PADM530Wk7&

Social Problems, 2016, 63, 284-301 doi: 10.1093/socpro/spw007 Article OXFORD

Incentives or Mandates? Determinants of the Renewable Energy Policies of U.S.

States, 1970-2012 Michael Vasseur

RAND Corporation

ABSTRACT Why might states adopt policy instruments of one type over another, and how does this

choice impact the overall portfolio of policy instruments a state adopts? To address these questions this article examines renewable energy policy instrument adoption by U.S. states and argues that states adopt instruments of different types based on their state-level eco­ nomic, political, institutional, and cultural characteristics. I test these claims by examining the tax incentive- and regulatory mandate-based policy instruments adopted to promote re­ newable energy generation by U.S. states over a 40-year period. Using random effects Poisson regression analysis, I find that state affluence, environmental movement organiza­ tion density, and fossil fuel production predict the number of policies a state is likely to adopt, while an affinity for a neoliberal ideology, U.S. senators environmental voting re­ cords, and prior policy actions predict the types of policies a state adopts. These results re­ inforce perceptions of economic factors as key predictors of renewable energy policy, but also highlight the importance of less frequently examined cultural factors for explaining a state’s portfolio of policies. These analyses offer a robust picture of the relationship between tax incentive and regulatory mandates, the two types of programmatic approaches that have dominated many policy domains in the United States over the past 40 years.

KEYWORDS: renewable energy; fiscal policy; subnational politics; regulation; neoliberalism.

Research on renewable energy policy adoption has tended to focus on how U.S. states come to adopt a particular policy instrument at a given point in time (Chandler 2009; Coley and Hess 2012; Daley and Garand 2005; Fowler and Breen 2013; Huang et al. 2007; Vachon and Menz 2006; Yi and Feiock 2012), or how a specific energy industry develops in a state (Campbell 1988; Jasper 1990; Podobnik 2006; Sine and Lee 2009; Vasi 2006, 2009, 2011). Given that most U.S. states adopt mul­ tiple, often very different, policy instruments within the same domain, this scholarly focus on single policy instrument adoption sidelines important questions. For all we have learned about how states adopt individual policy instruments, we know much less about the processes that shape the overall

The author wishes to thank Brian Steensland, Clem Brooks, Patricia McManus, Fabio Rojas, and the anonymous Social Problems re­ viewers for insightful comments and suggestions on prior drafts. This research was assisted by a fellowship from the Dissertation Proposal Development Fellowship Program of the Social Science Research Council with funds provided by the Andrew W. Mellon Foundation. A prior version was presented at the 2014 American Sociological Association Annual Meeting. Direct correspondence to: Michael Vasseur, RAND Corporation, 4570 Fifth Avenue, 5600, Pittsburgh, PA 15213. E-mail: mvasseur(S)rand.org.

© The Author 2016. Published by Oxford University Press on behalf of the Society for the Study of Social Problems. All rights reserved. For permissions, please e-mail: journals.permissions(2)oup.com

. 284

Determinants of the Renewable Energy Policies of U.S. States • 285

portfolio of policies that states adopt. An accounting of a state’s portfolio of policies is vital to under­ standing how policy instruments relate to each other. This omission is potentially significant because the factors that influence how many policy instruments a state adopts in a policy domain may not be the same as those influencing the types of policy instruments a state adopts. Furthermore, examining the adoption of individual policy instruments in isolation muddles the relationships between the in­ struments themselves. Accounts of policy making focused on consistent state action emphasize the process by which a state comes to favor policy instruments of one type over all others. In this view only a limited number of policy instruments will be adopted in a state, and the expectation is that they will be similar to each other. At best, the resulting instruments might be unrelated to each other, but it is seen as more likely that the adoption of a policy instrument of one type decreases the odds of a state adopting a different type of policy instrument. A shift to an analysis of a state’s entire port­ folio of policies allows for the possibility that state policy actions might be consistent with a more general orientation.

Questions of policy instrument type are of increasing importance in the polarized political context of the contemporary United States. Specifically, in a political climate where a policy’s relationship to broader narratives surrounding taxation or regulation is central to its feasibility, it is important to understand how states have adopted their portfolio of policies vis-a-vis incentives and mandates. The renewable energy policies of U.S. states offer an opportunity to examine incentive- and mandate- based policies in a well-delimited policy domain. This distinction is especially important given the central role of states in addressing global climate change in the U.S. context. Even when spurred on by federal pressure to date it has been left up to various states to craft their own plans for emission re­ duction (EPA 2014). With no signs that climate change will emerge as an issue addressed by federal legislation in the near future a more complete understanding of state’s portfolios of policies will be key. To do this, I build on and extend existing accounts of the determinants of energy policy.

I follow a perspective that argues that patterns of state policy instrument adoption result from a combination of state-level political interests, policy-making institutions, and ideational factors such as culture, discourse, or policy ideas (Beland and Cox 2011; Campbell 2002; Lieberman 2002; Padamsee 2009). Interest-based explanations focus on political actors, including both politicians and outside forces. These accounts examine how actors use policy to improve their chances at reelection. Institutional accounts show that the structure of formal policy-making institutions influences policy adoption. Ideational accounts focus on how actors interpret the problems and policy solutions being considered, and how views of legitimacy both constrain and enable policy choices.

Research on energy policy has typically not accounted for political explanatory factors beyond those concerning electoral interests. This omission is not surprising; existing research on renewable energy has been historically dominated by economically deterministic views. Until relatively recently the distribution of energy producing resources was all that was considered important for understand­ ing a state’s energy policy (see Lowry 2008 for a discussion and critique of these views). Building upon this existing approach and its insights, a logical next step is to account for the politics of the elites controlling access to these resources and the potential importance of states’ broader institu­ tional and cultural contexts.

Research in other policy domains has shown that political institutions and culture impact the kind of policy instruments a state adopts. To the extent that policy actions are consistent with a broader orientation or strategy of action, institutional and cultural factors shape a state’s portfolio of policies through a variety of mechanisms. Whether though specifying the means seen as appropriate to reach some policy end (Dobbin 1994), establishing separate policy tracks (Hacker 2002), forming styles of policy action (Jasper 1990), or locking states into a path through feedback mechanisms (Pierson 2000), institutional and cultural factors influence policy adoption. Recognizing these types of influ­ ences means that the content of a state’s portfolio of policies is not simply a hasty assembly of what­ ever problems require attention at the moment and what solutions are in the air at a given point in time (Baumgartner and Jones 1993; Kingdon [1995] 2004). Instead, policy content is a product of

286 . Vasseur

the structure of a state’s economy, the features of its political landscape, and the affinities established by its institutions and political culture. As analysis moves beyond the adoption of individual policies and to an examination of the broader portfolio of policies held by states, assessing the impact of these institutional and cultural factors is vital.

RENEWABLE ENERGY POLICIES IN U.S. STATES Renewable energy policy in the United States serves as excellent opportunity to examine policy port­ folios, as states have taken a variety of actions in this domain. These actions tend towards one of two strategies of policy action: tax incentives or regulatory mandates.1 These policy strategies assume dif­ ferent views of how social change occurs. Incentives rely on an aggregation of individual actions to create change, while mandates rely on top-down demands levied by the state. Importantly, both strat­ egies are centered on the goal of producing more renewable energy within a state, even if they do so by different means.

How do states choose between incentives and mandate strategies when adopting policy instru­ ments? This question is central in research within the framework of fiscal sociology (Martin, Mehrotra, and Prasad 2009). Research in this area examines how governments, especially in the United States, forge social policy through the tax code, specifically incentives defined as tax expend­ itures (Hacker 2002; Howard 1997). Tax expenditures are income that would be collected by the government if not for exemptions added to the tax code for behaviors the government wishes to pro­ mote, such as buying a home or having children. While generally thought of as equivalent to direct state spending from a budgetary standpoint (see Prasad 2011, however), tax expenditure policies arise in the area of social policy through different conditions than more traditional state interventions. Factors such as divided partisan government, the relatively limited involvement of interest groups or social movements, and the use of tax expenditures in past policy domains are all associated with an increased likelihood of a tax expenditure being adopted (Hacker 2002; Howard 1997, 2007, 2009).

