What Drives States to Support Renewable Energy?
Steffen Jenner*a,b, Gabriel Chana, Rolf Frankenbergerb, and Mathias Gabelb
Why do states support electricity generation from renewable energy
sources? Lyon/ Yin (2010), Chandler (2009), and Huang et al. (2007) have answered this question for the adoption of renewable portfolio standards (RPS) at
the U.S. state level. This article supplements their work by testing the core hypotheses on the EU27 sample between 1990 and 2010. Furthermore, the article asks
why the majority of EU states rely on feed-in-tariffs (FIT). The study conducts
logistic time series cross-section regression analyses that run on a hazard model.
Evidence in support of private interest theory and public interest theory
is provided. (a) The existence of a solar energy association increases the probability of a state to adopt regulation. (b) Solar radiation, and (c) the unemployment
rate also increase the odds. (d) Electricity market concentration decreases the
probability of transition.
Keywords: Energy policy, Renewable energy, Electricity, Feed-in-tariff, Hazard
model, Public Choice
http://dx.doi.org/10.5547/01956574.33.2.1
1. INTRODUCTION
Sky rocketing electricity prices, sky fogging greenhouse gas emissions,
looming ‘peak oil’, trade dependencies, emerging economies eating up a rapidly
increasing share of global energy demand, and ever depleting global energy supply have put conventional energy sources under pressure. To make things worse,
in March 2011, the Japan fallout gave an unexpected blow to nuclear power, a
conventional zero-emission power source.
*
a
b
Corresponding author: E-Mail: jenner@fas.harvard.edu
Harvard University
The University of Tübingen.
We would like to thank Professor Richard B. Freeman, from Harvard University, for his time and
suggestions that substantially improved this article. In addition, we appreciated the helpful comments
from two anonymous referees. Remaining errors are ours alone.
The Energy Journal, Vol. 33, No. 2. Copyright 䉷 2012 by the IAEE. All rights reserved.
1
2 / The Energy Journal
Table 1: Years of Policy Enactment in Europe
DE
IT
DK
LU
1990
1992
1993
GR
ES
AT
FR
IT
PT
1994
1998
2001
BE
CZ
HU
LT
GB
BG
EE
NL
SE
MT
SI
IE
RO
SK
BG
CY
2002
2003
2004
2005
2006
Bold and standard letters indicate a quota system and a feed-in-tariff, respectively. Source: Res-legal
(2010) and REN21 (2009).
One would expect international alliances, such as the United Nations
(UN), to counter challenges that threaten humanity on a global scale. But they
struggle in their effort to make agreements binding. It is the state level’s task
instead. Even if a club of countries, the European Union (EU) for instance,
achieves to negotiate collective targets, e.g. the Directive 2001/77/EC on RESE, the Directive 2003/30/EC on biofuels or the Directive 2009/28/EC on both,
the state is urged to do the work on the ground.
With conventional power technologies scrutinized or their resources running out, state policymakers promote investment in electricity generation from
renewable energy sources (RES-E). As can be reviewed in Table 1 some states,
such as Germany and Italy, have been front runners implementing RES-E policies
in the early 90th. Others have been more reluctant, to put it mildly.
Investigating this gap of enthusiasm, the article supplements similar research on the U.S. states. Answering the question of ‘what drives states to adopt
regulation that supports RES-E?’ we run logistic time series cross-section regression analyses. Section two reviews the literature on the puzzle. Section three
presents a proportional hazard model. Section four elaborates the variables. Section five lists the findings which are discussed in section six.
2. LITERATURE REVIEW
Academia has tackled the question before. Most empirical research applied different types of process tracing to the question why political RES-E support evolved.1 Besides these small-n case studies, a few large-n studies exist as
well. Huang et al. (2007) run a logistic model on the U.S. states. They find that
Republican party dominance decreases the likelihood of state RES-E support
while the educational level of citizens correlates positively. Chandler (2009) specializes in inter-state learning. He finds that neighboring effects exist in U.S. state
1. See Heymann (1998) on wind power in the U.S., Germany and Denmark; Lauber and Mez
(2004) on RES-E in Germany; most recently Laird and Stefes (2009)—and Keller´s (2010) response—
on the paths of German and U.S. energy policy; and Rabe (2004) on U.S. states’ role in federal
climate policy making.
Copyright 䉷 2012 by the IAEE. All rights reserved.
