• Footnotes
• Chapter
• University of Michigan Press
• View Citation
• Additional Information

. A manifestation of the federal government’s dysfunction, and arguably a driver of further plummeting faith among the public in the ability of the federal government to perform even basic tasks, is the increasing inability of the federal government to even adopt a semblance of meaningful policy. Afflicted by party polarization (McCarty 2019) and enmeshed in a culture of brinkmanship, the federal government is unable to effectively produce policy. One key area where such federal government inability has been exhibited is with respect to green energy policy. Although the planet’s changing climate is largely acknowledged by the scientific community (Gross 2018) and by major transnational institutions (United Nations 2021), and although a warming climate could wreak economic, ecological, and political havoc on the United States (Overpeck and Udall 2020; Koubi 2019), the federal government has not been able to substantially commit to policies that favor the emergence of cleaner sources of energy over traditional fossil fuels, especially in the electricity sector where fossil fuels and particularly dirtier coal retain significant utilization (Allison and Parinandi 2020).1 Given that the United States accounts for much of the global production of carbon dioxide emissions (Hickel 2020), the American federal government’s inaction regarding fostering the use of green or clean energy has been recognized as a pressing international problem (Vigilone 2020).

. In this book, I answer the above questions by delving into a meticulous study of state-level inventing and borrowing with respect to renewable portfolio standards, or RPS. An RPS generally specifies that electricity utility companies operating in a given state procure some of the electricity they provide to end-use consumers in that state from renewable sources. Appearing on the scene in the last quarter of the twentieth century and disseminating through a wide swath of states, RPSs have become the preeminent way in which the states have tried to spark renewable energy development (Rabe 2004, 2007; Carley and Miller 2012; Carley, Nicholson-Crotty, and Miller 2017; Parinandi 2020; Parinandi et al. 2020; and Stokes 2020). Unlike policies such as carbon taxes that have been shown to have limited appeal across the states, RPSs have been described as “climate policy that can actually win” and have been described by energy economist Michael Greenstone as “the biggest [carbon policy] we have in this country” (Meyer 2019). RPSs have also been a state-level rather than federal phenomenon, meaning that we can study and distinguish state inventing and borrowing in the absence of strong federal impulses that may distort when such inventing and borrowing occurs.10 Lastly, RPSs have been the site of tremendous state-level diversity in policy adoption, meaning that we actually observe instances of state invention and borrowing that can be dissected and analyzed.11 Simply put, focusing on the RPS area permits us to study state responses to federal inaction in Brandeis’s mold—that is, inventing by adopting novel policy or borrowing by adopting policy that has been tested in other states—in what is possibly one of the most important problems facing humanity in the twenty-first century.

. One measure of ideology—government ideology—captures the ideological comportment of state governmental officials on a conservative-liberal spectrum. Berry et al. (1998, 2010) construct a governmental ideology score for each state/year by assuming first that the ideological comportment of a state’s legislators of a given party can be gleaned from the ideological comportment of that state’s national legislators who belong to the same party.100 The authors also assume that the ideological comportment of a state’s governor can be gleaned from the estimated ideological comportment of a state’s legislators who belong to the same party as that state’s governor.101 Lastly, the authors use a weighted sum combining the ideological scores of major and minor party delegations from each chamber of a state’s legislature along with that of the governor to calculate a governmental ideological score for that state.102 In figure 5, Page 97 →I plot a state’s average government ideology calculated over the full time span of the study.

. One measure of ideology—government ideology—captures the ideological comportment of state governmental officials on a conservative-liberal spectrum. Berry et al. (1998, 2010) construct a governmental ideology score for each state/year by assuming first that the ideological comportment of a state’s legislators of a given party can be gleaned from the ideological comportment of that state’s national legislators who belong to the same party.100 The authors also assume that the ideological comportment of a state’s governor can be gleaned from the estimated ideological comportment of a state’s legislators who belong to the same party as that state’s governor.101 Lastly, the authors use a weighted sum combining the ideological scores of major and minor party delegations from each chamber of a state’s legislature along with that of the governor to calculate a governmental ideological score for that state.102 In figure 5, Page 97 →I plot a state’s average government ideology calculated over the full time span of the study.

. One measure of ideology—government ideology—captures the ideological comportment of state governmental officials on a conservative-liberal spectrum. Berry et al. (1998, 2010) construct a governmental ideology score for each state/year by assuming first that the ideological comportment of a state’s legislators of a given party can be gleaned from the ideological comportment of that state’s national legislators who belong to the same party.100 The authors also assume that the ideological comportment of a state’s governor can be gleaned from the estimated ideological comportment of a state’s legislators who belong to the same party as that state’s governor.101 Lastly, the authors use a weighted sum combining the ideological scores of major and minor party delegations from each chamber of a state’s legislature along with that of the governor to calculate a governmental ideological score for that state.102 In figure 5, Page 97 →I plot a state’s average government ideology calculated over the full time span of the study.