A more socio-legal perspective highlights that these differences are what lead to the more desirable qualities of using the tax code to promote social change. These include the fact that using existing political infrastructure in the tax code, compared to creating a new governance program, is more effi­ cient, as it does not require new administration. It is also more equitable as it is harder for any one interest group to dominate the program’s administration (Zelinsky 1993). In this view, the primary difference between direct spending and tax expenditures is the reaction a given policy form elicits from lawmakers and citizens (Zelinsky 2004). Regardless of the paradigm employed, scholars of so­ cial policy have identified important differences in the process of adopting tax incentive-based policies compared to more direct state actions.

U.S. states have adopted policies to promote the generation of renewable energy that fit two broader categories. The first category of policy widely adopted by U.S. states is tax incentives. These include deductions, credits, or rebates on personal or business taxes for renewable energy generation, or the remittance of property or sales taxes for the purchase of renewable energy generation systems. These policies use the tax code to provide incentives for the generation of renewable energy through the state avoiding the collection of revenues it would otherwise be entitled to. These policies are de­ signed to provide an enticement for individuals or corporations to enter the renewable energy mar­ ket, without the direct state intervention in shaping such a market.

The second category of widely adopted policy is regulatory mandates. These policies directly intervene in the relationship between utility companies and consumers. One example is energy port­ folio standards that mandate that electric utilities produce a certain percentage of their electricity from renewable sources or face a penalty. Other mandate policies include those that require a disclos­ ure of the sources used in a utility company’s electricity generation, and those that force utility

1 As of 2012, all but four states (Alabama, Arkansas, Mississippi, and Wyoming) have adopted at least one of these policies. Incentive policy adoption began in the mid-1970s. Mandate-based policies became widely adopted in the mid to late 1990s.

Determinants of the Renewable Energy Policies ofU.S. States 287

Table 1. Categories of Renewable Energy Policies adopted by U.S. States

Tax Incentives Regulatory Mandates

Income tax Energy portfolio standards Corporate tax Cap and trade program Sales tax Public benefits funds Property tax Generation source disclosure Tax rebate Green power purchasing option

companies to offer consumers the option to purchase power from green sources for a fee. Two other policies in this area, public benefit funds and cap-and-trade programs, are known as consumption taxes on electricity. In the case of public benefits funds, this is a surcharge on each unit of electricity consumed. This surcharge is normally set aside for a fund whose sole purpose is to fund new renew­ able energy projects within the state. Cap-and-trade policies do not directly tax energy consumption, but instead tax carbon emissions, producing the same end result. All of these policies involve more coercive state action compared to incentive-based policies.

Importantly, regardless of the type of action used, these policies represent the exercise of state power; in neither case is the development of renewable energy left entirely to the market. Currently, producing energy from renewable sources costs more than doing so from fossil fuels. Policy elites rec­ ognize this and know that without state action renewable energy production is unlikely to increase at a substantial rate. If states want to produce more renewable energy, they must act. These actions vary in their content, but all of these policy instruments incorporate aspects of market logic and are funda­ mental examples of state action. Table 1 summarizes the policies and categories examined in this article.

Impact of Interests, Institutions, and Ideational Factors on Renewable Energy Policy Research on energy policy has overwhelmingly focused on economic and political interests in predict­ ing the adoption of regulatory mandate policies (for an exception, see Fowler and Breen 2013). States are thought to be more likely to adopt renewable energy policies if they have available renew­ able energy resources (Chandler 2009; Vasi 2006) and do not produce fossil fuels (Coley and Hess 2012; Huang et al. 2007; Vachon and Menz 2006). Other research highlights the importance of state affluence as a predictor of energy and environmental policies (Daley and Garand 2005; Ringquist 1993; Vachon and Menz 2006). It is clear that economic concerns impact energy policy.

Beyond economic factors, scholars have identified two political interests associated with the adop­ tion of renewable energy policy: the partisanship of state political elites and the capacity of environ­ mental social movements. Studies of the determinants of state renewable energy policy have frequently noted the importance of state politicians on policy adoption. States with more Democratic (Coley and Hess 2012; Fowler and Breen 2013) or liberal (Chandler 2009; Vachon and Menz 2006; Yi and Feiock 2012) politicians are more likely to enact renewable energy policies. Beyond polit­ icians, more environmental social movements in a state increase the likelihood of renewable energy policy adoption (Podobnik 2006; Vasi 2006, 2009, 2011). This research suggests that states adopt re­ newable energy policies through a combination of political interests and a desire to avoid disrupting the prevailing economic conditions of a state. From this perspective we should expect that states with favorable political interests and economic conditions will be more likely to adopt renewable energy policies. As these studies have focused overwhelmingly on mandate outcomes, it remains uncertain whether the policies adopted will be incentives or mandates.

Institutional and cultural factors have received less attention in research on renewable energy pol­ icy than interest-based accounts, but are likely to be important predictors of renewable energy policy

288 . Vasseur

content. I expect political institutions to influence policy adoption through two mechanisms: the openness of the political process, and as a measure of a state’s capacity to administer policies effect­ ively. State legislative systems differ in the degree to which legislative change is possible outside of the legislature, especially in the realm of citizen engagement in the legislative processes. Some states keep legislative power tightly held by their legislative bodies, while others allow for a range of citizen actions to advance legislation without the legislatures. These actions, known as direct democracy, in­ clude referendums, initiatives, and related ways to bypass state legislatures. States that allow direct democratic actions represent not just a different system of legislative action, but also a broader under­ standing of politics than states with a more closed process. As such we might expect different policy actions to be taken on the basis of direct democracy.

The other way political institutions impact renewable energy policies is through a state’s capacity to administer regulations and governance programs. Capacity to govern has long been recognized as a key factor in pollution control policies (Ringquist 1993), and there is reason to expect it will matter for renewable energy policies as well. If a state lacks sophisticated political institutions, or has not in­ vested in the development of bureaucratic organizations, it would be unlikely for it to engage in ef­ forts to promote social change that require new regulatory infrastructures in order to function. These states simply would not have the capacity to construct and administer the additional infrastructure required to manage these new regulatory programs. For example, states with legislatures that are rarely in session, or with governors who lack mechanisms to influence the policy process, have less of an ability to administer new regulatory programs than states with more sophisticated political institutions.

Beyond political interests and institutions, I expect that ideational factors, specifically political cul­ ture, will also influence renewable energy policy adoption. By political culture I refer to normative paradigms held by political elites (Campbell 2002). These paradigms influence policy adoption by limiting the range of policy instruments that are perceived as desirable or legitimate solutions (Campbell 2002; March and Olsen 1989; Padamsee 2009; Skrentny 1996; Steensland 2008). Rather than assess these paradigms in general terms, I isolate specific, and measurable, aspects of political culture to compare alongside traditional interest and institutional factors. In place of a general meas­ ure of political culture I address two specific dimensions, environmental culture and neoliberal ideol­ ogy. Environmental culture assesses the degree to which state policies consistently favor issues of environmental protection. In these states norms regarding the environment, and the appropriate lengths the government should go to protect it, will differ from less environmentally friendly states. These norms will then lead to different sets of policy instruments being adopted.

