Public policy influence on renewable energy investments –
a longitudinal study across OECD countries
This version: June 2014
Friedemann Polzin*1,2, Michael Migendt1, Florian A. Täube1 & Paschen von Flotow2
Abstract:
This paper examines the impact of public policy measures on renewable energy (RE) diffusion
through corresponding investments in electricity-generating capacity made by institutional investors
(i.e. investment/pension funds, banks and insurance companies). Capacity investment data is gathered
from Bloomberg New Energy Finance (BNEF) and policy indicators from the IEA/IRENA Policy and
Measures Database. The authors observe the influence of different policy measures on a sample of
OECD countries during an 12-year period (2000-2011) to suggest an effective policy mix which could
tackle existing path dependencies and failures in the market for clean energy. The results call for
technology-specific policies which take into account actual market conditions and the position in the
technology life cycle. To improve the environment for institutional investments, advisable policy
instruments include economic/fiscal incentives such as feed-in tariffs (FIT) with grants and subsidies
being less effective. Additionally market-based instruments such as greenhouse gas emission trading
systems for mature technologies should be included. These policy measures directly impact the risk
and return structure of RE projects. Supplementing these with regulatory measures such as codes and
standards (e.g. RPS), and long term strategic planning could further strengthen the environment for RE
investments.
Keywords: renewable energy, public policy mix, institutional investors, longitudinal analysis
JEL codes: G28, O33, O38, Q42, Q48
*Corresponding author: polzin@sbi21.de
1
EBS Business School, Strascheg Institute for Innovation and Entrepreneurship (SIIE), Rheingaustr. 1,
65375 Oestrich-Winkel, Germany
2
Sustainable Business Institute (SBI), Burgstr. 4, 65375 Oestrich-Winkel, Germany
1
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
1
Introduction
Climate change has been aggravating in recent years as CO2 emissions continue to grow in
developed and developing countries alike (IPCC, 2014). OECD countries have a larger
responsibility to address these issues, as they accounted for almost 50% of the global carbon
emissions in 2010, most of which were linked to the energy sector (BNEF, 2013; Müller et al.,
2011; OECD, 2013)
While developed countries bear a historically larger responsibility to address climate change
issues, it is now clear that transitioning towards a low-carbon society will involve decoupling
economic growth and prosperity from growing CO2 emissions, respectively encouraging
“green growth”. One possibility is to increase the share of energy (especially electrical
energy) generated by renewable energy (RE) sources (Foxon and Pearson, 2007; Jefferson,
2008)
Investments in clean energy1 companies, projects and infrastructure have been growing in the
last decade and total up to a relevant amount. Remarkable investment markets have been
created based on RE technologies (Figure 1). Yet the volumes are still relatively small when
compared to investments in conventional fossil fuel-based power (BNEF, 2013).
Figure 1 here
This is partially due to the challenge which the diffusion of RE technology presents (Friebe et
al., 2014, 2013; Veugelers, 2011). Market failures occur related to inherent characteristics of
the energy sector (designed for fossil fuels-based power plants) on the one hand, as well as the
particular nature of RE technologies on the other (Helm, 2002). Long payback periods and
illiquid assets coupled with high regulatory dependencies and corresponding uncertainties
often make RE unattractive or even unsuitable for investors. While this holds true even for
mature technologies and conventional power plants, REs still face some technological
uncertainty (Kenney and Hargadon, 2012).
To mitigate market failures and compensate for technological and economic weaknesses,
policy makers were historically confronted with a variety of options to stimulate diffusion and
“green growth” while de-linking it from CO2 emissions. This article attempts to uncover the
effects of different policies to support clean energy applications in the future. It adds to recent
academic and political discussions about the choice among feed-in tariffs (FITs) and
1
Throughout this paper, we will use renewable energy (RE) and clean energy interchangeably.
2
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
alternative mechanisms as well as the corresponding overall effectiveness and efficiency of
these policy instruments (Carley, 2009; Lesser and Su, 2008; Mathews et al., 2010). Building
on previous work on RE investor behavior (Bergek et al., 2013; Wüstenhagen and Menichetti,
2012), it investigates the influence of public policies on subsequent RE investments by
institutional investors across OECD countries over a time period of 12 years (2000-2011).
The overarching research question therefore is: Which policies have proven (most) conducive
to investments in renewable energy assets?
The remainder of this paper is structured as follows: Section two briefly describes the
conceptual motivation. Section three introduces the analytical method and data used. Section
four depicts the results while section five presents the discussion, conclusion, limitations and
next steps in the research process.
2
Conceptual background
2.1. Clean energy innovation and diffusion
RE technologies are a prime example of issues associated with commercialization and
diffusion of innovations. Innovative technologies face difficulties from the moment they are
introduced to markets, as their potential and quality cannot be evaluated ex-ante (Arrow,
1962; Jaffe et al., 2005). These difficulties lead to market or system failures which provide
grounds for policy intervention (Dodgson et al., 2011; Martin and Scott, 2000). In addition,
other externalities, such as path-dependencies and external costs (i.e. for CO2 emissions),
occur. Path dependencies are particularly high in heavily regulated sectors such as energy, and
need special attention from the responsible policy makers (Brown, 2001; Mowery et al., 2010;
Sandén and Azar, 2005). Hence, a multitude of short-term instruments need to be found for
long-term climate targets, which according to Sandén and Azar (2005) range from R&D
support to technology-specific subsidies and niche market creation. Historically, mitigation
policies that target the early stage (i.e. generation phase) have been deployed (CárdenasRodríguez et al., 2013; Veugelers, 2012, 2011).
In later stages (i.e. diffusion phase) the distinct features of RE are subject to severe market
failures due to high path dependency in fossil fuels (Müller et al., 2011; Popp et al., 2011).
Here, policy support is limited so other regulatory measures to stimulate markets need to be
established, even though diffusion and application of RE technologies is socially and
politically desirable (Jefferson, 2008). In addition, the highly regulated environment for
3
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
diffusion of mature RE might require new forms of regulation compared to the
commercialization of RE in order to overcome market failures and dissolve path dependencies.
Literature on energy policy has analyzed these relationships in a number of different ways,
generating evidence for policy makers to support their decisions (Jacobsson et al., 2009).
Fundamental questions such as influence of price or quantity-based mechanisms have been
addressed (Menanteau et al., 2003). Scholars argue that in the sense of carbon and energy
market liberalization, preference should be given to market-based instruments as first best
solutions, e.g. carbon cap and trading systems (Helm, 2002; Rogge and Hoffmann, 2010;
Rogge et al., 2011; Smith and Swierzbinski, 2007). However in the absence of global carbon
markets and their inability to address incubation efforts (e.g. rapid development) and costefficiency, second-best policy instruments need to be considered (Jacobsson et al., 2009).
Within the range of policy choices, evidence points towards the superiority of FIT (Butler and
Neuhoff, 2008; Couture and Gagnon, 2010; Jenner et al., 2013; Lesser and Su, 2008) over
quota and obligation schemes (Butler and Neuhoff, 2008), renewable portfolio standards
(RPS) (Delmas and Montes-Sancho, 2011; Carley, 2009) and tax-based incentives (Cansino et
al., 2010; Quirion, 2010). One solution to address the problems encountered so far, apparently,
is a ‘policy mix’ consisting of complementary instruments. However, there is no scholarly
consensus on what the optimal policy mix could look like (Foxon and Pearson, 2007) or on
which criteria should be applied to determine it (Carley, 2009). Relatedly, the question arises,
if there is one optimal policy mix or several ones, contingent on different other factors?
2.2. Investor perspective on RE and public policy
In an attempt to answer these questions, scholars investigated the implications for private
sector investment of policy instrument usage (Bergek et al., 2013; Delmas and MontesSancho, 2011; Masini and Menichetti, 2012; Wüstenhagen and Menichetti, 2012), given that
private finance mobilization is regarded a core component of strategies to foster development
of clean technologies and their complementary infrastructure (Mathews et al., 2010; Mowery
et al., 2010). Mathews et al. (2010, p. 3263) note that “the issue of public vs. private financing
is not yet adequately explored”, but add that there is consensus among policy makers that the
transition to a low carbon economy will not happen without the involvement of private
institutional investors.
Investments by institutional investors are typically hindered by a number of factors; high
upfront costs, risks and uncertainty regarding long-term viability of the technology, long
4
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
payback periods, high regulatory and infrastructural dependency as well as public acceptance
(Cárdenas-Rodríguez et al., 2013; Haley and Schuler, 2011; Kenney and Hargadon, 2012;
Müller et al., 2011). These factors directly influence the risk/return profile of an RE
investment, which is a major determinant for institutional investors.
To mitigate the economic barriers, de Río and Bleda (2012) underline the superiority of feedin tariffs to spur deployment and technological diversity and lower risks associated with RE
technologies. They argue that a variety of policies, consisting of technology-specific and
technology-neutral measures (“systemic package” – knowledge generation and diffusion) is
needed to not only enhance deployment of mature technologies but also the subsequent
creation of new technologies.
Bergek et al. (2013) consider the evaluation criteria used by the heterogeneous group of RE
investors, such as (overall or portfolio) cost, perceived (market) uncertainty and political risk.
They argue that purely economic analyses (i.e. focusing on risk and return) fall short of
capturing the wide range of factors influencing the decisions and processes for investing into
RE technologies, with one being the reaction to different policy instruments. CárdenasRodríguez et al. (2013) distinguish between instruments that directly depend on public
budgets (e.g. tax incentives or grants and subsidies), support measures oriented towards
mitigating the externalities (regulations) and price-based systems (FIT).
The ultimate requirement for a sustainable RE policy is a reduction of capital costs to create a
level playing field with fossil fuel-based technologies which have been heavily subsidized in
the past (Szabó and Jäger-Waldau, 2008). Thus a monitoring of these costs is crucial. Szabó
& Jäger-Waldau (2008) suggest a more competitive financial environment could actually
reduce the costs of capital for RE projects, given that capital markets function efficiently.
