensive Survey. Decis. Sci.

10, 302–328. doi:10.1111/j.1540-5915.1979.tb00026.x

Yin, R.K., 2003. Case Study Research: Design and methods. Thousand Oaks: Sage Publications.

Yin, R.K., 2011. Applications of Case Study Research, 3rd Editio. ed. Sage Publications.

Yuan, X., Zuo, J., Ma, C., 2011. Social acceptance of solar energy technologies in China—End

users’ perspective. Energy Policy 39, 1031–1036. doi:10.1016/j.enpol.2011.01.003

Zhai, P., 2013. Analyzing solar energy policies using a three-tier model: A case study of

photovoltaics adoption in Arizona, United States. Renew. Energy 57, 317–322.

doi:10.1016/j.renene.2013.01.058

Zhai, P., Williams, E.D., 2012. Analyzing consumer acceptance of photovoltaics (PV) using fuzzy

logic model. Renew. Energy 41, 350–357. doi:10.1016/j.renene.2011.11.041

Zhang, S., Andrews-Speed, P., Ji, M., 2014. The erratic path of the low-carbon transition in China:

Evolution of solar PV policy. Energy Policy 67, 903–912. doi:10.1016/j.enpol.2013.12.063

Zhang, X., Shen, L., Chan, S.Y., 2012. The diffusion of solar energy use in HK: What are the

barriers? Energy Policy 41, 241–249. doi:10.1016/j.enpol.2011.10.043

Zhang, Y., Song, J., Hamori, S., 2011. Impact of subsidy policies on diffusion of photovoltaic

power generation. Energy Policy 39, 1958–1964. doi:10.1016/j.enpol.2011.01.021

Zimmerman, D., 2003. Peer effects in academic outcomes: Evidence from a natural experiment.

Rev. Econ. Stat. 85, 9–23.

Zott, C., Amit, R., Massa, L., 2011. The Business Model: Recent Developments and Future

Research. J. Manage. 37, 1019–1042. doi:10.1177/0149206311406265

135

Zubaryeva, A., Thiel, C., Zaccarelli, N., Barbone, E., Mercier, A., 2012. Spatial multi-criteria

assessment of potential lead markets for electrified vehicles in Europe. Transp. Res. Part A

Policy Pract. 46, 1477–1489. doi:10.1016/j.tra.2012.05.018