What Drives States to Support Renewable Energy? / 3
electricity policy. Positive coefficients indicate that RPS systems spread across
state borders. Knittel’s (2006) regressions reveal a link between conventional
electricity market regulation and strong industrial interests. Lyon and Yin (2010)
find significant correlations between private interest as well as party indicators
and the odds of RPS adoption. According to their study, the intensity of solar
radiation increases the probability of policy adoption while the unemployment
rate has a negative impact. Further findings are shown in Table 3 as part of the
discussion section.
Since none of the quantitative studies we know run their regressions on
the EU27 sample, this article contributes to the discourse by testing the common
hypotheses on the EU member countries. Choosing the state level for analysis
makes sense on both continents. In the U.S., the Obama Administration is only
slowly moving forward on introducing federal legislation to support RES-E. In
the EU, the European Parliament and the European Commission urge their member countries to foster renewable energy exploitation by setting targets as mentioned above. However, the EU did not employ a policy mechanism to attract
investment in electric systems that run on renewable energy sources. To put it in
a nutshell, in the U.S. as well as in the EU, RES-E is a matter of state-level
legislation (Joskow, 2009).
The second contribution lies in extending the pool of policies from RPS
to feed-in-tariffs. Prior studies distinguish between states that apply a RPS and
the rest of the states. Therefore, they mesh states that do not support RES-E with
states that support RES-E but do not employ a RPS. Although, a 36 state majority
solely relies on the RPS, there is a 10 state minority2 that also introduced a FIT
in different years. And there are two more states3 that do not employ a RPS but
a FIT (DSIRE, 2010). It is controversial to merge these 12 states with the 14
states minority4 that neither support RES-E by means of a RPS nor by a FIT.
Of course, RPS and FIT differ with regard to the regulatory design. The
RPS is a mode of quantity regulation that stimulates investment in RES-E by a
command and control type of policy towards the utilities. It promotes RES-E
technologies that come with a low cost since utilities can choose any available
RES-E technology to meet their quota requirement. In contrast, the FIT is a mode
of price regulation that stimulates investment by giving fixed price incentives to
producers. It applies technology specific subsidies in order to equalize attractiveness for investment among different RES-E technologies. Therefore, investors
can expect (more or less) the same return of investment regardless of their actual
technology specific cost of electricity generation.5 A FIT challenges the utilities’
2. Year of FIT enactment: 1997: MA, MN; 2004: PA, NJ; 2005: WA; 2006: CA; 2007: MD, SC;
2009: ME, OR, HI, VT (DSIRE, 2010).
3. 2004: NJ, SC (DSIRE, 2010).
4. AL, AK, AR, FL, GA, ID, IN, KY, LA, MS, NE, OK, TN, WY (DSIRE, 2010).
5. See Menanteau et al. (2003) for further details on price and quantity driven RES-E policy tools.
In addition, see Couture and Gagnon (2010) for a comprehensive depiction of different types of FIT
systems.
Copyright 䉷 2012 by the IAEE. All rights reserved.
4 / The Energy Journal
grip on the production park because it enables new producers to feed electricity
into the grid. Thus, many renewable power plants with little capacity can substitute the mostly large power plants operated by the energy utilities.
Despite these differences in design, there are two essential similarities.
First—and this is important for the consumer side—the end user pays the bill that
may increase because utilities forward additional costs to the end user. Second—
and this is important for the kind of research—both policies share the goal of
promoting investment in RES-E capacities. This aspect matters since we are asking why (not how) policymakers implement RES-E support schemes. Therefore,
we should sample our cases by their policymakers’ intention not the type of policy
tool they use. Results might change if the U.S. studies assembled their data according to this criterion. The reason is that the year of enactment would change
in countries that apply regulation regardless of the regulatory design. The change
in time is crucial to the regression coefficients since the studies do not conduct a
cross-section analysis but a time series cross-section analysis. New Jersey and
South Carolina would switch groups from non-regulation to regulation since they
apply a state level FIT but do not employ a RPS. Minnesota and Washington
would change from 2007 and 2006 to 1994 and 2005, respectively, because their
FIT systems have been enacted prior to the RPS implementation (DSIRE, 2010).
As a consequence, the study at hand runs regressions with a combined
dependent variable, putting the implementation of feed-in-tariffs and quotas into
one basket. Nonetheless, the discussion section also provides findings from models that test dichotomous variables of feed-in-tariffs and quotas separately.