Decreasing support to encourage technological improvements as well as tradability of RE
certificates might further spur the deployment of renewables. Above all policy instruments
that influence the risk and return structure of RE projects impact investor behavior (CárdenasRodríguez et al., 2013).
In addition, investment decisions can be related to policy instruments that do not directly
impact the risk and return structure of RE projects. For example, the perception of investment
opportunities and preference for short term or long-term incentives, also influence the
decision to invest (Masini and Menichetti, 2012). In this respect, Wüstenhagen & Menichetti
(2012, p. 9) specifically call for investigating the role of policy on the perception of
institutional investors: “[…] how can policy build on this knowledge to devise more effective
5
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
regulatory frameworks in an evolutionary and boundedly rational world? What is, for example,
the relative importance of ‘‘symbolic’’ policies compared to the actual monetary value of
incentives?” Thus policy measures that do not directly impact the risk and return structure of
RE projects but influence investor behavior indirectly need to be considered as well.
In sum, the literature has been searching for an integral overview of sustainable energy to spur
RE diffusion by private investments. In order to contribute to this debate, our research aims at
uncovering the effectiveness of different policy instruments to induce private finance in RE
assets.
3
Methods and data
3.1. Research design
Investigating the diffusion of a particular technology and corresponding investments requires
a longitudinal research design (Angrist and Pischke, 2008; Hair, 2010). Major works in the
field of energy policy have been of a qualitative nature, presenting in-depth evidence from
case studies in a small number of countries. On the other hand, energy economists apply
sophisticated econometric methods, mostly panel data regressions (e.g. Cárdenas-Rodríguez
et al., 2013; Jenner et al., 2013 for FIT; Johnstone et al., 2010; Popp et al., 2011). Most
quantitative studies have been carried out in the EU (e.g. Marques and Fuinhas, 2012a) with a
few comparing OECD or BRIC countries (Cárdenas-Rodríguez et al., 2013; Johnstone et al.,
2010; Popp et al., 2011).
To provide more generalizable results we cover a variety of countries, thus conducting a panel
regression. We use the time period from 2000 to 2011 to explain the influence of policy
instruments on the diffusion of clean energy technologies. This time frame is chosen because
it covers substantial developments in the worldwide renewable energy sector. Globally the
wind sector grew from 18 GW installed capacity in 2000 to 238 GW installed capacity in
2011, while the solar sector grew from 1.5 GW installed capacity in 2000 to 67 GW installed
capacity in 2011 (IEA, 2014). As policy instruments do not exhibit an immediate effect on
technology application, we add a lag procedure (Wooldridge et al., 2009).
Our descriptive analysis of the data reveals a strong relationship between the use of RE policy
and investments in RE projects by institutional investors. Initial case studies of Italy and the
UK show, on one hand, the high dependency of investments on regulation and on the other
hand, a time-dependent phenomenon, i.e. investments lagging behind policy measures.
Figure 2&3 here
6
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
Italy has favorable conditions for solar energy generation. Thus it is one of the earlier
adopters of solar energy and has built up a sizable amount of capacity. This growth is
supported by a policy mix targeted at the solar sector. The first policies targeting the solar
sector were established in the year 2000. The “Renewable Electricity support - PV Roofs
Programme” classified as a regulatory instrument and promoted the installation of solar PV,
thereby liberalizing electricity production from small solar PV installations. According to the
decree from the Ministry of Environment, the sale price of excess power to the grid was set
equal to the purchase price from the grid, independent of time and season. The next bigger
step in supporting the market was the “Feed-In Premium for photovoltaic systems”
established in the year 2005. This financial incentive, which is still in force today (albeit after
several alterations), initiated the rapid expansion of solar investments in the year 2006 and
following. This FIT policy included a set of tariffs, valid for a period of 20 years. Electricity
producers got a fixed price premium in addition to the price of the electricity sold. Further
policies introduced in 2007 and 2008 spurred the investments into solar even further.
The United Kingdom has similar positive natural conditions for the wind sector. The first
measure implemented to foster wind energy was the “Offshore Wind Capital Grants Scheme”,
set up in 2002. This grants scheme was implemented to stimulate early deployment of
offshore wind through direct support, and the grants could cover up to 40% of the eligible
project costs. The “Energy Act of 2004”, a regulatory instrument, awarded licenses for wind
farm sites in designated Renewable Energy Zones. The “Climate Change and Sustainable
Energy Act” of 2006 further enriched the regulatory landscape by reporting on greenhouse
gas emissions and actions taken to reduce them. In 2010, a “Feed-in Tariffs for Renewable
Electricity” was added, which mandated additional payments for renewable energy fed into
the grid or consumed on-site. This FIT varies according to technology and project size and
covers 20 years of electricity production.
We observe a growing trend of more policies over the years. These policies are lagged by
investments until a sudden drop of active policies as well as investments in 2010 and further
decreasing investments in 2011. This drop is also likely related to the financial crisis from
2008 and later.
3.2. Data
Data was collected from two independent data sources. Investments in RE capacity have been
drawn from Bloomberg New Energy Finance (BNEF), which possesses one of the most
comprehensive databases in the field of clean technology financing (Cárdenas-Rodríguez et
7
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
al., 2013). It contains information on installed electricity generating capacity, date, transaction
type, financing type, equity and debt during the years 2003-2012. The database includes 5840
Solar Investments, 9643 Wind Investments and 2889 Biomass & Waste Investments. These
three RE subsectors account for 72% of the RE investments and with 74 GW installed for
75% of all capacity additions in the period, have therefore been selected for further analysis.
Policy indicators were drawn from the IEA/IRENA Policy and Measures Database with other
data sources2 consulted to check the robustness of the descriptions3. These indicators have
been used by scholars to analyze the evolution and clustering (Nicolli and Vona, 2012; Vona
et al., 2012) as well as the impacts of aggregated policy instruments in Europe (CárdenasRodríguez et al., 2013; Marques and Fuinhas, 2012a, 2012b). A priori we included all
indicators ‘Economic Instruments’, ‘Information and Education’, ‘Policy Support’,
‘Regulatory Instruments’, ‘Research, Development and Deployment (RD&D)’ as well as
‘Voluntary Approaches’. The aggregation procedures resulted in 7835 policy data points
resulting from 957 distinct policy measures (see appendix for a detailed description of the
indicators). This allows comparing the experience of many countries and decomposing the
effect of distinct factors econometrically (Cárdenas-Rodríguez et al., 2013).
We structured the data according to sectors (i.e. Multiple RE sources, Wind, Solar, Biomass)
and applied additional data processing: First, we limited the timeframe for IEA policy
measures to 2000-2011. BNEF investment data is available from 2003. This approach
permitted us to include the lag structure (i.e. investments lagging behind policy measures).
We further decided not to include the year 2012, as data quality did not meet the standards of
previous years. This is due to the fact that BNEF staff continuously updates investments even
for the previous year. We removed cases with missing values and included only completed
deals. Second, regarding the selection of cases (i.e. countries), our interest in the influence of
policy measures meant the exclusion of countries with less than three consecutive years of
investor activities (DV). The case-selection is carried out for each sector we analyze (see
section 3.3. for the procedure and the Table A.1 appendix for a list of excluded countries).
The selection of policy instruments to be included in the model is described in section 3.3.2.
2
3
REN21, DSIRE, OpenEI, Clean Energy Sources
The timeframe here included the years 1990 to 2012.
8
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
3.3. Model
We investigate the influence of different policy measures on subsequent investments into RE
capacity by institutional investors. The components of this model are explained in Figure 4.
The model set up permits, depending on the sector, to estimate from 176-435 parameters (i.e.
coefficients for the different policy instruments). The included variables are log-transformed
to correct for the skewed distribution of both dependent and independent variables (Hair,
2010).
Figure 4 here
3.3.1. Dependent variable
The dependent variable for the overall model is measured as aggregated installed capacity (in
MW) in a certain country and year in a specific subsector (e.g. solar, wind, biomass). We use
capacity indicators since they represent the most accurate proxy for the deployment of a
technology (Popp et al., 2011). This variable is constructed by aggregating the newly installed
electricity generating capacity of all projects financed by institutional investors per year and
country.
3.3.2. Independent variables
The main independent variables are constructed using distinct policies (regulatory instruments,
grants, …) which are active per country per year. These are measured by the number of active
instances regulating the RE sector (Johnstone et al., 2010; Marques and Fuinhas, 2012a).
We can further distinguish between technology specific instruments (e.g. specific targets for
certain energy source) and instruments that apply to all types of renewable energy (e.g.
German FIT). The scheme is taken from the IEA Policies and Measures database 4 and
included in the appendix. This database provides relevant information on characteristics (Title,
country, year (started and ended), Policy Status (e.g. in force, ended, superseded), policy type
(see policy scheme), policy target (e.g. subsector such as solar, wind, etc.), geographical
scope (Supranational, national, regional), policy sector (e.g. electricity, multi-sectoral,
framework policy), size of plant targeted (Large, small or both) and funding (partially,
depending on instrument)). We selected distinct policies based on previous studies. The
selected instruments include (Table 1):
4
http://www.iea.org/policiesandmeasures
9
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
Table 1: Policy Scheme (adapted from IEA Policies and Measures) here
3.3.3. Controls
To account for technological progress, economies of scale as well as the fact that the installed
capacity gains momentum (leading to variance from the previous years), we include year and
country dummy variables. To rule out alternative explanation for RE investments we included
a number of control variables in the regressions.
Technology advancement: Economies of scale are picked up by time dummy variable and
therefore not included among the control variables.
Economic indicators: Further economic indicators that might drive capacity additions
including GDP (c_GDP)
Energy system: To account for differences in energy use and consumption we include energy
dependency (CO2 intensity – Metric Tons of Carbon Dioxide per thousand year 2005 U.S.
dollars GDP) as well as electricity consumption in the regression. (c_CI, c_TEC)
Investor behavior: To account for factors influencing investor behavior, we include interest
rates (c_LIR) as well as share prices (c_SP) of local indices as these might render an
investment into RE vs. non-RE more or less attractive.