. Figure 5 reveals observations that conform to conventional wisdom but also defy it. It is not surprising that Maryland and Massachusetts have liberal state governments or that Utah and Idaho have conservative state governments. However, California is not as liberal as one might expect, and West Virginia has historically had a very liberal state government.103

. Recently, scholars studying the diffusion of policy across the U.S. states Page 98 →have uncovered a role for electoral vulnerability in influencing adoption. Policy diffusion scholars have consistently argued that a key reason diffusion occurs is that electorally vulnerable policymakers in later adopting states free ride off of the experiences of pioneering states and tell constituents that they will re-create desirable outcomes from those pioneering states (Volden 2006; Volden, Ting, and Carpenter 2008; Shipan and Volden 2014; Butler et al. 2017).104 This line of argument establishes a possible link between electoral vulnerability and borrowing—which is by definition the act of replicating an earlier state’s policy action—but what link might exist between electoral vulnerability and invention? Do electorally vulnerable lawmakers see invention as a “Hail Mary” and do so in hopes of rejuvenating their reelection chances? Evaluating whether electoral vulnerability differentially impacts invention and borrowing would help us determine whether lawmakers’ desire for reelection leads to novel policy creation or the copying of existing policy.

. I make two main arguments in this chapter. First, I contend that government ideology corresponds with legislative invention more than it corresponds with legislative borrowing.105 And second, I contend that electoral vulnerability corresponds with legislative borrowing more than it corresponds with legislative invention. I also evaluate whether legislative professionalism influences invention differentially from borrowing but fail to find meaningful evidence with respect to it mattering and therefore do not elaborate on this relationship outside of speculating as to why a negative finding might exist in the conclusion of the chapter.

. I elaborate on my claim by appealing to the burgeoning literature on ideological diffusion. Work on diffusion has revealed that governments are more likely to learn from and adopt policies when such policies have already been implemented by other ideologically similar governments compared to when such policies have been implemented by ideologically dissimilar governments (Grossback, Nicholson-Crotty, and Peterson 2004; Gilardi 2010; Carley, Nicholson-Crotty, and Miller 2017; Parinandi, Langehennig, and Trautmann 2020). In ideologically based diffusion, beliefs override evidence in the sense that a borrowing government pays attention to the source (“who adopts”) rather than the message (“what is being adopted” (Butler et al. 2017). An untested or novel policy feature does not have analogues in other states, leading to the situation where an inventing government cannot observe other ideologically similar governments across the states to figure out whether to adopt a novel policy feature. But yet, with respect to invention, ideology can override evidence with regard to what is being adopted in the sense that different issue areas (or domains) of policy-making have different comportments ideologically. Regulations concerning abortion, for example, are associated with the ideological right (Kreitzer 2015) while regulations concerning clean energy are associated with the ideological left (Potrafke 2010). Compared to an ideologically moderate legislature, an ideologically extreme legislature may be more willing to overlook a lack of evidence and adopt a novel policy feature if that policy feature falls in a domain that is congruent with the more extreme legislature’s worldview. Ultimately, while all legislatures at some point utilize cues in lieu of evidence to facilitate policy-making (Lupia 1994), ideologically extreme legislatures are more willing to rely on cues—in this case, cues about whether crafting policy in a particular domain corresponds to their ideological comportment—and ignore a lack of evidence during policy adoption than are their ideologically moderate peer legislatures.106

. And what of the relationship between government ideology (governmental liberalism in our case, since we are dealing with the RPS policy domain) and borrowing? I hypothesize that more extreme government ideology plays a less prominent role in influencing borrowing than it does invention. Unlike in the case of invention, borrowed policy features have a track record and that track record can be accessed in order to inform adoption decisions. Although the zeal of ideologically predisposed legislatures (liberal legislatures given that we are analyzing renewable energy Page 103 →legislation) can lead to the ignoring of evidence during policy adoption, more moderate legislatures possess less zeal with respect to the cause of the policy and therefore will rely more heavily on evidence in guiding adoption decisions and may ultimately show a comparatively reduced preference for inventing during policy adoption. Additionally, compared to an ideologically liberal legislature, an ideologically moderate legislature will include more individual (e.g., in this case, conservative) legislators who may be predisposed against adopting RPS policy, and these individuals will give an especially important role to evidence in helping them potentially override their own predispositions against the policy domain and adopt RPS policy features (Butler et al. 2017). This suggests that ideological extremism (in this case, liberalism) should explain invention more than it explains borrowing, where the existence of a track record can ostensibly bring legislators of more diverse ideological persuasions into the policy-making enterprise.107

. The importance of an observable track record—in other words, evidence—to borrowing processes suggests how electoral concerns could relate to invention and borrowing during policy feature adoption. The typical member of a state legislature is accountable to the state’s median voter and tries to show that he or she is acting in the interest of this voter. Legislative members seek reelection, and one way that legislative members signal that they are serving the median voter is by adopting policies—and policy features—that they claim best match the interests of the median Page 104 →voter (Mayhew 1974).108 In deciding to adopt a policy, a legislature pays attention to each possible policy feature that, taken together with all other policy features, makes up a policy, and the legislature can invent by adopting a novel policy feature or borrow by adopting a policy feature that already exists in another state.