The other aspect of political culture I focus on is a state’s normative orientation toward involve­ ment in market activity. In some states these orientations manifest as a general inclination against any form of regulatory policy, regardless of the policy domain involved. I deem states in which this orien­ tation manifests as having an affinity for a neoliberal ideology. These affinities arise based on histor­ ical patterns of party politics and alignment with business interests (Prasad 2006). Neoliberal ideology is usually thought to contain three primary dimensions: a focus on reducing the welfare state, opposition to redistributive tax policies, and minimal regulations on industrial policies and labor markets (Campbell and Pedersen 2001; Prasad 2006). Each of these dimensions represents a distinct, but related, avenue by which a preference for the market over the state is expressed and how re­ sources should be distributed in a society. This preference stems in part from a view of the market as not only a legitimate means of distributing resources in society but a moral one (Somers and Block 2005). States conforming to this ideology would be unlikely to adopt any form of mandate that inter­ feres with market activity in any policy domain. In order to ascertain the impact of neoliberal ideol­ ogy on policy adoption, any measure used must draw on policy experiences with policies far enough removed from renewable energy so as not to conflate a general inclination against regulations with a lack of regulation in one specific instance.

Determinants of the Renewable Energy Policies of U.S. States . 289

Taken together, these accounts suggest three broad arguments: one based on the impact of eco­ nomic and electoral interests, another focused on state institutional capacity, and a third that empha­ sizes a state’s cultural orientations towards regulation and the environment. I expect to see different political factors, or combinations thereof, influence the adoption of policy instruments associated with different strategies. Incentive policies are generally thought of as interest based and will be heav­ ily influenced by economic and political factors. As incentives use an existing governance infrastruc­ ture, the tax code, they will be less reliant on sophisticated political institutions than mandate-based policies. More broadly speaking, mandates must contend with broader cultural views of regulations and the market if a state is going to adopt them. Assessing these expectations requires examining the adoption of incentive- and mandate-based policies while accounting for economic, interest-based, in­ stitutional, and cultural factors simultaneously.

DATA AND MEASURES To test these claims this article draws on an original longitudinal data set of U.S. states’ characteristics and policy actions. Drawing from a variety of secondary sources data was assembled every year start­ ing in 1970, the start of what is generally considered the "environmental decade” and an era that saw sustained pressure on environmental issues enter the political sphere. Data collection stops in 2012 as the most recent available data for many predictors. The unit of observation in these data is the state-year, with 43 observations per state and a final analysis sample of 2,103 state-years.2 3 Descriptive statistics and a pairwise correlation matrix of all variables are presented in Table 2.

I use two dependent variables in the analysis to follow. Each is a count of policy instruments of each type adopted by a state in a given year. I chose to use a count of the policy instruments as a meas­ ure of strength given the uncertainty surrounding the effectiveness of these policies. Given that at this point it is unclear to scholars, let alone policy elites, which of these mechanisms will be most effective I find it reasonable to assume that states that adopt a wider range of possible solutions are more com­ mitted to renewable energy generation. This suggests that the surest way to measure a state’s commit­ ment to renewable energy is by counting the different methods they are employing to promote it. ’

Tax incentives include personal or corporate income tax expenditures, property or sales tax ex­ emptions, and tax rebates for the production of renewable energy or renewable energy systems. Regulatory mandates include renewable energy portfolio standards, mandatory generation disclos- ures, green power purchasing options as well as public benefits funds and cap and trade programs. All data on these policies are drawn from the Database of State Incentives for Renewables & Efficiency (DSIRE), a collection of state-level regulatory environments maintained by the North Carolina Solar Center and the Interstate Renewable Energy Council (DSIRE 2012). A state was considered to have a policy instrument if a review of the corresponding bill or statute included the specified mechanism for increased generation of renewable energy from wind, solar, or geo-thermal sources.4 Policies focused solely on other potential sources of renewable energy generation, most prominently

2 As mandate policy instruments do not spread widely until the 1990s additional models were fit for this outcome that omit data for 1970-1989. These models produce substantively similar results.

3 Another possible metric could be the relative strength of the policies adopted. This is certainly an interesting possibility, and highly appropriate for some policy domains, but largely not for renewable energy policy. It is certainly true that early adopting states adopt weaker policies to start than later adopters, but the early adopting states frequently amend their policies upward to match the new standards. Examining states in 2012 I find no significant variation in the strength of these policies. I suspect that much of this change is technological in nature. As renewable energy generation technologies become more advanced policy elites modify proposals to keep up with these developments. Given this lack of variation I continue to use the count of policy instru- merits adopted as the best longitudinal measure available.

4 If a state later repeals or defiinds a policy instrument the count for that state decreases. During the time period under study in­ stances of this are relatively rare. Several states have seen bills introduced that would restrict renewable energy generation, but most of these repeal efforts have not been successful. This will surely be an issue for future research in this policy domain, but does not greatly impact the time period examined in this study.

290 Vasseur

o o O <N

Oo Os l> —i tJ-

<3 i0/*-•

roO

On

fi 0

J h

&oU n3C3

1 /3u

flj•M co<y>

o> Q r$ ju 3 cdH

O NO NO 1-o o o rO

o •'t ON NO OOn o o <N <N o 1 i 1

O rO ON NO o <NOO O *—1>—Io o * * 1 1

o o NO 1" ro NOIN o o o rO o O i-H 1 i

o r\ rO ro O r\ 00 <NNO o o ro O i-H <N o roi-H i i o NO NO NO OJ NONO o i-H o O ’-H o (N '—1<si-H 1 1 1

o ON (N <N o NO NO i-H 00 00o <N <—(O o NO o (N <N <N l l l

O 00 NO 00 NO NO rO NOfO O o o O o ’—1o o Ĥ O ’—1 l 1 1 !

o NO NO <N 00 rO 00 ON rS <N o O <N <N •-H o o O (N rrj r/)

1 1 i 1 i O On NO (N ON o 00 ON ro O O O'—( o O <s <s o '“lo 1- (N*—i i

i-H 1 o

1 00

l i

X ON *—i 003 NO NO ro o K r-H ro r-H rO "T i-H r-H ONO 00 >—i t> O

■1" O O O-—1 ON rO »/>< o o O o NO o NO o i-H O O o 00 00 ON NO NO o ON ro ON o (NQ i—H ON o 1- ro ’T 00 NO I> t> 'T NO (Nco r— f 00 On r-H (N ONi-H (N

c 00 NO t> ON O NO NO O NO NO 00<3 t> [> rO !■ NO NO '—1 r—1 NO ONO On orS ro NO

"O2ctfa c<DU.gi3oH

b g 2 u 43 .a S i g <u h -5o6<DQ

+3 G t / 3 Q )

- aao 3 | £< w

c 2OJ ,-G H3 <*>w C/3 g g .2 2 « ^ n 5'S -2S uoos s1/3

« g cu •-<« cu H3 .> G -M•5 -2 ̂ on

a s ^

u a; 3 ^3n3OtH S I0> £3 00** 0/ "cd0)

>i *- < co

00<D kJ On

00 .G 'O G 4»

2j O.5 1/3 00 0/

>u

‘T im

e in

va ria

nt m

ea su

re .

Determinants of the Renewable Energy Policies of U.S. States . 291

hydroelectric and bio fuels, are omitted from analysis given controversy surrounding their environ­ mental impacts.