3.3.4. Lag structure
To account for the time-dependent influence of policy measures on investor behavior, we
include a lag structure in the analysis (Angrist and Pischke, 2008; Wooldridge et al., 2009).
On the one hand, it is possible that investors anticipate the regulation and already have their
projects ready when it is passed, as the regulatory process is mostly open. On the other, there
are factors delaying the investment process such as the time needed to build the wind farm or
solar park and to gain access to the grid (Jenner et al., 2013).
3.4. Analysis
The evaluation of the outcomes of RE policies has differed in the last 25 years. Konidaris &
Mavrakis (2007) construct a multi-method evaluation tool, highlighting cost effectiveness and
(economic) efficiency (Finon and Perez, 2007; Foxon and Pearson, 2007) as the main criteria
for measuring the impact of climate change-mitigating policies. Still, questions arise over the
quantification of support mechanisms (Sukumar et al., 2010).
Determining the influence of policy measures on investments in RE capacity proves to be a
complex task. A priori, we expect spatial as well as temporal effects. We anticipate the
10
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
number of countries and the number of policies and measures to be reinforced over time.
Assuming that a set of countries (e.g. EU countries) is exposed to common policy guidance on
RE, we expect contemporaneous correlations. Academic literature describes some drivers for
RE use, which are included in the model (see section 3.3.2). To fulfill all the demands on a
model, the best fitting econometric approach is to use a panel data approach, which is flexible
enough to work under the complex conditions of RE deployment (Marques and Fuinhas,
2012a).
We estimate Random Effects, and Panel Corrected Standard Error (PCSE) models. We
perform our panel data estimation upon complex error compositions. Heteroskedasticity,
panel autocorrelation, and contemporaneous correlation phenomena are addressed through
fitting approaches (Reed and Ye, 2011). Thus we circumvent inefficiency in coefficient
estimation and biasedness in the estimation of standard errors. We do not focus on random
effects estimators (REE) as they do not address serial correlation and contemporaneous
correlation, however we include the estimates for reasons of robustness, as suggested by
Marques and Fuinhas (2012a, 2012b).
We use the following procedure: 1. We observe the quality and nature of the data; 2. We test
the presence of heteroskedasticity, panel autocorrelation, and contemporaneous correlation; 3.
If deviation from the classical assumptions is noticeable, we apply the PCSE estimator, which
is a suitable solution with the presence of panel-level heteroskedasticity and contemporaneous
correlation of observations among panels; 4. We compare the results with those derived from
the REE to check the robustness (Marques and Fuinhas, 2012a).
3.4.1. Panel data analysis
Table 2 presents results from the estimations and confirms that especially the policy data is
heteroskedastic (i.e. has a common variance) and that panel autocorrelation and
contemporaneous correlation is present.
Table 2: Specification tests here
3.4.2. Panel data regressions
As public policy effects differ across RE subsectors (e.g. solar, wind, biomass) we carried out
the analysis sector by sector and also aggregated the data (Multiple RE sources) to analyze
effects that are similar across sectors. Thereby we can also distinguish policy instruments
between the sectors as well as policies that apply to all sectors.
11
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
Additional analyses have been performed using aggregated independent policy measures
following the IEA policy scheme (Table 1) and the corresponding categories (e.g. Economic
Instruments - Direct investment, Economic Instruments – Fiscal/financial incentives etc.)
Panel data estimation without lag procedure (I)
∑
Panel data estimation with lag procedure (II-IV)
∑
is the aggregated installed capacity financed by institutional investors per country j per
year k.
is a vector of i explanatory variables representing policy measures based on the
IEA scheme (per country per year).
consists of a number of control variables. For the
analyses of time-dependent phenomenon we include lags “l” of one to three years in the
regressions. The PCSE estimator allows the error term
allows the use of first-order autoregressive process for
to be correlated over the countries,
over time and allows
to be
heteroskedastic (Cameron and Trivedi, 2009; Marques and Fuinhas, 2012a).
4
Results
4.1. Descriptive statistics
Following the initial analysis of selected countries (see chapter 3.1) we further aggregated
investments in capacity of multiple RE and policy measures taken over time to gain an
overview about the relationship between the dependent and independent variable. Figure 5
provides the relationship between the two. One dot represents one country. It becomes
apparent that some countries have neither attracted any investments nor instituted policy
measures whereas other countries have implemented an above average number of policy
measures with regard to attracted investments. This confirms our assumptions that some
policy measures are more effective than others in attracting RE projects.
Figure 5 here
Furthermore the correlation among explanatory variables has been subject to analysis. The
simultaneous use of several drivers leads to the hypothesis of collinearity among explanatory
variables. Table A.2 and Table A.3 (in the appendix) show the summary statistics and the
12
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
correlation coefficients for our analysis. The analysis suggests the absence of collinearity
among variables.
4.2. Panel-Regression
We estimated all models separately using the PCSE and the random effects estimator. We
conducted the analysis for Multiple RE data and distinct sectors. Analyses have been
performed structuring the policy measures according to the IEA policy scheme (Table 1). The
estimation results are displayed in order of the categories and different policy structures.
4.2.1. Multiple RE sources
The results of our complete policy variable analysis using Multiple RE are presented in Table
3. Findings confirm that six of the variables show a significant influence on capacity additions
in all energy sectors combined. PCSE shows positive, highly significant contributions by
Feed-in tariffs/premiums and GHG allowances, which directly impact the risk and return
structure of RE projects and thus provide an incentive for investors. Grants and subsidies
temporally reduce the cost of finance for a project, and directly depend on a public budget.
Interestingly, the presence of a GHG trading system has a stronger effect on RE capacity then
an FIT or grants and subsidies. Feed-in tariffs have been implemented in a range of countries,
starting with Germany and Austria in 2000, while GHG emission trading systems have been
introduced in the Australia, UK, Italy and Norway since 1991. In addition, supportive
instruments such as a clear long term energy strategy (Strategic planning), as well as
regulatory instruments (Codes and standards),especially RPS, for the use of RE sources prove
to be effective.
The PCSE and random effects estimators display consistent results, however when adding
further restrictions such as robust standard errors or autocorrelation assumptions, some policy
instruments become insignificant (e.g. Codes and standards and Strategic planning). This
might be due to the fact that statistical noise is more pronounced in the cross-sectoral data.
Table 3: Results for Multiple RE here
4.2.2. Solar
The results of our complete policy variable analysis using solar energy data are presented in
Table 4. Findings confirm that seven of the policy variables show a significant influence on
capacity additions in all energy sectors combined. PCSE estimator shows positive
contributions with high significance of Feed-in tariffs/premiums as a policy instrument that
13
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
guarantees a certain return of investment of time. FIT for solar technologies have been
particularly successful in countries such as Germany and Italy. As a supporting instrument,
investors favor a long term framework with a clear vision (Strategic planning). Essentially all
countries in our study had policy measures containing a strategic component, however only a
few incorporated it in many policy initiatives.
On the other hand we found two instruments which prove to be ineffective. GHG emissions
allowances show a negative impact on the capacity financed by institutional investors, which
might be due to the lack in maturity and the low amount of certificates generated per capital
invested compared to other sectors, such as wind. Green certificates that permit trading the
obligatory RE capacity in a national scheme do not incentivize institutional investors, as these
depend on the total quantity in the market which might vary (especially in sunny periods).
Without the assumption of auto-correlation other instruments are significant as well. These
results are mostly consistent across estimators (PCSE and REE).
Table 4: Results for Solar here
4.2.3. Wind
The results of the complete policy variable analysis using Wind energy data are presented in
Table 5. Findings confirm that three policy variables show a significant influence on capacity
additions in all energy sectors combined. The PCSE estimator shows the positive
contributions of Feed-in tariffs/premiums, as a measure that directly impacts the risk and
return structure of wind projects. Interestingly the presence of GHG emissions allowances
have a stronger impact on the capacity financed by institutional investors than FIT, as
investors prefer market-based systems which are less dependent on policy changes. GHG
emissions allowances have been introduced in the US, UK and Italy. Institutionalization of
markets in the form of Codes and standards is further conducive to RE capacity additions for
mature technologies. These RPS have mainly been deployed throughout North America and
in parts of Europe. The random effects estimator confirm these findings.
Table 5: Results for Wind here
4.2.4. Biomass
The complete policy variable analysis using Biomass energy data is presented in Table 6.
Findings confirm that five policy variables show a significant influence on capacity additions
in all energy sectors combined. Positive contributions with high significance include Funds to
sub-national governments (direct investments with federal money with regional, local or
14
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
municipal level entities as intermediaries or targets) as biomass markets tend to be regionally
dispersed. Grants and subsidies prove to be effective as short term measures to alleviate
finance constraints. The presence of a GHG emissions trading system as a market based
system did also spur the investments in biomass plants. This might be due to the fact that
these plants generate a constant flow of certificate an exhibit base-load characteristics.
Institutional creation such as the forming an energy agency, further accelerates the capacity
additions in the biomass sector. Infrastructure investments do not increase the capacity
installed in the biomass sector which might be due to the fact that bioenergy can be used
locally. Here both estimators display consistent results as well.
Table 6: Results for Biomass here
Finally, the disaggregated results (i.e. individual policy instruments) were cross-checked with
aggregated results (categories of policy instruments) for the entire analysis. An overview is
given in figure 6.
Figure 6 here
5
Discussion
RE technologies are subject to market insufficiencies and system failures. Therefore scholars
have been discussing the policy mix for an effective support for RE technologies (Foxon and
Pearson, 2007; Jacobsson et al., 2009). This discussion was transformed by questions
concerning energy system transformation (Jefferson, 2008). Theoretically, market based
solutions (such as emissions trading systems) would be preferable. However, due to the
absence of such functioning global markets for carbon emissions, second best solutions have
been implemented across countries with mixed effects. Research found that FITs proved
particularly successful in countries such as Germany with some exceptions in other countries
(e.g. Spain) (Büsgen and Dürrschmidt, 2009; Couture and Gagnon, 2010; Jenner et al., 2013).