3. THE PROPORTIONAL HAZARD MODEL
In formal terms, this article’s aim is to ask what drives a country c to
adopt regulation in a certain year t . Therefore, the dependent variable is a binary
code distinguishing between regulation c(t)⳱1 and non-regulation c(t)⳱0. The
proportional hazard model tests the impact of the independent variables X on the
relative odds k of the conditional probability P1⳱P(t,X) of a country to adopt
regulation in a certain year, given the country did not adopt such regulation before:
P0⳱1– P(t,X). The relative odds k of a country to adopt regulation for the first
time are expressed by P1/P0.
k(t)⳱
P(t,X)
1– P(t,X)
(1)
If X is assumed to have no impact, we receive the relative odds k0 of
the conditional baseline probability.
k0(t)⳱
P0(t)
1– P0(t)
for X⳱0
(2)
The baseline is an estimate of the mean values of the independent metrics. According to Kiefer’s (1988; 1989) widely used proportional hazard model,
Copyright 䉷 2012 by the IAEE. All rights reserved.
What Drives States to Support Renewable Energy? / 5
the relative odds k depend on a product of the conditional baseline probability
k0(t) and an individual scaling factor ␾(X,b).
k(t,X,b,k0)⳱k0(t) ␾(X,b)
(3)
Please note that t does not impact the individual scaling factor ␾(X,b)
but the baseline hazard k0(t) . Lyon and Yin (2010) as well as this article proceed
by replacing the individual scaling factor by the specification (4) that is often
used in labor economics and other research that work with duration data.
␾(X,b)⳱exp(X⬘b)
(4)
Applying (1), (2) and (4) to the proportional hazard model produces the
relative odds of transition (5). The logit is shown in (6).
P(t,X)
P (t)
⳱ 0
1– P(t,X) 1– P0(t)
exp(X⬘b)
logit(P(t,X))⳱logit(P0(t))ⳭXb
(5)
(6)
Deriving X yields the first order condition of the proportional hazard
model including the specification case.
d logk(t,X,b,k0)
⳱b
dX
(7)
According to (7), a change in X affects k by the constant term of b. In
other words, a one unit change of an independent variable alters the probability
of a state to transit to regulation in a certain year by b.
4. INDEPENDENT VARIABLES
There are various ways to explain the puzzle. Reviewing the literature
mentioned above, we distinguish between public interest theory, private interest
theory and a group of controls.
Private interest theory assumes social groups that are affected by regulation to mobilize in order to influence policymaking according to their preferences. In turn, the policymaker responds to the efforts of the groups. Peltzman
(1976) introduced the focus on interest groups as an explanation for the adoption
of regulation. Becker’s (1983) model of interest group competition relates a
group’s lobbying benefits to its lobbying ability. As Knittel (2006) puts it, the
stronger a certain interest group, the higher the probability of a state to adopt
regulation that benefits this group and vice versa. From an empirical stance, RESE support schemes affect two types of stakeholders: groups that seek rents from
Copyright 䉷 2012 by the IAEE. All rights reserved.
6 / The Energy Journal
RES-E support and groups that would lose their previous rents from fossil electricity sources (FES-E) because RES-E substitutes them. To capture this, we take
quantifications similar to the ones applied in the U.S. studies.
• ISES is a variable that indicates the years of existence of a state chapter
of the International Solar Energy Association. Lyon and Yin (2010)
test for its sister organization, the American Solar Energy Association
(ASES).
• UTIL is a proxy variable representing the market power of utilities on
state electricity markets. Data is taken from a OECD survey that quantifies power market concentration on a horizontal and vertical dimension on a one to six scale (Conway and Nicoletti, 2006).
• ATOM is a variable that indicates the years of existence of a national
nuclear association that is allied with the European Atomic Forum
(FORATOM), a trade association that rallies for nuclear energy in
Europe.
Private interest theory supposes ISES to have a positive effect on the
dependent variable because RES-E lobby groups and their industries are potential
beneficiaries of RES-E supportive regulation. In contrary, UTIL and ATOM are
predicted to counter this effect since conventional power suppliers would be challenged by the new policy supported RES-E suppliers.