. I therefore assume that reelection-seeking legislators do not adopt a policy feature that they believe will not be supported by the median voter regardless of whether that feature is novel and borrowed. For legislators, however, an important criterion in choosing between invention and borrowing relates to how credibly the legislators can sell the projected benefits of a given action and how the median voter can credibly perceive of (or visualize) the projected benefits of that action. The capability of legislators to credibly sell policy benefits combined with the capability of the median voter to perceive them is something I refer to as observability.109 While a less observable adoption has benefits that are harder to credibly sell to constituents and harder to credibly perceive for median voters, more observable adoptions have benefits that can more credibly be sold because they are more perceivable to median voters. Since borrowed policy features have an observable track record, they are on balance more discernible to the median voter than are untested policy features. This is the case since borrowing gives both legislators and the median voter a template—a state or states that already adopted the policy feature under consideration—that can be used to ascertain the possible benefits of adopting the already-tested policy feature (Volden 2006; Pacheco 2012). In contrast, no template exists for invention, suggesting that legislators will have greater difficulty credibly selling, and the median voter quickly perceiving, the possible benefits of inventing.

. In this book, I answer the above questions by delving into a meticulous study of state-level inventing and borrowing with respect to renewable portfolio standards, or RPS. An RPS generally specifies that electricity utility companies operating in a given state procure some of the electricity they provide to end-use consumers in that state from renewable sources. Appearing on the scene in the last quarter of the twentieth century and disseminating through a wide swath of states, RPSs have become the preeminent way in which the states have tried to spark renewable energy development (Rabe 2004, 2007; Carley and Miller 2012; Carley, Nicholson-Crotty, and Miller 2017; Parinandi 2020; Parinandi et al. 2020; and Stokes 2020). Unlike policies such as carbon taxes that have been shown to have limited appeal across the states, RPSs have been described as “climate policy that can actually win” and have been described by energy economist Michael Greenstone as “the biggest [carbon policy] we have in this country” (Meyer 2019). RPSs have also been a state-level rather than federal phenomenon, meaning that we can study and distinguish state inventing and borrowing in the absence of strong federal impulses that may distort when such inventing and borrowing occurs.10 Lastly, RPSs have been the site of tremendous state-level diversity in policy adoption, meaning that we actually observe instances of state invention and borrowing that can be dissected and analyzed.11 Simply put, focusing on the RPS area permits us to study state responses to federal inaction in Brandeis’s mold—that is, inventing by adopting novel policy or borrowing by adopting policy that has been tested in other states—in what is possibly one of the most important problems facing humanity in the twenty-first century.

. I test my theoretical expectations by evaluating whether government ideology and electoral vulnerability differentially influence the likelihood that legislatures invent or borrow when adopting a policy feature or a combination of policy features. In chapter 3, I walked through how invention and borrowing are identified in policy-making documents and also described how invention and borrowing datasets (regardless of the institutional actor doing the inventing or borrowing) could be created and then analyzed through a pooled event history procedure (Boushey 2016). This means that I can extract all instances where legislatures invent (or alternatively borrow) along with corresponding instances of non-adoption for invention (or borrowing) of a given policy feature or combination of policy features by a given state. I extract instances of non-adoption so that the true probabilities of a legislature inventing or borrowing are reflected in the structure of the datasets. Having an event history data structure that accurately captures the probability of an event’s occurrence—here, legislative adoption that can take two forms: invention or borrowing—is important since I adhere to convention within event history analysis and use a statistical technique called logistic regression to examine which independent variables make a given event more or less likely (Hosmer and Lemeshow 2000; Box-Steffensmeier and Jones 2004).110