Other policies such as quotas, renewable portfolio standards and tax based measures have
been applied as well, revealing mixed evidence (Butler and Neuhoff, 2008; Cansino et al.,
2010; Carley, 2009; Quirion, 2010). In sum, single policies have been subject to analysis on a
country or industry level.
A second stream of literature deals with investor behavior regarding RE technologies, as
investors provide funds for large scale deployment (Bergek et al., 2013). This perspective has
not been adequately explored (Mathews et al., 2010). Evidence by Río and Bleda (2012)
underlines the superiority of FITs to spur deployment and to lower risks associated with RE
technologies, however the authors argue that a variety of policies, consisting of specific and
15
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
technology-neutral measures provide fertile ground for RE technology deployment. Calls for
future research explicitly state the need for an advanced (quantitative) analysis of the
influence of policy measures on investor decision making (Wüstenhagen and Menichetti,
2012)
With our analysis we contribute to both realms of literature, providing an integral picture of
RE policies and their influence on RE capacity investments by institutional investors. The
analysis is conducted on a sectoral basis to allow differentiated policy recommendations. In
the following discussion, we highlight significant effective and ineffective policy measures.
First, the three RE subsectors have been aggregated to allow for an overall policy analysis.
The results highlight the robust positive influence of tradable permit systems as a market
based instrument and the less pronounced positive influence of fiscal incentives such as FIT
as well as grants and subsidies on subsequent capacity investments. Regulatory mechanisms
such as codes and standards (especially RPS) also attract institutional investors. These
findings confirm earlier work by Jenner et al. (2013) who highlight FIT as effective measures
to increase the overall capacity of RE. We also provide new insights in the discussion about
RPS schemes. Looking at US states, Carley (2009) found a positive effect, whereas Delmas
and Montes-Sancho (2011) did not find a significant positive contribution of RPS to RE
investments. Adding to this literature we underline positive effects of GHG allowance system.
Finally, policy support mechanisms such as strategic planning prove to be effective. These
results confirm conceptual and empirical works by Wüstenhagen & Menichetti (2012) and
Lüthi & Prässler (2011), which hold clear strategic long-term economic instruments to be
conducive to RE investments. While the multiple RE analysis includes different technologies
at different stages in their life cycle, it only allows for a rough overview on policy instruments.
Therefore, in a second step, we conduct an individual sectoral analysis for solar, wind and
biomass.
A detailed analysis of the solar sector confirms the strong role of FITs for market
development (Couture and Gagnon, 2010; Jenner et al., 2013). This instrument is a strong
signal to investors as it addresses the capital market restrictions by adjusting the risk/return
structure (Cárdenas-Rodríguez et al., 2013). We add empirical support for the strong role
which a long-term policy commitment (strategic planning) plays in an effective policy mix
(Bergek et al., 2013). The German “Energiewende” provides a successful example of this type
of policy measure.
16
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
Solar energy technologies, being less mature and more diverse than wind, are more heavily
dependent on regulation, although grid parity is almost reached. These developments are
reflected in our results regarding the policy mix. Market based incentives (such as GHG
emission trading systems) prove to be ineffective in this case. A possible explanation lies in
the fact, that solar technologies require stronger signals towards investors due to their relative
novelty compared to wind technologies. Furthermore, green certificates are a second
ineffective policy instrument due to limited implementation, insufficient mobilization of funds
and high regulatory uncertainty, all of which make them unattractive for institutional investors.
In addition, quota-based systems tend to be opaque as they involve over-the-counter
transactions for certificates (Cárdenas-Rodríguez et al., 2013).
Wind sector results add empirical evidence to the debate revolving around FITs and tradable
permits (Butler and Neuhoff, 2008; Cansino et al., 2010). Our results show a stronger
influence of tradable permits as market based system than FITs for wind. Regulatory
measures such as codes and standards (e.g. RPS) have also proven to be effective, perhaps
because the wind sector shows elements of a developed market. The cost-effectiveness of this
technology is proven so it can compete with fossil fuel-based electricity generation in certain
environments.
Our analysis of the biomass sector, however, yields inconclusive results. On one hand, grants
and subsidies as direct economic instruments prove to be effective, however these are only
short-term policy measures. Long term market based instruments (GHG emission trading
system) complement these and prove to have positive effect as well. On the other, indirect
measures such as “funds to subnational governments” and institutional creation seem to be
effective as well. According to our results, infrastructure investments to provide grid access
and strategic planning to develop a long-term energy supply based on bioenergy are
ineffective for channeling investor’s money into biomass technologies. Reasons for deviating
results in the biomass energy sector lie in the different structure which can be characterized by
a strong regional focus and usually the small scale of power plants (Upreti, 2004). This might
attract a different set of investors that focus less on overall market conditions.
6
Conclusions
Our research calls for technology specific policies, taking into account the actual market
conditions and the position in the technology life cycle to design a supportive and effective
policy mix. We thus provide empirical support for policy implications by the IEA. “Policy
needs to take into account the overall maturity of the technology and the state of its market on
17
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
a global scale” (Müller et al., 2011, p. 31). To foster investments by institutional investors
during the diffusion phase the policy mix needs to contain elements of economic/fiscal
incentives (such as FITs) as well as market based incentives such as GHG emission trading
system which directly impact the risk/return structure of RE projects. Grants and subsidies are
helpful in the R&D and commercialization phase yet prove to be less effective in the later
stages as they depend on public/fiscal budgets. Regulatory instruments such as codes and
standards (especially RPS) accelerate the diffusion of RE technologies. Our results further
indicate support for policy measures such as strategic planning which provide investors with
the necessary prospects and goals concerning the different technologies to adjust their
investment strategy.
With our research, we firstly extend work on effectiveness of renewables policy (Marques and
Fuinhas, 2012a; Marques et al., 2010) and secondly provide empirical evidence for investor
influences (Masini and Menichetti, 2012; Wüstenhagen and Menichetti, 2012). Our results
also have implications for strategies of transitioning towards a clean energy system (Huberty
and Zysman, 2010; Jefferson, 2008), which is not possible without the integration of
institutional investors (Mowery et al., 2010). In this respect we indirectly complement
research on how to heal market failures (Mowery et al 2010, Dodgson et al 2011).
Our results reveal policy implications based on patterns for a suitable policy mix. Policy
makers interested in improving their country’s transition towards RE should implement
measures for attracting institutional investors, as the capital required for large-scale RE
projects by far surpasses the available funds of utility companies as well as the public budgets.
Institutional investors’ capital played an important role in the development of the RE sector,
and establishing a favorable environment for them, including specific policies, should
increase capacity additions in the future. A policy mix which contains the following elements
proved to be suitable for institutional investors:
Above all, our results strongly suggest the establishment of a reliable framework with a clear
vision and long-term policy objectives regarding the RE capacities to be installed in the future
as well as complementary transitions in the energy sector. Ex-post changes to the
remuneration of existing projects should be avoided. However, as technological progress
continues, the measures taken need to be adjusted, taking the market and technological
conditions (i.e. life cycle) into account.
Within this framework, monetary/fiscal and economic incentives are the most relevant policy
measures for investors. These directly impact the risk/return profile of RE projects and thus
18
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
their attractiveness. Investors are positive about long term reliable support mechanisms that
cannot be revoked and provide a highly predictable revenue stream. Feed-in tariffs thus
provide the more reliant and long-term signal than grants which depend on public budgets
however these funds, directly influencing the direct and early project cash flows, are also seen
favorable.
According to our results, market based incentives (such as GHG emission trading systems)
can also have strong influence on investments by institutional investors. These measures
support the need of investors for a highly reliable environment, best accompanied by a
diminished risk exposure. However, for an emission trading system to become an effective
anchor for institutional investors, the technology should have reached maturity, ruling out
technological difficulties.
Supportive regulatory measures such as codes and standards (especially RPS) further
accelerate the diffusion process of RE technologies by further reducing technological and
regulatory risk associated with investments in RE projects. Thus we recommend the
streamlining and strengthening of legislation and a transparent setting of renewable energy
targets
There are a number of limitations regarding study design and modeling. The use of dummy
variables for the policy measures does not allow for statements concerning policy
implementation or policy uncertainty (Bergek et al., 2013; Lüthi and Wüstenhagen, 2012;
Müller et al., 2011). Unlike e.g. Jenner et al. (2013) we thus cannot reflect on the design
features of the policy instruments.
Amending our fine grained policy analysis, future studies could look at the influence of these
and other policies on general capacity additions, including non-institutional investments (such
as households and utilities). Geographically our research could be extended to the BRICS
and/or less developed countries (LDC) which might alter the results due to an different
institutional setting (Friebe et al., 2014, 2013). It would be interesting to close the link
between early stage and later stage financing along the finance value chain for RE
technologies, thus analyzing venture capital and private equity investments in the early and
later stages which might lead asset finance investments.
Finally, with our analysis we also open up a discussion about the effectiveness and efficiency
of certain policy instruments for a low-carbon carbon economy. Thus policy makers might
want to consider both options for a relevant policy mix, depending on the stage of the
19
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
technology. Future research has to further clarify social costs for alternative policy
instruments, such as FIT and GHG emission trading systems (see also Müller et al. (2011).
7
Acknowledgements
The authors are grateful for the time and support of Martin Kenney and Donald Patton
(University of California, Davis) and Alex Coad and Paul Nightingale (SPRU – University of
Sussex). In addition the discussion at the ZEW Energy Conference 2014 helped us in further
refining our arguments. We would like to thank the Federal Ministry of Education and
Research (BMBF), Germany, for their financial support as part of the research project
‘‘Climate Change, Financial Markets and Innovation (CFI)’’.
20
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
8
References
Angrist, J.D., Pischke, J.-S., 2008. Mostly Harmless Econometrics: An Empiricist’s
Companion. Princeton University Press.
Arrow, K.J., 1962. Economic Welfare and The Allocation of Resources for Invention, in:
Nelson, R.R. (Ed.), The Rate and Direction of Inventive Activity, Princeton University
Press and NBER.
Bergek, A., Mignon, I., Sundberg, G., 2013. Who invests in renewable electricity production?