Public interest theory assumes the policymakers to have a vital interest
in helping to produce public goods such as clean air, employment and affordable
energy. According to pure private interest theory, the policymaker is the agent of
the interest groups. In public interest theory, the policymaker’s principal is the
society as a whole instead. Energy consumers and taxpayers as huge parts of
society make up the electorate to which the policymaker is accountable. Thus,
the adoption of RES-E support schemes may also depend on their costs for society. In order to quantify the degree to which a society can afford to replace
relatively cheap FES-E by costlier RES-E, we fall back on GDP per capita and
the electricity price for private end users. To cover the industrial dependence on
affordable power, we test for the energy consumption to GDP ratio. Data has been
provided by Eurostat (2010).
• GDP represents the gross domestic product per capita.
• EPPC stands for the electricity price per kwh for private consumers.
• EINT quantifies the energy intensity as the ratio of total energy consumption in Btu per 1,000 US-Dollars of GDP.
Following the U.S. studies and public interest theory, the poorer a country, the lower its likelihood of RES-E support. Conversely, we predict a positive
link between GDP per capita and the dependent variable. The same logic can be
applied to private and industrial consumers. Hence, we expect EPPC and EINT
to lower the odds.
Copyright 䉷 2012 by the IAEE. All rights reserved.
What Drives States to Support Renewable Energy? / 7
Beside these cost aspects, there are possible benefits for society that may
increase the likelihood of policy adoption. RES-E investment stimuli created approx. 550,000 new workplaces in Europe since 1990; 150,000 in 2008 and 2009
alone (EREC, 2010). The replacement of PM10 rich FES-E by RES-E technologies is another public good. Thus, reducing the public bad of air pollution can
attract policymakers to adopt regulation. Therefore, we predict the likelihood to
increase with a rise in UEMP or AIR, respectively.
• UEMP is the state unemployment rate as a percent of total work force.
Data has been provided by Eurostat (2010).
• PM10 is a metric for national air pollution. It represents the total particulate formation as an equivalent to 10,000 tonnes of PM10. Data
has been provided by the European Environment Agency (2009).
In an array of regressions, we control for further variables that have
frequently been brought up by qualitative analyses and the U.S. studies mentioned
above. We run robustness checks with controls separately and in combined models.
• NBOR is the ratio of neighbor states that have already implemented
RES-E support schemes and the total number of neighbor states. Chandler (2009) introduced learning effects between U.S. neighbor states.
We applied his indicator to the EU27 sample.
• GREEN stands for the number of parliamentary seats that are occupied
by the national Green party. Data has been provided by the European
Green Party (2010). This indicator mirrors the party indicators for the
Democratic or Republican party variables in the U.S. studies.
• EFAM is a categorical metric for the state’s electoral family. It distinguishes between majoritarian, combined and proportional electoral
systems (Norris, 2009). We control for this polity aspect because the
Greens, as a supposedly small party, have a structural disadvantage in
two-party systems.
• EU is a dichotomous indicator representing the EU Directive 2001/
77/EC on RES-E. It has been the EU’s first binding directive that
obliges state legislators to support RES-E. Marques et al. (2010) employs the very same dummy to control for the EU’s impact on RESE development.
• SOL quantifies the solar energy potential as global radiation in kwh
per square meter. Data has been provided by the Global Energy Network Institute (2010). The solar potential is classified into low (⬍1050
kwh/m2a), middle (1050–1599 kwh/m2a), and high (⬎1599 kwh/
m2a) degrees of radiation.
Copyright 䉷 2012 by the IAEE. All rights reserved.
8 / The Energy Journal
Table 2: Logit Regression Results on the Relative Odds of Policy Adoption
Variable
Model (1)
Model (2)
Model (3)
ISES
UTIL
ATOM
1.331** (2.05)
–2.836* (–1.8)
–0.244 (–0.96)
3.766*** (2.94)
–2.701* (–1)
0.672 (1)
1.948*** (4.27)
–2.396* (–1.68)
–0.214 (1.31)
GDP
EPPC
EINT
UEMP
PM10
0.001 (1.6)
–0.018 (–0.19)
–2.712 (–1.41)
1.374*** (2.79)
0.042 (1.63)
0.001 (0.86)
–0.074 (–0.42)
–5.870 (–1.62)
2.619*** (2.68)
0.054 (0.54)
NBOR
GREEN
EFAM
EU
SOL
–5.060 (–0.7)
–0.335 (–0.89)
25.741*** (2.82)
6.076 (0.78)
49.715*** (4.94)
0.905 (0.18)
–0.203 (–1.45)
4.081 (1.37)
2.777 (0.96)
16.797*** (3.95)
* p⬍0.1, ** p⬍0.05, *** p⬍0.01; t statistics in parantheses.