Economic concerns are a key factor driving energy policy. Data on a state’s economic conditions was assembled from three sources. The first, production of fossil fuels, is a time-invariant covariate measuring if a state has produced coal, oil, or natural gas for the majority of the years under examin­ ation. This was determined by consulting the total production from these energy types as listed in the State Energy Data System, a repository on energy production and consumption by states from 1970 through 2010 assembled and hosted by the U.S. Department of Energy (2010). I choose to use fossil fuel production as a measure of a state’s energy economy as it is the most inclusive measure available. Past research has focused primarily on coal (Huang et al. 2007; Vachon and Menz 2006) to the exclu­ sion of other energy sources. To avoid giving undo weight to a single energy type when all are import­ ant drivers of energy politics in the United States, I utilize fossil fuel extrachon as the broadest measure available. I use a binary measure to highlight the greater difference between producing and non-producing states, rather than focus on the more limited variation within producing states.s

Another key economic factor is the potential resources for renewable energy in a state. Given that the majority of renewable energy produced in the United States comes from wind power, I operation­ alize renewable energy potential as a state’s average annual wind speed.5 6 Wind speed data is time in­ variant and measured as of 1978. This measure is time invariant as there is very little change in wind speed over time, but far more variation between states. These data are drawn from the U.S. Statistical Abstract (U.S. Census Bureau 2012). For states with multiple measurement sites, averages were taken when calculating state totals. As a final economic factor I control for the affluence of a state as past studies have pointed to average income as a key predictor of state energy and environmental policy (Daley and Garand 2005; Ringquist 1993; Vachon and Menz 2006). Affluence is measured as a time varying measure of median annual household income of a state in thousands of 2011 dollars (U.S. Census Bureau 2012).

Other factors that influence energy policy are political in nature. Key among these is political inter­ ests, which I define as the partisanship of political elites and the presence of environmental move­ ments. Following past research (Coley and Hess 2012), control of state political institutions by sympathetic elites is operationalized as control of a state’s political institutions by the Democratic Party. Drawing on data from the Book of the States (Council of State Government 1979-2012) I counted the number of political institutions, including both legislative chambers and the office of the governor, ranging from 0 to 3, controlled by the Democratic Party in a given state-year. Control of a chamber was defined as having the majority of the legislators in that chamber.7

5 Using a continuous time-varying measure of fossil fuel extraction, logged to account for the skewed nature of the data, does not substantively change the results presented.

6 Given its dominance relative to other forms of renewable energy I focus this measure on wind energy. I avoid the use of a meas- ure of sunny days in area, data also collected by the National Oceanic and Atmospheric Administration (NOAA), as the link be­ tween this measure and an area’s suitability for solar power generation is less clear than wind speed and wind power. Additionally, despite its present surge, solar power is simply not a major source of power generation in the United States. According to the Energy Information Administration’s (2015) data wind energy accounted for just over 4 percent of total energy generation in 2013. By comparison, solar accounted for roughly a quarter of one percent (U.S. Census Bureau 2012), the smallest value listed. Nevertheless, a more comprehensive measure for renewable energy potential, including solar, wind, geothermal, and other sources (Deyette et al. 2003), was tested and did not change the conclusions offered. While more encompassing this meas­ ure is not available for all states (Alaska and Hawaii). Out of a desire to keep as many observations in the analysis as possible I maintain the wind speed item in presentation.

7 The federated system of government in the United States introduces diversity in the political systems enacted by states. In order to facilitate comparisons across states an attempt was made to standardize political institutions. Common among these differ­ ences are how states deal with a chamber that results in an equal number of legislatures from each party. States use a variety of techniques for deciding control in these cases, for ease of comparison I coded control based on the party affiliation of the major­ ity leader of the chamber. Furthermore, independent legislatures were counted as part of the Democratic Party only if they held Green or progressive affiliation. Additionally, two states do not record the partisanship of their state legislators during all or part of the period under study. When this information could not obtained those state years were excluded from analysis. This removes all of Nebraska’s observations and the first four years of observations for Minnesota.

292 . Vasseur

I also account for the density of environmental movements with a count of the number of envir­ onmental movement organizations (EMOs) active within a given state-year. These counts come from the Policy Agendas Project Encyclopedia of Associations Database.8 The Encyclopedia of Associations, Volume 1, National Organizations of the U.S. (EA) (Gale Research Inc. 1956-2010) pro­ vides a census of national social movement organizations, and has been used by many social move­ ment scholars to establish measures of social movement density (Andrews and Edwards 2005; Johnson 2006, 2008; Johnson, Agnone and McCarthy 2010; Johnson and Frickel 2011; Martin, Baumgartner, and McCarthy 2006; Minkoff 1997, 1999; Van Dyke and Soule 2002; Vasi and Strang 2009; Walker, McCarthy, and Baumgartner 2011). The Policy Agendas Project data set provides a count of these movement organizations every five years, which I expanded to 2010. Following past studies, (Johnson 2008; Johnson and Frickel 2011) groups were selected if their primary purpose was environmental conservation/protection or the production of renewable energy. This was established through use of keywords, headings, association name, and organizational description.9 Once a group was located I obtained a state address, and calculated state-year counts of movement organizations. While there is some concern about bias in the groups reported in the EA (see Brulle et al. 2007 and Johnson and Frickel 2011) the EA remains a valid measure of EMO density within a state-year. Concerns over bias in the EA center on the fact that an organization’s size and its proximity to Washington, DC, increase the probability of being listed in the EA (Brulle et al. 2007). Certainly there are movement organizations not reported in the EA that are active in the United States, but aside from a potential DC bias, there is no reason to expect any bias on the basis of organization size is not constant across states.

In order to assess a state’s capacity to adopt new legislation and administer new programs I ac­ count for the openness and sophistication of a state’s political institutions. I measure a state’s political institutions in three dimensions: the availability of direct democracy, the professionalism of the state’s legislature, and the governor’s power. These measures account for the relative capacity of the two most important political institutions in a state, and the ability of citizens to engage in direct legislative actions. Direct democracy is important to account for as some states have adopted renewable energy policies through this process. These are relatively few in number. To be clear, the vast majority of policy instruments are adopted through legislative channels, but I believe it is still important to ac­ count for this alternative avenue. Direct democracy is measured as a non-time varying indicator vari­ able for whether a state allows citizen initiatives during the period its renewable energy policies were adopted. This data is drawn from the Initiative and Referendum Institute (2013). I focus on the gen­ eral category of legislative initiatives as a way for citizens to bypass the legislature rather than the more restrictive constitutional amendments as this is the most likely direct democracy avenue renew­ able energy policies take.

The relative sophistication of a state’s political institutions helps determine their capacity to adopt new regulatory programs. More sophisticated legislatures have greater time and expertise to engage in the often technical debates surrounding energy policy, and are associated with increased resources for the administration and maintenance of regulatory policy instruments once adopted. Simply put, it is easier for a highly professional state to adopt a technically demanding policy, such as cap-and-trade, for example, than it is for a legislature composed of understaffed part-time legislators. Legislative pro­ fessionalism is measured using the Squire index (Squire 2007). This measure is a ratio of the

8 The data used here were originally collected by Frank R. Baumgartner and Bryan D. Jones, with the support of National Science Foundation (NSF) grant numbers SBR 9320922 and 0111611, and were distributed through the Department of Government at the University of Texas at Austin. Neither NSF nor the original collectors of the data bear any responsibility for the analysis re- ported here.

9 Keywords used following Johnson 2008 were: conservation, wildlife conservation, environment, environmental quality, environ­ mental protection, environmental health, toxic exposure, nuclear energy, ecology, pollution control, and hazardous waste. Given my focus on renewable energy policy I added keywords: solar energy, wind energy, and organizations in the general energy key­ word with a focus on efficiency or renewable/sustainable energies.

Determinants of the Renewable Energy Policies of U.S. States . 293

professionalism of a state s legislature relative to the U.S. Congress across three dimensions: staff, compensation, and time in session. Changes in level of professionalism are not drastic over time, but nevertheless the index is calculated at five points in time (1979, 1986, 1996, 2003, 2009) based on available component data. I utilize a measure of governor’s power following Krupnikov and Shipan (2012) that averages five different factors related to the governor’s ability to directly influence fiscal issues in their state. Governors’ power is potentially relevant as many of the incentive policy instru­ ments are adopted as part of broader fiscal or budgetary policies.