Empirical evidence and suggestions for further research. Energy Policy 56, 568–581.
doi:10.1016/j.enpol.2013.01.038
BNEF, 2013. Bloomberg New Energy Finance (BNEF). Bloomberg.
Brown, M.A., 2001. Market failures and barriers as a basis for clean energy policies. Energy
Policy 29, 1197–1207. doi:10.1016/S0301-4215(01)00067-2
Büsgen, U., Dürrschmidt, W., 2009. The expansion of electricity generation from renewable
energies in Germany: A review based on the Renewable Energy Sources Act Progress
Report 2007 and the new German feed-in legislation. Energy Policy 37, 2536–2545.
Butler, L., Neuhoff, K., 2008. Comparison of feed-in tariff, quota and auction mechanisms to
support wind power development. Renewable Energy 33, 1854–1867.
Cameron, A.C., Trivedi, P.K., 2009. Microeconometrics using stata. Stata Press College
Station, TX.
Cansino, J.M., Pablo-Romero, M. del P., Román, R., Yñiguez, R., 2010. Tax incentives to
promote green electricity: An overview of EU-27 countries. Energy Policy 38, 6000–
6008.
Cárdenas-Rodríguez, M., Johnstone, N., Haščič, I., Silva, J., Ferey, A., 2013. Inducing private
finance for renewable energy projects: Evidence from micro-data. OECD
Environment Working Papers.
Carley, S., 2009. State renewable energy electricity policies: An empirical evaluation of
effectiveness. Energy Policy 37, 3071–3081. doi:10.1016/j.enpol.2009.03.062
Couture, T., Gagnon, Y., 2010. An analysis of feed-in tariff remuneration models:
Implications for renewable energy investment. Energy Policy 38, 955–965.
doi:10.1016/j.enpol.2009.10.047
Del Río, P., Bleda, M., 2012. Comparing the innovation effects of support schemes for
renewable electricity technologies: A function of innovation approach. Energy Policy
50, 272–282. doi:10.1016/j.enpol.2012.07.014
Delmas, M.A., Montes-Sancho, M.J., 2011. U.S. state policies for renewable energy: Context
and effectiveness. Energy Policy 39, 2273–2288. doi:10.1016/j.enpol.2011.01.034
Dodgson, M., Hughes, A., Foster, J., Metcalfe, S., 2011. Systems thinking, market failure, and
the development of innovation policy: The case of Australia. Research Policy 40,
1145–1156. doi:10.1016/j.respol.2011.05.015
Finon, D., Perez, Y., 2007. The social efficiency of instruments of promotion of renewable
energies: A transaction-cost perspective. Ecological Economics 62, 77–92.
Foxon, T.J., Pearson, P.J.G., 2007. Towards improved policy processes for promoting
innovation in renewable electricity technologies in the UK. Energy Policy 35, 1539–
1550. doi:10.1016/j.enpol.2006.04.009
Friebe, C.A., Flotow, P. von, Täube, F.A., 2013. Exploring the link between products and
services in low-income markets—Evidence from solar home systems. Energy Policy
52, 760–769. doi:10.1016/j.enpol.2012.10.038
Friebe, C.A., von Flotow, P., Täube, F.A., 2014. Exploring technology diffusion in emerging
markets – the role of public policy for wind energy. Energy Policy 70, 217–226.
doi:10.1016/j.enpol.2014.03.016
21
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
Hair, J., 2010. Multivariate data analysis : a global perspective., 7th ed. /. ed. Pearson
Education, Upper Saddle River N.J., London.
Haley, U.C.V., Schuler, D.A., 2011. Government Policy and Firm Strategy in the Solar
Photovoltaic
Industry.
California
Management
Review
54,
17–38.
doi:10.1525/cmr.2011.54.1.17
Helm, D., 2002. Energy policy: security of supply, sustainability and competition. Energy
Policy 30, 173–184.
Huberty, M., Zysman, J., 2010. An energy system transformation: Framing research choices
for
the
climate
challenge.
Research
Policy
39,
1027–1029.
doi:10.1016/j.respol.2010.05.010
IEA,
2014.
IEA
FAQ
Renewable
Energy
[WWW
Document].
URL
http://www.iea.org/aboutus/faqs/renewableenergy/
IPCC, 2014. Climate Change 2014: Mitigation of climate change - IPCC Working Group III
Contribution to AR5. Int, Berlin.
Jacobsson, S., Bergek, A., Finon, D., Lauber, V., Mitchell, C., Toke, D., Verbruggen, A.,
2009. EU renewable energy support policy: Faith or facts? Energy Policy 37, 2143–
2146. doi:10.1016/j.enpol.2009.02.043
Jaffe, A.B., Newell, R.G., Stavins, R.N., 2005. A tale of two market failures: Technology and
environmental policy. Ecological Economics, Technological Change and the
Environment Technological Change 54, 164–174. doi:10.1016/j.ecolecon.2004.12.027
Jefferson, M., 2008. Accelerating the transition to sustainable energy systems. Energy Policy
36, 4116–4125. doi:10.1016/j.enpol.2008.06.020
Jenner, S., Groba, F., Indvik, J., 2013. Assessing the strength and effectiveness of renewable
electricity feed-in tariffs in European Union countries. Energy Policy 52, 385–401.
doi:10.1016/j.enpol.2012.09.046
Johnstone, N., Haščič, I., Popp, D., 2010. Renewable Energy Policies and Technological
Innovation: Evidence Based on Patent Counts. Environ Resource Econ 45, 133–155.
doi:10.1007/s10640-009-9309-1
Kenney, M., Hargadon, A., 2012. Misguided Policy? California Management Review 54,
118–139. doi:10.1525/cmr.2012.54.2.118
Konidari, P., Mavrakis, D., 2007. A multi-criteria evaluation method for climate change
mitigation
policy
instruments.
Energy
Policy
35,
6235–6257.
doi:10.1016/j.enpol.2007.07.007
Lesser, J., Su, X., 2008. Design of an economically efficient feed-in tariff structure for
renewable energy development. Energy Policy 36, 981–990.
Lüthi, S., Prässler, T., 2011. Analyzing policy support instruments and regulatory risk factors
for wind energy deployment—A developers’ perspective. Energy Policy 39, 4876–
4892. doi:10.1016/j.enpol.2011.06.029
Lüthi, S., Wüstenhagen, R., 2012. The price of policy risk — Empirical insights from choice
experiments with European photovoltaic project developers. Energy Economics 34,
1001–1011. doi:10.1016/j.eneco.2011.08.007
Marques, A.C., Fuinhas, J.A., 2011. Do energy efficiency measures promote the use of
renewable sources? Environmental Science & Policy 14, 471–481.
doi:10.1016/j.envsci.2011.02.001
Marques, A.C., Fuinhas, J.A., 2012a. Are public policies towards renewables successful?
Evidence from European countries. Renewable Energy 44, 109–118.
doi:10.1016/j.renene.2012.01.007
Marques, A.C., Fuinhas, J.A., 2012b. Is renewable energy effective in promoting growth?
Energy Policy 46, 434–442. doi:10.1016/j.enpol.2012.04.006
Marques, A.C., Fuinhas, J.A., Pires Manso, J.R., 2010. Motivations driving renewable energy
in European countries: A panel data approach. Energy Policy 38, 6877–6885.
22
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
Martin, S., Scott, J.T., 2000. The nature of innovation market failure and the design of public
support for private innovation. Research Policy 29, 437–447. doi:16/S00487333(99)00084-0
Masini, A., Menichetti, E., 2012. The impact of behavioural factors in the renewable energy
investment decision making process: Conceptual framework and empirical findings.
Energy Policy 40, 28–38. doi:10.1016/j.enpol.2010.06.062
Mathews, J.A., Kidney, S., Mallon, K., Hughes, M., 2010. Mobilizing private finance to drive
an energy industrial revolution. Energy Policy 38, 3263–3265.
Menanteau, P., Finon, D., Lamy, M.-L., 2003. Prices versus quantities: choosing policies for
promoting the development of renewable energy. Energy Policy 31, 799–812.
Mowery, D.C., Nelson, R.R., Martin, B.R., 2010. Technology policy and global warming:
Why new policy models are needed (or why putting new wine in old bottles won’t
work). Research Policy 39, 1011–1023.
Müller, S., Brown, A., Ölz, S., 2011. Renewable energy: policy considerations for deploying
renewables. International Energy Agency.
Nicolli, F., Vona, F., 2012. The Evolution of Renewable Energy Policy in OECD Countries:
Aggregate Indicators and Determinants (SSRN Scholarly Paper No. ID 2123094).
Social Science Research Network, Rochester, NY.
OECD, 2013. OECD factbook: economic, environmental and social statistics. OECD.
Popp, D., Hascic, I., Medhi, N., 2011. Technology and the diffusion of renewable energy.
Energy Economics 33, 648–662. doi:10.1016/j.eneco.2010.08.007
Quirion, P., 2010. Complying with the Kyoto Protocol under uncertainty: Taxes or tradable
permits? Energy Policy 38, 5166–5173.
Reed, W.R., Ye, H., 2011. Which panel data estimator should I use? Applied Economics 43,
985–1000. doi:10.1080/00036840802600087
Rogge, K.S., Hoffmann, V.H., 2010. The impact of the EU ETS on the sectoral innovation
system for power generation technologies - Findings for Germany. Energy Policy 38,
7639–7652.
Rogge, K.S., Schneider, M., Hoffmann, V.H., 2011. The innovation impact of the EU
Emission Trading System — Findings of company case studies in the German power
sector. Ecological Economics 70, 513–523.
Sandén, B.A., Azar, C., 2005. Near-term technology policies for long-term climate targets—
economy wide versus technology specific approaches. Energy Policy 33, 1557–1576.
doi:10.1016/j.enpol.2004.01.012
Smith, S., Swierzbinski, J., 2007. Assessing the performance of the UK Emissions Trading
Scheme. Environmental and Resource Economics 37.
Sukumar, S.., Shankar, M., Olama, M., Hadley, S., Protopopescu, V., Malinchik, S., Ives, B.,
2010. Quantifying state-policy incentives for the renewable energy investor. Energy
Conversion Congress and Exposition (ECCE), 2010 IEEE 386–391.