5. FINDINGS
The estimations base on the EU27 sample from 1990 to 2010. The actual
N of 237 is lower than the total N of 567 because cases are dropped from the
sample as soon as a states adopts regulation in order not to bias the coefficients
of the time series cross-section regression. Table 2 provides the outcome of three
logit models. The first column presents the full model. The second model introduces the controls. Model three treats private interest variables and controls exclusively.
Various additional model runs have verified the findings to be robust for
separate and combined exclusion of private interest variables, public interest variables and controls. The “variance inflation factor” (VIF) tests for multicollinearity
since the model runs on duration data. Studenmund (2009) suggests a VIF alert
level of five. O’Brien (2007) finds designs which allow an even higher level.
Neither the tested metrics nor their mean VIF ranges at or above the margin of
five. Furthermore, the variables do not correlate among each other at or above
the 0.7 level.
We find four coefficients remaining at a significant level throughout the
regression table at hand and the robustness check: ISES, UEMP, and SOL have
a positive impact on policy adoption while UTIL comes with a negative effect
onto the likelihood of policy adoption.
6. DISCUSSION
The significant results provide evidence in support of private interest
theory. In particular, the existence of lobby groups is associated with RES-E
Copyright 䉷 2012 by the IAEE. All rights reserved.
What Drives States to Support Renewable Energy? / 9
Table 3: Comparison of EU27 and U.S. findings
Variable
EU
U.S.
Solar Industry
Nuclear Industry
Oil and Gas Industry
Coal Industry
Ⳮ**
–
–***1
Market Concentration
Restructured Market
–*
Private Interest
–1
–1
–**1
Public Interest
GDP
Ⳮ
Ⳮ*2
Electricity Price
–
–1
Energy Dependency
–
–3
Unemployment Rate
Ⳮ***
–**1
Air Pollution
Ⳮ
Ⳮ1,3
Neighbor Effect
Ⳮ
Ⳮ**3
Green Seats
Democratic Seats
Republican Seats
Liberal Government
–
Proportional Elec. System
Ⳮ
EU Directive 2001/77/EC
Ⳮ
Solar Potential
Wind Ⳮ Solar Potential
Wind Potential
Ⳮ***
Controls
Ⳮ***1
–**2
Ⳮ***3
Ⳮ***1
Ⳮ3
Ⳮ**1
* p⬍0.1, ** p⬍0.05, *** p⬍0.01. 1 Lyon and Yin (2010), 2 Huang et al. (2007), 3 Chandler (2009).
policy adoption. Both links are consistent with Becker’s (1983) competition
model. On the one side, solar interest groups come with a positive link to the
dependent variable. As possible beneficiaries of RES-E investment, they have an
incentive to lobby in favor of policy adoption. Thus, the coefficient is revealed
to be positive. On the other side, previous rent-takers will be challenged by the
legislation. Their efforts to curb motivation to support RES-E is expressed in the
negative coefficient of UTIL, a proxy for utilities’ market power. Utilities have a
disincentive to get RES-E legislation ratified because it allows new producers to
enter the market, thus challenging their segment by decentralizing the production
park. The ATOM indicator, representing another conventional power lobby, implies this interpretation to be correct, although it does not come at a significant
level. As can be reviewed in Table 3 the findings for the conventional energy
Copyright 䉷 2012 by the IAEE. All rights reserved.
10 / The Energy Journal
interest groups and the market proxies are most consistent with Lyon and Yin
(2010).
Both metrics, UTIL and ATOM, have an implication for the choice between RPS and FIT. In the FIT scenario, the market entry barriers are lowered
because new RES-E producers are legally entitled to feed electricity into the grid.
Thus, conventional power suppliers—in many states used to their regional monopolies—are challenged at the production level. Loosing market shares, they
have an incentive to lobby against FIT adoption. In the RPS scenario, they keep
their grip on the production park. Such a quota scheme requires utilities to forward
a certain RES-E share of their portfolio to the end user but it does not increase
competition on the power production market by lowering barriers. Therefore,
private interest theory would expect conventional power suppliers to favor a quota
as a second best solution over a FIT as a third best solution, in case their first
best solution of no regulation is no longer an option. Interestingly, the coefficient
of the ATOM indicator changes from negative to positive if the FIT cases are
excluded from the sample. The low number of quota cases in the EU27 sample
however, does not produce significant results. The UTIL indicator switches back
and forth which does not allow us to give stable interpretations. Intuition however
would suggest the same explanation. Conventional power suppliers, represented
by UTIL and ATOM, fight FIT adoption more decisively than RPS schemes.