A final set of factors, tapping political culture, measure a state’s environmental culture and its affin­ ity for neoliberal ideology. I measure environmental culture as the average of the state’s U.S. Senators League of Conservation Voters scores (LCV 2012). The League of Conservation Voters (LCV) has maintained a ranking of every U.S. Congress member’s environmental voting record since the early 1970s, and represents a longitudinal, state-specific measure of the degree to which statewide elected officials in a state vote in favor of environmental protection causes. I focus on senators to capture statewide, rather than district-based, environmental affinity, and to ensure that the votes tallied in each year are of similar scope in their environmental impact.10’11

I measure neoliberal ideology in two ways, each addressing a different aspect of the ideology.12 First, I account for states that engage in relatively little public welfare spending with a binary measure (U.S. Census Bureau various years).13 These low spending states reflect a neoliberal focus on less government involvement in society, especially as it relates to social safety net programs. Secondly, I also assess a state s opposition to regulation by including a measure of which states have adopted a right-to-work policy by a given year. Right-to-work policies make unionization more difficult in a state and are consistent with broader neoliberal ideology preferences regarding market supremacy. I do not expect that right-to-work policies themselves are associated with renewable energy policy, but in­ stead they serve as an indicator of states that seek to minimize state interference in the market.14 As right-to-work policies are associated with an inclination against regulations more broadly, and are not restricted to environmental policy, with this measure included I am better positioned to observe how political culture in general influences the policy orientation a state will pursue when adopting policies. I measure right-to-work policies as a time-variant binary variable for if a state has adopted a right-to-

10 Another commonly used measure of environmental culture is single-year scorecards of state polices produced by other environ­ mental interest groups. These measures might be more directly relevant to state politics, but they suffer two flaws compared to LCV scores. First, they are not longitudinal. Few environmental groups have existed continuously since the 1970s, and thus most scorecards cover only a single year from relatively late in the time period under study. Secondly, many of these indices draw on the exact policy instruments I am predicting adoption of. In order to avoid causality problems and to capture a time- variant measure I prefer the LCV scores.

I I One other measure examined was the share of state spending devoted to natural resource conservation, promotion, and devel­ opment (U.S. Census Bureau various years). The definition of this spending is broad, including everything from forest conserva­ tion to flood control efforts. This measure was never significant and did not change any of the results presented. Given the wide definition and lack of significance, and out of a desire to conserve degrees of freedom in the models presented, I omit this meas­ ure from presentation.

12 Another measure, examining the degree to which a state’s tax environment is regressive, was tested but ultimately not included in this article. I assembled data on state tax revenue to assess the degree to which states drew their revenue from tax mechanisms that did not offer variable rates based on income. These mechanisms include sales tax and licenses or user fees, compared to more progressive income or capital gains taxes. I focus on a state’s regressive tax environment rather than something like tax cuts, to ease comparisons between states. Many states already have no income tax, leaving politicians with nothing to cut. Instead, I focus on the underlying issue that neoliberal ideology drives towards, flattening the overall tax rate paid by individuals. For that reason I believe this measure of regressive tax environments captures at the state level a similar underlying concept to federal tax cuts. This measure was not significant in bivariate analysis and proved highly correlated with right-to-work status and low welfare spending. As such I omit it from the presented models.

13 Low welfare spending is defined as less than 20 percent of state revenues, roughly the average of state welfare spending during the time period under study. Slight modifications to this cut point do not substantively change the results presented. A binary measure is used to avoid the skewed distribution of this measure and to facilitate comparison to the right-to-work measure and over time.

14 Testmg an additional item, Elizar’s (1972) three-political culture model suggests that the right to work measures capture much of the same variation as what Elizar calls traditional political culture.

294 . Vasseur

work policy by a given year using data drawn from the National Council of State Legislators (NCSL 2014).15 These measures capture the two most prominent aspects of neoliberal ideology, a limited welfare state and a preference for a market-based industrial policy.

Logic of Analysis To assess the impact of these state-level factors on the number of policy instruments of the various

types adopted by a state I fit a series of random effects longitudinal Poisson regression models.16 For each outcome three models are fit. The first includes economic factors and political interests. The se­ cond model adds controls for institutional and cultural factors. The third model controls for the num­ ber of policy instruments of the other type adopted by a state. All models are fit with robust standard errors to correct for possible level-two overdispersion (Rabe-Hesketh and Skrondal 2012). In each model time is specified using indicator variables for decade.17

RESULTS AND DISCUSSION The results of the models predicting tax incentive policy instruments, presented in Table 3, are con­ sistent with a view of energy policy as economic and political interest based. Both higher levels of an­ nual income and increased environmental movement organization density predict adopting more incentive-based policy instruments. On the other hand, states that produce fossil fuels are predicted to adopt significantly fewer incentives. These results are consistent with the addition of the institu­ tional and cultural factors in Model 2, none of which prove significant for this outcome. This model is what would be expected from the energy policy literature; states that have higher levels of affluence and more environmental movement organizations are motivated to act, and those that produce fossil fuels are less likely to enact policies that challenge their current economic interests. A model based on economic and political interests, as dominates the broader energy policy literature, fits for incen­ tives, but this is not entirely the case for mandate-based policies.

Similar models are presented in Model 4 and Model 5 in Table 4. The pattern of results observed across these models is clear, but no longer entirely consistent with an economic and political interest- based account of energy policy. As with tax incentives, states with more environmental movement or­ ganizations and higher income adopt more regulatory mandates. Additionally, increasing LCV scores increases regulatory mandate adoption. On the other hand, having a right-to-work law and/or having low welfare spending significantly reduces the odds a state will adopt a mandate-based instrument. This indicates that states that have adopted a set of policies consistent with a neoliberal ideology are dramatically less likely to enact policies that intervene between the consumer and electricity produ­ cers when compared to other states. This effect is independent of a state’s more general environmen­ tal orientation.

15 Most right-to-work laws were adopted well before 1970. In fact, many states adopted them in the 1940s and 50s. Only six states adopted these laws during the period under study, and most of those did so well before adopting any renewable energy policies. As such, changing right-to-work to a time-varying binary variable does not change the results presented.

16 A Mundlak test for the suitability of random coefficients provides no evidence against using random coefficients: tax incentives (Chi2 = 3.96, df = 6, p = .5556); regulatory mandates (Chi2 = 6.18, df = 6, p = .2894). Additionally, examinations of the dis­ crete change in probabilities of a state-year having a zero, one, or two plus count for each outcome is not seen across the signifi­ cant independent variables. This suggests that there is not a substantively different process for going from a zero to a one count as there is for further policy action, thus I use the count models in place of event history models of these transitions.

17 Other specifications of time examined including omitting it entirely, using a linear specification, and using a second order poly­ nomial specification. Changing the specification of time in these models changes the substantive conclusion for only the Democratic power covariate in the mandate models. With some specifications of time and some alternative specifications of Democratic power (for example separating governors from legislative chambers, using percentages of Democrats) individual co­ efficients are significant at the .05 level. Given the dependence of this result on specification I believe this is an issue of colinear­ ity between predictors. As there is no evidence of a significant relationship between Democratic power and either type of policy adoption with other predictors omitted, regardless of the specification of time, and that removing it from the model does not change any of the other effects observed, I maintain the traditional 0 to 3 scale in presentation.