Szabó, S., Jäger-Waldau, A., 2008. More competition: Threat or chance for financing
renewable electricity? Energy Policy 36, 1436–1447.
Upreti, B.R., 2004. Conflict over biomass energy development in the United Kingdom: some
observations and lessons from England and Wales. Energy Policy 32, 785–800.
doi:10.1016/S0301-4215(02)00342-7
Veugelers, R., 2011. Europe’s clean technology investment challenge (Policy Contributions
No. 561). Bruegel.
Veugelers, R., 2012. Which policy instruments to induce clean innovating? Research Policy
41, 1770–1778. doi:10.1016/j.respol.2012.06.012
Vona, F., Nicolli, F., Nesta, L., 2012. Determinants of Renewable Energy Innovation:
environmental policies vs. market regulation. Documents de Travail de l’OFCE 5.
23
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
Wooldridge, J.M., Calhoun, J.W., Jung, B., Greber, S., Montgomery, R., 2009. Introductory
Econometrics, 4th ed. South-Western Cengage Learning, Mason, OH.
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
9
24
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
10 Appendices
10.1.
Definitions
Information and education: Policies and measures designed to increase knowledge,
awareness and training among relevant stakeholders or the general public. This can include
general information campaigns, targeted training programs for professionals and labeling
schemes that provide the purchaser with information on a product’s energy usage or emissions
performance.
Economic instruments: Policies and measures that stimulate certain activities, behaviors or
investments using financial supports and price signals to influence the market. These include
fiscal and financial policy instruments such as taxes, tax relief, grants or subsidies, feed-in
tariffs for renewable energy, and loans for the purchase or installation of certain goods and
services. They also include direct public funding and procurement rules, and market
mechanisms such as tradable permits.
Policy development and reform: Refers to steps in the ongoing process of developing,
supporting and implementing policies. This includes strategic plans that guide policy
development and the creation of specific bodies to support policy
Research, Development &Deployment (RD&D): Policies and measures aimed at
supporting technological advancement, through direct government investment, or facilitation
of investment, in technology research, development, demonstration and deployment activities.
Regulatory instruments: Covers a wide range of instruments with which a government
imposes targets, obligations and standards on actors requiring them to undertake specific
measures and/or report on specific information. Examples include energy performance
standards for appliances, equipment, and buildings; requirement for companies to manage
energy consumption, produce or purchase a certain amount of renewable energy or deliver
energy efficiency to customers; mandatory energy audits of industrial facilities; requirements
to monitor and report on greenhouse gas emissions or energy use.
Voluntary approaches: Refers to measures that are undertaken voluntarily either by public
agencies or by the private sector unilaterally, or by the two in a negotiated agreement.
Unilateral commitments are when entities set themselves environmental targets and
communicate successful compliance to their stakeholders. Public voluntary schemes invite
companies to meet specified environmental targets established by public authorities.
Negotiated agreements set environmental targets agreed between a government and a private
25
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
sector entity, and may require reporting information on energy use to the government, being
subject to audits, and undertaking measures to reduce energy use.
10.1.
Case selection
Table A.1 here
10.2.
Variable Selection
Economic instruments/ Direct investments: We excluded the variable Procurement rules as
there is no significant public market for RE technologies. Further, we excluded RD&D
funding as only five policies falling into this category have been implemented.
Economic instruments/ Fiscal/financial incentives: Here we excluded only user charges due to
data availability.
Economic instruments/ Market-based instruments: The IEA policy scheme included a white
certificate scheme (focusing on energy efficiency). Although research shows that these policy
measures might accelerate the diffusion of RE technologies (Marques and Fuinhas, 2011), we
found no causal link to investor behavior.
Information and Education: Here, Information provision, Advice/Aid in Implementation as
well as performance labels and professional training and qualification have been excluded,
firstly due to missing investor reference and secondly due to data availability.
Regulatory Instruments: Auditing, building codes and standards as well as monitory have
been removed from the list of explanatory variables as to our knowledge these might have no
effect on investments in RE capacity by institutional investors.
Research, Development and Deployment (RD&D): measures have been removed as these
measures target individual firms, which are not our unit of analysis.
Voluntary Approaches: These variables have been excluded as they target corporations, rather
than investors.
10.3.
Summary statistics
Table A.2 & A.3here
26
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
Category
Instruments
Variable Abbreviation
Literature
Economic Instruments - Direct investment
Funds to sub-national governments
EI_DI_FSGjk
(Ragwitz et al. 2008)
Infrastructure investments
EI_DI_IIjk
(Bergek et al. 2013)
Feed-in tariffs/premiums
EI_FI_FIjk
(Jenner et al. 2013)
Grants and subsidies
EI_FI_GSjk
(Bergek et al. 2013)
Loans
EI_FI_Ljk
(Bergek et al. 2013)
Tax relief
EI_FI_TRjk
(Cansino et al. 2010; Quirion 2010)
Taxes
EI_FI_Tjk
(Cansino et al. 2010; Quirion 2010)
GHG emissions allowances
EI_MI_GAjk
(Rogge et al. 2011; Rogge & Hoffmann 2010)
Green certificates
EI_MI_GCjk
(Rogge et al. 2011; Rogge & Hoffmann 2010)
Institutional creation
PS_ICjk
(Wieczorek & Hekkert 2012)
Strategic planning
PS_SPjk
(Lüthi & Wüstenhagen 2012)
Codes and standards
RI_CSjk
(Carley 2009)
Obligation schemes
RI_OSjk
(Butler & Neuhoff 2008),
Other mandatory requirements
RI_MRjk
(Butler & Neuhoff 2008)
Institutional creation
PS_ICjk
(Wieczorek & Hekkert 2012)
Strategic planning
PS_SPjk
(Lüthi & Wüstenhagen 2012)
Economic Instruments – Fiscal/financial incentives
Economic Instruments – Market-based instruments
Policy Support
Regulatory Instruments
Policy Support
Table 7: Policy Scheme (adapted from IEA Policies and Measures)
27
copy available at: https://ssrn.com/abstract=2423310
Electronic copy available at:Electronic
https://ssrn.com/abstract=2423310
Random effects
Fixed effects
Multiple RE
Solar
Wind
Biomass
Multiple RE
Solar
Wind
Biomass
Modified Wald test
-
-
-
-
236.84***
91.24***
1052.72***
141.05***
Pesaran’s test
19.550***
5.544***
16.264***
8.702***
15.020***
4.127***
10.148***
4.611***
Frees’ test
2.997***
1.029***
2.491***
0.900***
2.344***
0.216
1.872***
0.631**
Notes: The Modified Wald Test Chi2 distribution and tests the null hypothesis:
, for c = 1,.,N; Pesaran and Frees’ tests test the null hypothesis of crosssection
independence; Pesaran’s test is a parametric testing procedure and follows a standard normal distribution; Frees’ test uses Frees’ Q-distribution; xtcsd command was used (Hoyos
& Sarafidis 2006). ***, *, denote 1 and 10% significance level, respectively.
Table 2: Specification tests
28
copy available at: https://ssrn.com/abstract=2423310
Electronic copy available at:Electronic
https://ssrn.com/abstract=2423310
PCSE
Independent
variables
EI_DI_FSGjk
EI_DI_IIjk
EI_FI_FIjk
Random effects
No autocorrelation
AR(1)
CSE
AR(1)
(I)
(II)
(III)
(IV)
Coeff.
SD
-.6516576*
.39985
.1480486
.2566804
Coeff.
SD
Coeff.
SD
Coeff.
SD
-.8630866 .5901286
-1.12362
.7074356 -.6923118 .6345557
.2910476
.1199704
.5227986 -.013469
.3534862
.5025643
.6871031*** .2806932 .3448074** .4361374 .6460938* .3767379 .4565576 .3533344
EI_FI_GSjk
.3262012
EI_FI_Ljk
-.3590348 .2642972
.0480538
.3678096 -.3814668 .5468806 -.1682956 .5119435
EI_FI_TRjk
.4008434** .1780084
.3064977
.3259957
EI_FI_Tjk
-.8995866* .4812785 -.7008466 .5409997 -1.428411** .5878148 -.7806197 .5767547
EI_MI_GAjk
1.476**
.2525002 .7721907** .3423242 .983853*** .3595374 .6290736** .3265612
.3614985
.4265634 .2877263 .3978518
.6289598 1.422642* .7595094 2.060666*** .7979844 1.366636* .793817
EI_MI_GCjk
-.0147581 .3439044
.3990277
PS_ICjk
-.407161
.2785661
.0366566
.4038071 -.5492414 .4767925 -.2247933
PS_SPjk
.7037612* .3738054
.4092362
.4437024 1.199293*** .4226628 .5398187 .4043298
RI_CSjk
.4475644*** .1712941
.1326882
.3174074 .8248669* .4792686
.2805152
.4863
.36328
.153039
.5558723 .1682016 .5317527
.461503
.474983
.4229809
.359739
.4520459
RI_OSjk
.2783923
.2619837
.3480405
.4777855
RI_MRjk
.5228704
.3580947 .8174922** .4244155
.7293992
.4620257 .5434511 .4384469
c_TECjk
.6369263
.2235642
.5377107
.3651499 .6155304**
c_CIjk
-.0896215 .6683637
1.03767
.7754292 -.6598328 1.594438 -.2695809 1.262449
Control
variables
c_LIRjk
.6704916***
c_SPjk
c_GDPjk
_cons
Observations
R2
Wald
.1896642
.6700271***
.27879 .5122237** .2315727
.185451 -.7077997** .3075835
.6700027**
.2868721
1.841445*** .5574079 1.547328** .6609779 1.840681*** .3614505 1.55531*** .3780566
-.0211656 .0781152 -.0409273 .0761332 -.0361108 .0843065 -.0087939 .0721356
-
-
-7.118729** 3.28976
-5.843901 3.640319
330
330
330
330
0.3811
0.3173
0.3632
0.3741
258.81***
222.09***
205.74***
106.10***
7.542619***
2.787254
5.936144**
2.540102
Notes: The Wald test distribution and tests the null hypothesis of non-significance of all coefficients of
explanatory variables; panel corrected standard errors are reported in brackets. ***, **, *, denote significance at
1, 5 and 10% significance levels, respectively; AR(1) - first-order autoregressive error; CSE – Conventional
Standard Errors. Estimations include both country and time dummies.