The regressions also provide support for the public interest hypothesis
that weak labor markets encourage policymakers to shift incentives in favor of
jobs in the renewable energy sector. At first glance, this is surprising because we
would have expected policymakers to subsidize conventional power production
in economically bad times. As can be reviewed in Table 3, the U.S. sample
produces such a link. For the EU27 sample, the UEMP coefficient implies the
contrary. The likelihood of policy adoption goes up with an increase in unemployment rate. This connection suggests that policymakers have left the stage of
paying lip service to RES-E technologies in downturn phases only, now seriously
betting on RES-E to create jobs. If the 2000 to 2010 data is excluded from the
sample, the coefficient turns negative. An explanation can be that prior to 2000,
policymakers relied on FES-E. With growing struggles that have been mentioned
in this article’s very first sentence, there has been a shift in labor market policy
in favor of job creation in the renewable energy sector in Europe. Continued
jobless industrial growth may also shift strategies of U.S. unemployment policy
in the future.
Other public interest indicators do not stand up to scrutiny. They produce
the same signs as have been presented in the U.S. studies. On both sides of the
Atlantic, GDP per capita seems to slightly increase the probability of RES-E
support, allowing for the claim that a state needs to be prosperous enough to
support RES-E. This interpretation is supported by the negative link between the
electricity price and the likelihood of adoption. Air pollution, quantified by the
PM10 indicator, captures what public theory claims about a policymakers task to
help producing the public good of clean air. Although GDP, EPPC, and PM10 do
Copyright 䉷 2012 by the IAEE. All rights reserved.
What Drives States to Support Renewable Energy? / 11
not meet significance criteria, the convergence with the significant U.S. findings
indicates that the theory itself is on the right track.
The control variables neither reject predictions nor do they provide significant connections. The solar potential is the exemption to the rule. Sunny places
appear to be more attractive for policymakers to adopt RES-E regulation. The
impact is even higher if we take the PV feed-in-tariffs as the dependent variable.
High solar radiation increase the return of investment in solar power systems of
course. This hypothesis has been verified by Lyon and Yin (2010) as well. As
has been pointed out by Chandler (2009) for the U.S. states, we can also speak
about state-to-state learning effects in the EU. The low significance level does
not allow for verification however. The same can be said about EFAM, a categorical metric that classifies the electoral system, and the EU dummy. Green seats
do not seem to make significant difference—a claim that has been put forward
by Laird and Stefes (2009) in a comparative case study for the U.S. and Germany.
Very generally speaking, the article concludes that private interest hypotheses come close to reality. They emphasized that supporter groups boost while
opposed groups block RES-E legislation. (a) The existence of a solar energy
association increases the probability of a state to adopt regulation. (b) Solar radiation, and (c) the unemployment rate also increase the odds. (d) Electricity
market concentration decreases the probability of transition. Interestingly, the
nuclear society indicator as well as the market concentration indicator turn positive if we exclude the FIT cases from the sample. We suggest that utilities favor
quotas over feed-in-tariffs because the latter can challenge their grip on the production park.
REFERENCES
Becker, G.S. (1983). “A theory of competition among pressure groups for political influence.” The
Quarterly Journal of Economics 98(3): 371–400. http://dx.doi.org/10.2307/1886017.
Chandler, J. (2009). “Trendy solutions. Why do states adopt Sustainable Energy Portfolio Standards?”
Energy Policy 37(8): 3274–3281. http://dx.doi.org/10.1016/j.enpol.2009.04.032.
Conway, P. and G. Nicoletti (2006). “Product market regulation in the non-manufacturing sectors of
OECD countries: Measurement and highlights.” OECD Publishing, OECD Economics Department
Working Paper #530.
Couture, T. and Y. Gagnon (2010). “An analysis of feed-in tariff remuneration models: Implications
for renewable energy investment.” Energy Policy 38(2): 955–965. http://dx.doi.org/10.1016/
j.enpol.2009.10.047.
DSIRE (2010). “Database of State Incentives for Renewables and Efficiency.” July. Accessed at http://
www.dsireusa.org/.