Determinants of the Renewable Energy Policies of U.S. States . 295

Table 3. Odds Ratios from Random Effects Longitudinal Poisson Models of Tax Incentive Adoption

(1) Base Model

(2) Institutions and Culture

(3) Policy Portfolio

Policy portfolio Total m andate policies .894

( -1 .3 2 ) Econom ic factors H ousehold incom e0 1.018** 1.018** 1.019**

(2.94) (2.89) (3.14) Fossil fuel production .426** .538+ .521*

( -2 .6 0 ) ( - 1 .9 1 ) ( -2 .0 8 ) Average wind speed6 1.029 1.035 1.041

(.28) (.31) (.36) Political interests D em ocratic pow erc 1.007 .998 1.019

(.11) ( - .0 3 ) (.34) EM O density 1.075* 1.069* 1.066*

(2.29) (2.30) (2.36) Political institutions Direct democracy .638 .649

( - 1 .3 0 ) ( - 1 .2 4 ) Legislative professionalism 1.000 1.001

( - .0 5 ) (.24) G overnor power .980 .964

( - .0 9 ) ( - .1 6 ) Neoliberal affinity Right-to-work 1.081 1.027

(.17) (.06) Low welfare spending .798 .821

( - 1 .3 8 ) ( -1 .3 5 ) Environm ental culture LCV scores 1.003 1.003

(.96) (1.19) State-level intercept .174 .111 .118

(.60) (.37) (.41) State-year intercept 1.091 1.051 1.061 BIC 3,467.716 3,501.051 3,498.456 N = 2,103 state-years

Notes: Exponentiated coefficients; z statistics in parentheses, models fit with robust standard errors. Controls for decade included in model (not shown). “Median annual household income in thousands of 2011 dollars. Miles per hour.

Cham bers of state legislator plus governorship controlled by the Democratic Party. ip < .10 *p < .05 **p < .01 ***p < .001 (two-tailed tests)

296 Vasseur

Table 4. Odds Ratios from Random Effects Longitudinal Poisson Models of Regulatory Mandate Adoption

(4 ) (5) (6) Base Model Institutions and Culture Policy Portfolio

Policy portfolio Total incentive policies 1.245***

(4.19) Economic factors Household income11 1.053*** 1.063*** 1.059***

(3.85) (4.62) (4.00) Fossil fuel production .319+ .655 .642

(-1 .8 3 ) (- .8 0 ) (- .8 4 ) Average wind speed*" 1.102 1.164 1.150

(.62) (.91) (.91) Political interests Democratic powerc 1.050 .997 .957

(1.35) (- .0 9 ) (-1 .1 6 ) EMO density 1.044** 1.036* 1.038*

(2.81) (1.98) (1.97) Political institutions Direct democracy .939 .987

(- .1 2 ) (- .0 2 ) Legislative professionalism 1.002 1.000

(.50) (.02) Governor power .584 .569

(-1 .2 9 ) (-1 .4 3 ) Neoliberal affinity Right-to-work .321* .303*

(-2 .2 3 ) (-2 .3 7 ) Low welfare spending .502*** .515***

(-4 .2 0 ) (-4 .2 0 ) Environmental culture LCV scores 1.006** 1.004*

(2.83) (2.07) State-level intercept 1.256 .755 .647

(3.44) (1.91) (1.70) State-year intercept 1.874 1.459 1.382 BIC N = 2,103 state-years

1,732.840 1,737.512 1,730.746

Notes: Exponentiated coefficients; z statistics in parentheses, models fit with robust standard errors. Controls for decade included in model (not shown). ‘'Median annual household income in thousands of 2011 dollars. ^Miles per hour. cChambers of state legislator plus governorship controlled by the Democratic Party. ip < .10 *p < .05 **p < .01 ***p < .001 (two-tailed tests)

Determinants of the Renewable Energy Policies of U.S. States . 297

Table 5. Predicted Number of Policies Adopted by Policy Type, Fossil Fuel Production, and Neoliberal Status0

Total Incentives Mandates Difference

Neither 3.40 2.09 1.31 .78 Neoliberal 2.09 2.00 .09 1.91 Fossil fuel only 2.11 1.04 1.07 -.03 Both 1.06 .99 .07 .92

“Predictions based on institutional and cultural models for each outcomes (Models 2 and 5) with all other covariates held at their mean. Neoliberal states are those with right-to-work laws and low welfare spending. Changing levels of environmental movement organizations and income changes the overall level of activity, but does not change the ordering of categories or the relative difference between incentive and mandate policies. Predictions are for the final year of data (2012) to model ending conditions for states.

The final models in each table, Models 3 and 6, present the relationship between these two out­ comes as net of the predictors. These policies are not unrelated, nor does adopting one kind of policy decrease the odds of adopting a policy of the other type. Instead there is a significant increase in the odds of a state adopting a regulatory mandate for each tax incentive it has already adopted, but the re­ verse does not yield a significant relationship. This suggests that these policy orientations are not zero sum, especially when seen in their historically contingent context. Tax incentive policies first ap­ pear in the mid-1970s, while regulatory mandates were not widely adopted until almost 20 years later. Understanding this result requires a reorientation of views of policy sequencing away from a focus on policy instrument type and towards broader policy orientations. It is certainly true that conditions in some states lead to a noted preference for incentives or mandates, but other states adopt policies based on a commitment towards increasing renewable energy generation. These states are not ori­ ented toward a specific type of policy instrument but instead toward a broader goal of increasing re­ newable energy generation by any means available. They adopt incentives as they first become available and later add mandates as those policies develop. From a policy content point of view, these states have adopted a mixed approach to renewable energy policy.

This mixed approach to renewable energy policy is a distinct orientation unto itself, but is not one that could be identified by examining idiosyncratic policy adoption or only studying the first policy instrument adopted. This mixed approach is reflected by the significant relationship in Model 6. States are more likely to engage in mandate-based policies if they already have incentive-based poli­ cies. Therefore, two dominant strategies of action can be observed: some states adopt a mixed ap­ proach while others adopt a primarily (or exclusively) incentive-based orientation. What state-level factors distinguish these active states from those that take relatively few policy actions?

State policy action can be explained by traditional political interests and economic factors. Renewable energy policies are adopted in states in which they do not threaten existing interests. As seen in the prior results, more affluent states with more environmental movement organizations and that do not produce fossil fuels will adopt more renewable energy policies. These factors are consist­ ent for both incentives and mandates. This suggests that similar conditions spur states to policy ac­ tion in general, but what factors account for differences in the type of policy a state adopts?

The content of state policy is best explained by political culture in the form of an affinity for a neo­ liberal ideology beyond energy policy. Table 5 presents the expected policies adopted by states as of 2012, separated by policy type and by all possible combinations of fossil fuel production and neo­ liberal status. This table illustrates two key points. First, states that do not produce fossil fuels adopt more policies regardless of neoliberal status. Secondly, neoliberal states show the largest difference between their incentive and mandate-based policies. States that are not neoliberal and do not pro­ duce fossil fuels take a more balanced approach. Neoliberal states, on the other hand, skew heavily to­ wards incentive-based policies regardless of their fossil fuels status. States that do not produce fossil

298 • Vasseur

fuels but are neoliberal adopt an average of two incentive policies, no different from similar non­ neoliberal states. They differ sharply when it comes to mandates, from an average of 1.3 policies to less than a tenth of a policy. Thus even if it is a state’s economic characteristics and political interests that determine how many policies a state is likely to adopt, it is that state’s political culture that deter­ mines what kind of policies they enact. This divide reinforces the need to examine broader policy orientations, not simply individual policy adoption, in order to account for all facets of the policy adoption process.

CONCLUSION This project contributes to our understanding of policy adoption by examining a state’s broader port­ folio of policies rather than idiosyncratic policy adoption. This shift demonstrates that traditional eco­ nomic and electoral interest-based factors are excellent predictors of how many policies a state is likely to adopt, but do far less to explain what kind of policies a state will enact. To understand policy content, a broader conception of politics that includes institutional and cultural factors is required. Adding these factors clarifies the process by which states select the content of their policies. Specifically, even when accounting for a state’s general environmental orientation, a state’s affinity for neoliberal ideology by way of having a right-to-work law or low welfare spending is associated with adopting an incentive-only orientation. I show that states that are adverse to market intervention in union organizing carry that aversion over to the largely unrelated area of renewable energy politics. Given the centrality of the incentive versus mandate debate in contemporary American political dis­ course, using right-to-work laws as a quantifiable and easily comparable measure of political culture regarding regulation has the potential to be informative for future research across a variety of policy domains. Be it social program spending, the privatization of education, or the establishment of local healthcare markets, there are reasons to suspect that this measure will prove fruitful for analysis be­ yond renewable energy policy.