Table 3: Results for Multiple RE
29
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
PCSE
Independent variables
Random effects
No autocorrelation
AR(1)
CSE
AR(1)
(V)
(VI)
(VII)
(VIII)
Coeff.
SD
Coeff.
SD
Coeff.
SD
Coeff.
SD
-1.430191***
.452557
-.4345027
.6578232
-1.430191***
.4693717
-.7560549
.5740057
EI_DI_IIjk
.205142
.4027607
.1930198
.4221444
.205142
.5622184
-.2319487
.5570207
EI_FI_FIjk
1.183469***
.2614633
.9968022***
.2765806
1.183469***
.3108172
1.064136***
.3318779
EI_FI_GSjk
.6075715***
.2286481
.153594
.2578642
.6075715**
.2519058
.3574473
.296831
EI_FI_Ljk
-.9083116***
.3042927
-.3219862
.3829625
-.9083116**
.428827
-.5934808
.4710678
-.0101943
.2788787
.5962874
.3225234
-.0101943
.3439999
.3352309
.3777232
EI_FI_Tjk
.9986014**
.452132*
.1905455
.4228552
.9986014*
.5788794
.5929351
.6261464
EI_MI_GAjk
-1.174446**
.5967956*
-1.207017**
.521431
-1.174446
.8048623
-.7284159
.9053043
EI_MI_GCjk
-2.274715***
.4114937
-1.244204**
.5135853
-2.274715***
.4830896
-1.649163***
.5472138
PS_ICjk
-1.513661***
.3755918
-.1851111
.3821114
-1.513661***
.4327728
-.6471036
.4550913
PS_SPjk
2.347876***
.3042435
1.099465***
.3478684
2.347876***
.3444409
1.72752***
.3965018
RI_CSjk
.5437917**
.2699895
-.445115
.4370589
.5437917
.3625386
.0065249
.4092227
RI_OSjk
.1625115
.3058051
-.1337805
.3184731
.1625115
.4126382
.064317
.4607792
RI_MRjk
.7667914**
.3652028
.6711904
.4245183
.7667914
.4680079
.5862809
.4653205
c_TECjk
.9454781
.7403308
.502266
.8728845
.9454781
.7233382
.19094
.7615019
c_CIjk
-6.009266 2.240291***
-4.545095
2.778778
-6.009266**
2.582556
-1.444087
2.93114
c_LIRjk
-.5395957
.751136
.4752815
.7520795
-.5395957
.6670382
-.1201582
.6801035
c_SPjk
-.2380471
.5872644
-.0602443
.6734641
-.2380471
.4560057
-.2061717
.4829867
c_GDPjk
-.8238478
.7017687
-.1409385
.7989894
-.8238478
.6983512
.1140484
.732125
_cons
21.59223
17.18385
2.75846
18.93736
21.59223
17.01726
-2.987067
17.79872
EI_DI_FSGjk
EI_FI_TRjk
Control variables
30
copy available at: https://ssrn.com/abstract=2423310
Electronic copy available at:Electronic
https://ssrn.com/abstract=2423310
Observations
R2
Wald
176
176
176
176
0.4938
0.1043
0.4938
0.4668
161.69***
46.59***
152.19***
70.93***
Notes: The Wald test distribution and tests the null hypothesis of non-significance of all coefficients of explanatory variables; panel corrected standard errors are reported in
brackets. ***, **, *, denote significance at 1, 5 and 10% significance levels, respectively; AR(1) - first-order autoregressive error; CSE – Conventional Standard Errors.
Estimations include both country and time dummies.
Table 4: Results for Solar
31
copy available at: https://ssrn.com/abstract=2423310
Electronic copy available at:Electronic
https://ssrn.com/abstract=2423310
PCSE
Independent variables
Random effects
No autocorrelation
AR(1)
CSE
AR(1)
(XI)
(VII)
(III)
(IV)
Coeff.
SD
Coeff.
SD
Coeff.
SD
Coeff.
SD
EI_DI_FSGjk
-.3757999
.4309306
-1.115682***
.5215933 -1.117909
.7042944
-.490034
.5922703
EI_DI_IIjk
.3654253
.2427625
.5124566
.3962674 .5206531
.4555222
.2274922
.4362141
EI_FI_FIjk
.7034015*** .2594473
.3466685
.3705958 .6145793*
.3562496
.5717061*
.3242466
EI_FI_GSjk
.1299914
.2458559
.3584737
.2852773 .8257291** .3584267
.4697756
.3197792
EI_FI_Ljk
.3021586
.2754532
.9553349**
.3801406 .6636585
.5289154
.5423989
.4828976
EI_FI_TRjk
.1957176
.2003498
.0637804
.3724269 -.0227203
.4155139
.1331156
.36574
EI_FI_Tjk
-.5103487
.3414624
-.2289026
.4386519 -.5030176
.5816102
-.3150329
.5275276
EI_MI_GAjk
1.5911***
.411295
1.196165**
.6036883 1.460892*
.8265885
1.289026*
.7588826
EI_MI_GCjk
-.0326347
.3549096
-.1049938
.4631245 .2383117
.5159596
.2998673
.4947087
PS_ICjk
.1335157
.2622394
.079518
.4187491 .0165175
.4525562
.2484949
.4156641
PS_SPjk
.1607409
.4327854
.3294506
.4911224 .6631261
.4301623
.1116961
.4089962
RI_CSjk
.6264673** .1855897
.5738682*
.3321465 1.031217** .4704546
.6056064
.4105352
RI_OSjk
-.0883198
.3071059
.2490448
.3150301 .2585585
.4727774
.0770355
.4299847
RI_MRjk
.2113357
.3181797
.2007613
.3819612 .0357223
.4277984
.1372273
.4099641
.7544588**
.3608242 .8300047
.2988083
.6483789*** .2371395
Control variables
c_TECjk
.768095*** .223812
32
copy available at: https://ssrn.com/abstract=2423310
Electronic copy available at:Electronic
https://ssrn.com/abstract=2423310
c_CIjk
.9649845
.6596128
1.941291**
.9370116 -.3786435
c_LIRjk
-.748926*** .1827553
-.662133***
.187107
-.7451924** .311754
-.7195487** .286342
c_SPjk
2.103777*** .5937315
1.579907***
.616685
1.916384*** .368951
1.635179*** .3850877
c_GDPjk
-.0465362
_cons
-8.327666*** 3.172761
Observations
R2
Wald
.0787399
-.0578715
.0769572 -.047532
-6.743841**
1.798071
.0829624
3.474884 -8.641509*** 2.79598
.6928736
-.0240795
1.354673
.0710109
-6.851058*** 2.512694
319
319
319
319
0.3977
0.2739
0.3634
0.3868
398.37***
208.87***
198.02***
108.58***
Notes: The Wald test distribution and tests the null hypothesis of non-significance of all coefficients of explanatory variables; panel corrected standard errors are reported in
brackets. ***, **, *, denote significance at 1, 5 and 10% significance levels, respectively; AR(1) - first-order autoregressive error; CSE – Conventional Standard Errors.
Estimations include both country and time dummies.
Table 5: Results for Wind
33
copy available at: https://ssrn.com/abstract=2423310
Electronic copy available at:Electronic
https://ssrn.com/abstract=2423310
PCSE
Independent variables
Random effects
No autocorrelation
AR(1)
CSE
AR(1)
(XVI)
(XVII)
(XVIII)
(XIX)
Coeff.
SD
Coeff.
SD
Coeff.
SD
Coeff.
SD
EI_DI_FSGjk
1.506121*** .392019
1.219264** .5174145
1.506121*** .5247834
1.468757** .5802493
EI_DI_IIjk
-1.196352** .3852509
-1.040943** .4669735
-1.196352*** .4578357
-1.214129** .4882678
EI_FI_FIjk
.0855131
.3397968
.2857104
.0855131
.0383056
EI_FI_GSjk
.6062906** .2448567
.5218678*
.3076593
.6062906*** .2486713
.6356551** .274053
EI_FI_Ljk
-.1467847
.3211142
-.0673767
.3596578
-.1467847
.3941907
-.0481403
.422529
EI_FI_TRjk
.0978145
.182495
-.055498
.2731093
.0978145
.2681086
.0529121
.2944414
EI_FI_Tjk
-.2000953
.293675
-.1988765
.414443
-.2000953
.3874982
-.2509206
.4246686
EI_MI_GAjk
.9708445*
.4696504
.9970731
.6529303
.9708445** .5348302
1.000802*
.591107
EI_MI_GCjk
.0611662
.3388353
.1831551
.4952602
.0611662
.1213299
.4069643
PS_ICjk
.8325421** .3670397
.7382939*
.4009378
.8325421** .3553792
.7920043** .3846703
PS_SPjk
-.44445*
.2364515
-.267019
.371566
-.44445
.3023074
-.4376296
.330444
RI_CSjk
-.4141917
.374794
-.3366353
.4092963
-.4141917
.334913
-.4235416
.3691855
RI_OSjk
-.012409
.2674505
-.0289097
.3806128
-.012409
.3380002
.0439453
.3685454
RI_MRjk
.2794535
.2654414
.1784495
.3318189
.2794535
.3128969
.3191621
.3401591
.2354532
.2630173
.370753
.2877502
Control variables
34
copy available at: https://ssrn.com/abstract=2423310
Electronic copy available at:Electronic
https://ssrn.com/abstract=2423310
c_TECjk
-1.969388*** .6450672
-1.69437*
.6629631
-1.969388*** .5245612
-1.929471*** .5697671
c_CIjk
4.368859*** 1.560729
3.975855** 2.147203
4.368859*** 1.601548
4.250155** 1.752275
c_LIRjk
-.565306**
.2911711
-.5420741*
.3153329
-.565306*
.2952814
-.5717118*
.3110166
c_SPjk
.6105773
.4172689
.6457221
.4520086
.6105773*
.3712451
.6356115
.3993678
c_GDPjk
2.159608*** .5804663
1.928434*** .62993
2.159608*** .5019944
2.083474*** .5420153
_cons
-53.24435*** 13.25234
-48.23348*** 15.25634
-53.24435*** 11.51497
-51.45514*** 12.44432
Observations
R2
Wald
220
220
220
220
0.3825
0.2365
0.3825
0.3818
169.40***
61.50***
123.86***
98.50***
Notes: The Wald test distribution and tests the null hypothesis of non-significance of all coefficients of explanatory variables; panel corrected standard errors are reported in
brackets. ***, **, *, denote significance at 1, 5 and 10% significance levels, respectively; AR(1) - first-order autoregressive error; CSE – Conventional Standard Errors.