European Environment Agency (2009). “Greenhouse gas emission trends and projections in Europe
2009.” European Environment Agency. EEA Report 9/2009. Brussels.
European Green Party (2010). “National Results Archive.” September. Data provided by the European
Green Party. Brussels.
European Renewable Energy Council (2010). RE-thinking 2050. European Renewable Energy Council. Brussels.
Eurostat (2010). “Statistics Database.” July. Accessed at http://epp.eurostat.ec.europa.eu.
Global Energy Network Institute (2010). “Library. Solarenergy in Europe.” July. Accessed at URL
http://www.geni.org/globalenergy/library/.
Copyright 䉷 2012 by the IAEE. All rights reserved.
12 / The Energy Journal
Heymann, M. (1998). “Signs of hubris. The shaping of wind technology styles in Germany, Denmark,
and the United States.” Technology & Culture 39(4): 641–670.
Huang, M.-Y., J.R. Alavalapati, D.R. Carter and M.H. Langholtz (2007). “Is the choice of renewable
portfolio standards random?” Energy Policy 35(11): 5571–5575. http://dx.doi.org/10.1016/
j.enpol.2007.06.010.
Joskow, P.L. (2009). “Foreword: US vs. EU electricity reforms achievement” In F. Leveque and J.M. Glachant , Electricity reform in Europe: towards a single energy market. Chatenham: Edward
Elgar. 13–29.
Keller, S. (2010). “Sources of difference. In answer to the article about diverging paths of German
and US policies for renewable energies.” Energy Policy 38(8): 4741–4742. http://dx.doi.org/
10.1016/j.enpol.2009.10.035.
Kiefer, Nicholas M.; G.R. Neumann (1989). Search models and applied labor economics. Cambridge:
Cambridge University Press.
Kiefer, N.M. (1988). “Economic duration data and hazard functions.” Journal of Economic Literature
26(2): 646–679.
Knittel, C.R., 2006. “The adoption of state electricity regulation. The role of interest groups.” Journal
of Industrial Economics 54(2): 201–222. http://dx.doi.org/10.1111/j.1467-6451.2006.00280.x.
Laird, F.N. and C. Stefes (2009). “The diverging paths of German and United States policies for
renewable energy: Sources of difference.” Energy Policy 37(7): 2619–2629. http://dx.doi.org/
10.1016/j.enpol.2009.02.027.
Lauber, V. and L. Mez (2004). “Three decades of renewable electricity policies in Germany.” Energy
& Environment 15(4): 599–623. http://dx.doi.org/10.1260/0958305042259792.
Lyon, T.P. and H. Yin (2010). “Why do states adopt renewable portfolio standards? An empirical
investigation.” The Energy Journal 31(3): 133–157. http://dx.doi.org/10.5547/ISSN0195-6574-EJVol31-No3-7.
Marques, A.C., J.A. Fuinhas and J. Pires Manso (2010). “Motivations driving renewable energy in
European countries. A panel data approach.” Energy Policy 38(11): 6877–6885. http://dx.doi.org/
10.1016/j.enpol.2010.07.003.
Menanteau, P., D. Finon and M.-L. Lamy (2003). “Prices versus quantities. Choosing policies for
promoting the development of renewable energy.” Energy Policy 31(8): 799–812. http://dx.doi.org/
10.1016/S0301-4215(02)00133-7.
Norris, P. (2009). Democracy crossnational data. Tech. Rep. 3, Cambridge MA.: Harvard University.
O’Brien, R.M. (2007). “A caution regarding rules of thumb for variance inflation factors.” Quality
and Quantity 41(5): 673–690. http://dx.doi.org/10.1007/s11135-006-9018-6.
Peltzman, S. (1976). “Toward a more general theory of regulation.” National Bureau of Economic
Research. NBER Working Paper #0133.
Rabe, B.G. (2004). Statehouse and greenhouse. The emerging politics of American climate change
policy. Washington D.C.: Brookings Institution Press.
Renewable Energy Policy Network for the 21st Century (2009). “Renewables. Global status report.
2009 update.” Renewable Energy Policy Network for the 21st Century. Paris.
Res-legal (2010). “Legal sources on renewable energy.” July, accessed at http://res-legal.eu/.
Studenmund, A.H. (2009). Using econometrics. A practical guide, Boston: Pearson.
Copyright 䉷 2012 by the IAEE. All rights reserved.