In addition, viewing policies as a result of an overall policy orientation lends clarity to the diversity of actions undertaken by U.S. states in the domain of renewable energy policy. States adopt a variety of policies and do so with little consistency regarding individual policy types. It is rarely the case that a state adopting a policy of one kind will necessarily result in adoption of a related policy or not hav­ ing any policies of a different type. This could lead to a view of policy adoption as random or idiosyn­ cratic if the focus of analysis remains on the individual policies. If analysis instead focuses on the entire portfolio of policies, as I have done, state actions have an overall consistency with a broader policy orientation. In the contemporary United States, states adopt one of two strategies of action, ei­ ther incentive based or a mixed approach. Relatively few states adopt a mandate-only approach. If states are going to act in this domain they will almost certainly adopt incentive-based policies, and some will add mandates as well. This is likely based on the historical sequencing of policy adoption, with tax incentives first adopted in the 1970s and regulatory mandates not seeing widespread adop­ tion until the late 1990s. This insight further reinforces the continued importance of understanding the relationship between strategies based on tax incentive and regulatory mandate policies in contem­ porary American politics. Examining a state’s general policy orientations expands the definitions of consistency in state policy action to better reflect political reality.

The approach in this article demonstrates the ability to examine what kinds of policies a state adopts, but the question remains: why does the type of policy matter? Policy type matters for two reasons. First, ignoring policy type in favor of only policy action obscures the complex nature of polit­ ical reality. Prior research has noted this across other policy domains. For example, Jacob Hacker (2002) demonstrates that the United States is no longer a welfare laggard if more varieties of policies are considered social policy. In this case a focus on policy content redefines one of the central moti­ vating questions for policy scholars of the past decades. In order to gain an accurate picture of the policy process, it is vital to understand both policy action and the content of the adopted policies.

Determinants of the Renewable Energy Policies of U.S. States . 299

Another impact of policy type is seen when examining real-world efficacy. It is still unclear if the renewable energy policies studied here will actually be effective at increasing renewable energy gener­ ation or reducing carbon dioxide emissions in the states that adopt them (Carley 2009; Prasad and Munch 2012). The existing evidence on effectiveness clusters among the most coercive regulatory mandates. Consumption taxes have been shown to be more effective than other renewable energy policies (Prasad 2010; Prasad and Munch 2012). If these results hold as more state policies come into effect, it suggests that incentives alone are not enough; a top-down, mandate-based approach is required to meaningfully address the problems these policies were designed to solve.

Beyond a focus on policy portfolios and orientations of action, this study also contributes to a broader literature on fiscal sociology by expanding its focus on tax incentives to a policy domain that has received less attention from scholars. As fiscal sociology continues to develop, it will expand be­ yond the confines of social policy and into other domains in which tax incentives are favored instru­ ments of promoting social change. In applying the theories from fiscal sociology to renewable energy policy this project does just that. In the current American political climate, incentive-based policies are the only actions taken at a federal level regarding renewable energy, and are preferred across a range of policy domains. These insights into how states decide on the content of their policies are vital to understanding the future politics of climate change in the U.S. context. As renewable energy policy and climate change in general continue to be debated in the United States, research on the subject must expand and incorporate insights from other political fields in order to remain relevant.

REFERENCES Andrews, Kenneth T. and Bob Edwards 2005. The Organizational Structure of Local Environmentalism.” Mobilization

10(2):213-34. Baumgartner, Frank R. and Bryan D. Jones 1993. Agendas and Instability in American Politics. Chicago: University of

Chicago Press. Beland, Daniel and Robert Henry Cox 2011. Ideas and Social Politics in Social Science Research. New York: Oxford

University Press. Brulle, Robert, Liesel Hall Turner, Jason Carmichael, and J. Craig Jenkins. 2007. “Measuring Social Movement

Organization Populations: A Comprehensive Census of U.S. Environmental Movement Organizations.” Mobilizations 12(3): 195-211.

Campbell J. L. and O. K. Pedersen, eds. 2001. The Rise of Neoliberalism and Institutional Analysis. Princeton, NJ: Princeton University Press.

Campbell, John L. 1988. Collapse of an Industry: Nuclear Power and the Contradictions of U.S. Policy. Ithaca, NY and London, UK: Cornell University Press.

------. 2002. “Ideas, Politics, and Public Policy.” Annual Review of Sociology 28(l):21-38. Carley, Sanya. 2009. “State Renewable Energy Electricity Policies: An Empirical Evaluation of Effectiveness” Energy

Policy 37:3071-81. Chandler, Jess. 2009. “Trendy Solutions: Why Do States Adopt Sustainable Energy Portfolio Standards?” Energy Policy

37:3274-81. Coley, Jonathan S. and David J. Hess 2012. Green Energy Laws and Republican Legislators in the United States.”

Energy Policy 48(2012):576-83. Council of State Government. 1979-2012. Book of the States. Volumes 23-44. Chicago: Council of State Governments. Daley, Dorothy M. and James C. Garand 2005. Horizontal Diffusion, Vertical Diffusion, and Internal Pressure in State

Environmental Policymaking, 1989-1998.” American Politics Research 33(5):615-44. Database of State Incentives for Renewables & Efficiency (DSIRE). 2012. “Rules, Regulations & Policies for Renewable

Energy. Retrieved May 8, 2015 (www.dsireusa.org/). Deyette, Jeff, Steve Clemmer, and Deborah Donovan 2003. Plugging in Renewable Energy: Grading the States.

Washington, DC: Union of Concerned Scientists. Dobbin, Frank. 1994. Forging Industrial Policy: The United States, Britain, and France in the Railway Age. New York:

Cambridge University Press. Elizar, Daniel J. 1972. American Federalism: A View from the States. New York: Thomas Y. Crowell Company. Energy Information Administration. 2015 “U.S. Energy Generation by Source.” Retrieved May 8, 2015 (www.eia.gov/

tools/faqs/faq.cfm?id=427&t=3).

300 . Vasseur

Fowler, Luke and Joseph Breen 2013. “The Impact of Political Factors on States’ Adoption of Renewable Portfolio Standards.” The Electricity Journal 26(2):79-94.

Gale Research. Inc. 1956-2010. The Encyclopedia of Associations, Volume 1, National Organizations of the U.S. Detroit, MI: Gale Research Co.

Hacker, Jacob S. 2002. The Divided Welfare State: The Battle over Public and Private Social Benefits in the United States. Cambridge, UK: Cambridge University Press.

Howard, Christopher. 1997. The Hidden Welfare State: Tax Expenditures and Social Policy in the United States. Princeton, NJ: Princeton University Press.

----- . 2007. The Welfare State Nobody Knows: Debunking Myths About U.S. Social Policy. Princeton, NJ: Princeton University Press.

----- . 2009. “Maxing Taxes the Life of the Party.” Pp. 86-101 in The New Fiscal Sociology: Taxation in Comparative and Historical Perspective, edited by I. W. Martin, A. K. Mehrotra, and M. Prasad Cambridge, UK: Cambridge University Press.

Huang, Ming-Yuan, R. Janaki, R Alavalapati, Douglas R Carter, and Matthew H. Langholtz 2007. “Is the Choice of Renewable Portfolio Standards Random?” Energy Policy 35:5571-75.

Initiative and Referendum Institute. 2013, “State-by State List of Initiative and Referendum Provisions.” Retrieved May 8, 2015 (www.iandrinstitute.org/statewide_i%26r.htm).