Estimations include both country and time dummies.
Table 6: Results for Biomass
35
copy available at: https://ssrn.com/abstract=2423310
Electronic copy available at:Electronic
https://ssrn.com/abstract=2423310
Countries included in Multiple
Countries included in Solar
Countries included in Wind
Countries included Biomass
RE
Australia
Austria
Belgium
Canada
Chile
Czech Republic
Denmark
Estonia
Finland
France
Germany
Greece
Hungary
Ireland
Italy
Japan
Korea, Rep.
Mexico
Netherlands
New Zealand
Norway
Poland
Portugal
Slovak Republic
Spain
Sweden
Switzerland
AUS
AUT
BEL
CAN
CHL
CZE
DNK
EST
FIN
FRA
DEU
GRC
HUN
IRL
ITA
JPN
KOR
MEX
NLD
NZL
NOR
POL
PRT
SVK
ESP
SWE
CHE
Australia
Belgium
Canada
Czech Republic
France
Germany
Greece
Italy
Japan
Korea, Rep.
Netherlands
Portugal
Slovak Republic
Spain
Turkey
United
Kingdom
United States
AUS
BEL
CAN
CZE
FRA
DEU
GRC
ITA
JPN
KOR
NLD
PRT
SVK
ESP
TUR
GBR
USA
Australia
Austria
Belgium
Canada
Chile
Czech Republic
Denmark
Estonia
Finland
France
Germany
Greece
Hungary
Ireland
Italy
Japan
Korea, Rep.
Mexico
Netherlands
New Zealand
Norway
Poland
Portugal
Spain
Sweden
Switzerland
Turkey
AUS
AUT
BEL
CAN
CHL
CZE
DNK
EST
FIN
FRA
DEU
GRC
HUN
IRL
ITA
JPN
KOR
MEX
NLD
NZL
NOR
POL
PRT
ESP
SWE
CHE
TUR
Australia
Austria
Belgium
Canada
Chile
Czech Republic
Finland
France
Germany
Ireland
Italy
Japan
Mexico
Netherlands
Norway
Poland
Spain
Sweden
United
Kingdom
United States
AUS
AUT
BEL
CAN
CHL
CZE
FIN
FRA
DEU
IRL
ITA
JPN
MEX
NLD
NOR
POL
ESP
SWE
GBR
USA
36
copy available at: https://ssrn.com/abstract=2423310
Electronic copy available at:Electronic
https://ssrn.com/abstract=2423310
Turkey
United
Kingdom
United States
TUR
GBR
USA
United
Kingdom
United States
GBR
USA
Table A.1: Country selection
37
copy available at: https://ssrn.com/abstract=2423310
Electronic copy available at:Electronic
https://ssrn.com/abstract=2423310
Variable
Obs. Mean
Std. Dev. Min
BNEF_Capacity_ALL
360 3.607634 2.814445 0
9.428982
BNEF_Capacity_Biomass 240 2.251529 2.091209 0
6.800838
BNEF_Capacity_Solar
192 1.759968 2.268786 0
8.557951
BNEF_Capacity_Wind
348 3.337217 2.806457 0
9.168638
EI_DI_FSG
360 .1603185 .3238168 0
1.386294
EI_DI_II
360 .1676961 .3603024 0
1.609438
EI_FI_FI
360 .4791029 .5231468 0
1.94591
EI_FI_GS
360 .88213
.639799 0
2.302585
EI_FI_L
360 .2467513 .3814493 0
1.609438
EI_FI_TR
360 .4044712 .5213873 0
1.791759
EI_FI_T
360 .2651274 .3723063 0
1.098612
EI_MI_GA
360 .1019019 .2921595 0
1.386294
EI_MI_GC
360 .18623
1.609438
PS_IC
360 .4748967 .4910026 0
1.94591
PS_SP
360 .8731863 .5641254 0
2.079442
RI_CS
360 .5062903 .5155385 0
2.397895
RI_OS
360 .6436002 .5068814 0
2.079442
RI_MR
360 .4636214 .5625358 0
2.302585
c_TEC
360 3.501761 1.181402 1.226126 6.960218
c_CI
360 .3411723 .1685374 .0972628 .991836
c_LIR
360 1.517852 .5975912 0
c_SP
360 4.488321 .4851809 2.895912 5.699774
c_GDP
360 26.72129 1.999405 0
.3680391 0
Max
2.818398
30.33849
Table A.2: Summary statistics
38
Electroniccopy
copy available
available at:
Electronic
at: https://ssrn.com/abstract=2423310
https://ssrn.com/abstract=2423310
Variables
IC
EI_DI_~G EI_DI_II EI_FI_FI
EI_FI_
EI_FI_T
EI_MI_ EI_MI_
EI_FI_L
EI_FI_T
GS
R
GA
GC
PS_IC
PS_SP
RI_CS
RI_OS
RI_MR c_TEC
c_CI
c_
LIR
c_SP
IC
1.0000
EI_DI_FSG
0.1374 1.0000
EI_DI_II
0.1074 0.0576
1.0000
EI_FI_FI
0.1464 -0.1385
0.0967
1.0000
EI_FI_GS
0.2626 0.2780
0.3621
0.2118
1.0000
EI_FI_L
0.2259 0.2470
0.2334
0.1085
0.1724
1.0000
EI_FI_TR
0.3295 0.2282
0.2600
0.0871
0.2226
0.3433
1.0000
EI_FI_T
0.0924 0.1887
0.0245
-0.0096
0.0681
0.2758
0.0734
1.0000
EI_MI_GA
0.1968 0.0390
-0.0462
-0.2560
0.0054
0.2204
0.0831
0.5718
1.0000
EI_MI_GC
0.1779 -0.0904
0.0222
-0.1668
-0.0536 0.1898
0.1947
0.2310
0.2665
1.0000
PS_IC
0.1821 0.0842
0.2683
0.2234
0.3447
0.3205
-0.0057 0.1888
0.1434
-0.1353 1.0000
PS_SP
0.2919 0.2644
0.0530
-0.0299
0.2617
0.3440
-0.0084 0.3384
0.4124
0.3706
0.3970
1.0000
RI_CS
0.3240 0.3432
0.2854
0.1014
0.3798
0.2549
0.4373
0.2329
0.0516
0.2243
0.2739
0.3023
1.0000
RI_OS
0.3632 0.1689
0.0843
0.0829
0.2272
0.1701
0.3035
0.0090
0.1967
-0.0018 0.4586
0.1505
0.3443
1.0000
RI_MR
0.3292 0.3978
0.2592
-0.1101
0.0695
0.4104
0.4534
0.1751
0.0819
0.1633
0.0349
0.1478
0.4009
0.4023
1.0000
c_TEC
0.3988 0.3866
0.1253
-0.1240
0.3047
0.3059
0.3784
0.0415
0.1934
0.0639
0.1899
0.1259
0.3005
0.4479
0.5330
c_CI
-0.1768
0.1746
0.0773
-0.0436
-0.1826 -0.1740 -0.1024 -0.2211 -0.1904 -0.1298 -0.1899 -0.3430 -0.1639 -0.0001 0.0142
-0.2637 1.0000
c_LIR
0.0931
0.0063
0.1709
0.0753
0.1560
0.1596
0.1495
0.2521
0.0958
0.1347
0.1591
0.1808
0.2194
0.1160
0.1065
0.0492
-0.2801 1.0000
c_SP
0.3140 0.0830
0.0051
0.2664
0.1889
0.2100
0.2005
0.1923
0.0466
0.2229
0.2611
0.3705
0.2376
0.2189
0.0982
0.0666
-0.3268 0.2384
1.0000
c_GDP
0.3748 0.1673
0.1021
0.0198
0.2299
0.2079
0.3190
0.0929
0.1550
0.1573
0.0953
0.0992
0.3081
0.3267
0.3887
0.6997
-0.3101 0.0665
0.1594
1.0000
Table A.3: Pairwise correlation coefficients (Multiple RE)
39
copy available at: https://ssrn.com/abstract=2423310
Electronic copy available at:Electronic
https://ssrn.com/abstract=2423310
c_
GDP
1.0000
Figure 1: Aggregated sources of finance in the cleantech sector during the last 10 years in USDm (Bloomberg New Energy Finance 2013)
40
copy available at: https://ssrn.com/abstract=2423310
Electronic copy available at:Electronic
https://ssrn.com/abstract=2423310
Figure 2: Solar energy investments in Italy
Figure 3: Wind energy investments in UK
41
copy available at: https://ssrn.com/abstract=2423310
Electronic copy available at:Electronic
https://ssrn.com/abstract=2423310
Figure 4: Model
42
copy available at: https://ssrn.com/abstract=2423310
Electronic copy available at:Electronic
https://ssrn.com/abstract=2423310
Figure 5: Scatterplot of log(Investments in multiple RE) and log(Policy Measures)
43
copy available at: https://ssrn.com/abstract=2423310
Electronic copy available at:Electronic
https://ssrn.com/abstract=2423310
Figure 6: Overview of results
44
copy available at: https://ssrn.com/abstract=2423310
Electronic copy available at:Electronic
https://ssrn.com/abstract=2423310