Jasper, James M. 1990. Nuclear Politics: Energy and the State in the United States, Sweden, and France. Princeton, NJ: Princeton University Press.

Johnson, Erik W. 2006. “Changing Issue Representation among Major United States Environmental Movement Organizations.” Rural Sociology 7l(l):132-54.

----- . 2008. “Social Movement Size, Organizational Diversity and the Making of Federal Law.” Social Forces 86(3): 1-28. Johnson, Erik W., Jon Agnone, and John D. McCarthy 2010. “Movement Organizations, Synergistic Tactics, and

Environmental Public Policy.” Social Forces 88(5):2267-92. Johnson, Erik W. and Scott. Frickel 2011. “Ecological Threat and the Founding of U.S. National Environmental

Movement Organizations, 1962-1998.” Social Problems 58(3):305-29. Kingdon, John W. [1995] 2004. Agendas, Alternatives, and Public Policies. New York: HarperCollins. Krupnikov, Yanna and Charles Shipan 2012. “Measuring Gubernatorial Budgetary Power: A New Approach.” State

Politics and Policy Quarterly 12(4):438-55. League of Conservation Voters (LCV). 2012. “National Environmental Scorecard.” Retrieved May 8, 2015 (www.lcv.

°rg/). Lieberman, Robert C. 2002. “Ideas, Institutions, and Political Order: Explaining Political Change.” American Political

Science Review 96:697-711. Lowry, William R 2008. “Disentangling Energy Policy from Environmental Policy.” Social Science Quarterly

89(5): 1194-211. March, James G. and Johan P. Olsen 1989. Rediscovering Institutions: The Organizational Basis of Politics. New York:

The Free Press. Martin, Andrew W., Frank R Baumgartner, and John D. McCarthy 2006. “Measuring Association Populations

Using the Encyclopedia of Associations: Evidence from the Field of Labor Unions.” Social Science Research 35(3):771-78.

Martin, Isaac William, Ajay K. Mehrotra, and Monica Prasad 2009. The New Fiscal Sociology: Taxation in Comparative and Historical Perspective. Cambridge, UK: Cambridge University Press.

Minkoff, Debra C. 1997. “The Sequencing of Social Movements.” American Sociological Review 62(5):779-99. ------. 1999. “Bending with the Wind: Strategic Change and Adaption by Women’s and Racial Minority Organizations.”

American Journal of Sociology 104(6):1666-703. National Conference of State Legislatures (NCSL). 2014. “Right-to-Work States.” Retrieved May 8, 2015 (www.ncsl.

org/research/labor-and-employment/right-to-work-laws-and-bills.aspx). Padamsee, T. J. 2009. “Culture in Connection: Re-Contextualizing Ideational Processes in the Analysis of Policy

Development.” Social Politics: International Studies in Gender, State & Society 16(4):413-45. Pierson, Paul. 2000. “Increasing Returns, Path Dependence, and the Study of Politics.” American Political Science Review

94:251-67. Podobnik, Bruce. 2006. Global Energy Shifts: Fostering Sustainability in a Turbulent Age. Philadelphia: Temple University

Press. Prasad, Monica. 2006. The Politics of Free Markets: The Rise of Neoliberal Economic Policies in Britain, France, Germany

and the United States. Chicago: The University of Chicago Press.

Determinants of the Renewable Energy Policies of U.S. States . 301

----- . 2010. “Taxation as Regulatory Tool: Lessons from Environmental Taxes in Europe.” Pp. 363-90 in Governments and Markets: Toward a New Theory of Regulation, edited by E. J. Balleisen and D. A. Moss. Cambridge, UK: Cambridge University Press.

------. 2011. "Tax ‘Expenditures’ and Welfare States: A Critique.” The Journal of Policy History 23(2):2Sl-66. Prasad, Monica and Steven Munch 2012. “State-Level Renewable Electricity Policies and Reduction in Carbon

Emissions.” Energy Policy 45:237-42. Rabe-Hesketh, Sophia and Anders Skrondal 2012. Multilevel and Longitudinal Modeling Using Stata: Volume II

Categorical Responses, Counts, and Survival. College Station, TX: Stata Press. Ringquist, Evan J. 1993. Environmental Protection at the State Level. Armonk, NY: M. E. Sharpe. Sine, Wesley D. and Brandon H. Lee 2009. “Tilting at Windmills? The Environmental Movement and the Emergence

of the U.S. Wind Energy Sector.” Administrative Science Quarterly 54:123-55. Skrentny, John David. 1996. The Ironies of Affirmative Action, Politics, Culture, and Justice in America. Chicago:

University of Chicago Press. Somers, Margaret R and Fred F. Block 2005. “From Poverty to Perversity: Ideas, Markets and Institutions over 200

Years of Welfare Debate.” American Sociological Review 70:260-87. Squire, Peverill. 2007. “Measuring State Legislative Professionalism: The Squire Index Revisited.” State Politics and

Policy Quarterly 7:211-27. Steensland, Brian. 2008. The Failed Welfare Revolution: America’s Struggle over Guaranteed Income Policy. Princeton, NJ:

Princeton University Press. U.S. Census Bureau. 1970, 1975, 1980, 1990, 1995, 2000, 2010. “Annual Survey of State Government Finances.”

Retrieved May 8, 2015 (www.census.gov/govs/state/historical_data.html). ------. 2012. Statistical Abstract of the United States Congress, 131st Edition. Retrieved May 8, 2015 (www.census.gov/

compendia/statab/>). U.S. Department of Energy. 2010. “States Energy Data System.” Retrieved May 8, 2015 (www.eia.gov/state/seds/). U.S. Environmental Protection Agency (EPA). 2014. “Clean Power Plan.” Retrieved May 8, 2015 (www.epa.gov/clean

powerplan/clean-power-plan-exis ting-power-plants). Vachon, Stephan and Fredric C. Menz 2006. The Role of Social, Political, and Economic Interests in Promoting State

Green Electricity Policies.” Energy Policy 9(7):652-62. Van Dyke, Nella and Sarah A. Soule 2002. “Structural Social Change and the Mobilizing Effect of Threat: Explaining

Levels of Patriot and Militia Organizing in the United States.” Social Problems 49(4):497-520. Vasi, Ion Bodgan. 2006. “Organizational Environments, Framing Processes, and the Diffusion of the Program to

Address Global Climate Change among Local Governments in the United States.” Sociological Forum 2l(3):439-66. ----- . 2009. “Social Movements and Industry Development: The Environmental Movement’s Impact on the Wind

Energy Industry.” Mobilization: An International Journal 14(3):315-36. ------. 2011. Winds of Change: The Environmental Movement and the Global Development of the Wind Energy Industry.

New York: Oxford University Press. Vasi, Ion Bodgan and David Strang 2009. “Civil Liberty in America: The Diflbsion of Municipal Bill of Rights

Resolutions after the Passage of the U.S. Patriot Act.” American Journal of Sociology 114(6): 1716-64. Walker, Edward T., John D. McCarthy, and Frank Baumgartner 2011. “Replacing Members with Mangers? Mutualism

among Membership and Nonmembership Advocacy Organizations in the United States.” American Journal of Sociology 116(4):1284-337.

Yi, Hongtao and Richard C. Feiock 2012. “Policy Tools Interactions and the Adoption of State Renewable Portfolio Standards.” Review of Policy Research 29(2): 193-206.

Zelinsky, Edward A. 1993. “James Madison and Public Choice at Gucci Gulch: A Procedural Defense of Tax Expenditures and Tax Institutions.” The Yale Law Journal 102(5): 1165-207.

------. 2004. “Do Tax Expenditures Create Framing Effects? Volunteer Firefighters, Property Tax Exemptions, and the Paradox of Tax Expenditure Analysis.” Virginia Tax Review 24:798-825.

Copyright © 2016 The Society for the Study of Social Problems.