. Finally, it is worth reemphasizing why I focus on legislative decision-making by itself rather than comparing legislative invention and borrowing to that of state public utilities commissions. First, all legislators are elected and subject to electoral constraints, meaning that zeroing in on legislatures allows us to evaluate how changes in the amount of electoral vulnerability (in a population where everyone is theoretically exposed to some amount of electoral vulnerability) might differentially influence invention and borrowing. Many state public utilities commissioners are appointed rather than elected, meaning that including these officeholders alongside legislatures in our analysis precludes us from figuring out how changes in electoral vulnerability among a population that faces similar institutional constraints affects invention and borrowing. Moreover, even if some public utilities commissioners are elected, the nature of their electoral constraints is arguably vastly different from that of legislators for two reasons: first, since the commissioners are selected on the basis of their stewardship of utility-related regulatory issues while legislators are selected on the basis of their stewardship of a “basket” of issues of which utility regulation is just one component (Besley and Coate 2003); and second, because elected commissioners not only create but also execute regulatory policy, meaning that they can electioneer in a way that is not open to legislators since legislators do not also possess executive power (Parinandi and Hitt 2018). Therefore, focusing on legislatures allows for the determination of how electoral vulnerability affects invention and borrowing in a Page 109 →situation where the population under investigation faces a similar kind of electoral constraint.111

. I now describe my empirical testing. I have two dependent variables, legislative invention and legislative borrowing, that map onto each of the respective legislative inventing and legislative borrowing datasets. I use logistic regression to determine which independent variables make the dependent variables more likely, and there are two independent variables in the analysis: a state’s current year government ideology, which is taken from Berry et al. (1998), and the median incumbent vote share of a legislatorPage 110 → in a state’s legislature in the most recent election. The median incumbent vote share variable captures electoral vulnerability based on the idea that a lower vote share implies greater anti-incumbent mood among the electorate and hence greater electoral vulnerability.112

. I also include a host of control variables and describe the most important ones here. Legislative professionalism captures the resource capacity of a state’s legislature in a given year and comes from Squire (2007); I include this variable to test whether legislative resource capacity differentially influences invention and borrowing. Other noteworthy controls include a state’s per capita income (measured here as the fraction of state income against the federal baseline where the baseline is set to 100); the price of energy in a state (measured in 2011 dollars per million British Thermal Units); a state’s percentage of energy that is produced from fossil fuel sources; a state’s citizen ideology (where the ideology of a state’s citizens is measured on a 0–100 scale by Berry et al. where 0 is perfectly conservative and 100 is perfectly liberal); and whether or not a state has unified Democratic government. As noted, I utilize separate logistic regressions for each dependent variable and cluster standard errors at the state level to reflect the idea that decisions within a state are correlated. In table 3, I display regression results for the variables described above.113

. Results from table 3 indicate how a given independent variable relates to legislative invention and borrowing. The critical values or thresholds (denoted by asterisks) in the table are used to indicate the certainty of estimated relationships, or the probability that the associations shown in Page 111 →the estimated relationships were determined through chance. A few points are worth mentioning. First, the positive association between government ideology and legislative invention lends support for the claim that greater governmental liberalism corresponds with an increased likelihood of invention in RPS policy-making.114 Second, the more certain positive association between government ideology and invention compared to the association between government ideology and borrowing gives support for the claim that governmental liberalism among lawmakers does a better job of explaining invention in RPS policy-making than it does borrowing. And, third, the finding of a statistically significant (or more certain) negative association between median incumbent legislator vote share and borrowing along with the lack of a statistically significant association between median incumbent legislator vote share and invention suggests that legislator electoral vulnerability matters more with respect to borrowing than it does with respect to invention.

. In figure 6, solid lines refer to the main predicted probabilities of legislative invention and borrowing as the level of government ideology varies.115 Dashed lines of the same color refer to corresponding upper and lower bound predicted probability estimates based on a 95% confidence interval. It is perhaps unsurprising that increased governmental liberalism should drive greater RPS-related invention and borrowing by legislatures since the policy domain of RPS is largely liberal in ideological character. However, predicted probabilities related to invention are more certain than those related to borrowing, as is reflected in the observation that the Page 113 →lower bound predicted probability curve with respect to invention does not cross zero while the analogous curve with respect to borrowing does. This gives us some assurance in saying that support exists for the claim that increased governmental liberalism does a better job of predicting legislative invention than it does of predicting legislative borrowing. Looking at the curves of invention and borrowing, it appears as if the slope of borrowing rises at a steeper clip than that of invention. This is partly an artifact of placing invention and borrowing on the same scale: since the invention dataset contains fewer instances of invention but more opportunities to invent than the borrowing dataset (which contains more instances of borrowing and fewer opportunities to borrow), invention appears to rise more gently. A useful way to compare slopes here would be to calculate the percentage rise in predicted probability values for invention and borrowing across the full range of the government ideology variable. Over this range, the predicted probability value for invention increases in size by almost 350% (roughly 343%) while the predicted probability value for borrowing increases in size by a little over 200% (roughly 204%).

. In figure 7, there is clear evidence that electoral vulnerability influences borrowing more than it does invention, and this evidence can be seen in comparing how the predicted probability of borrowing quickly decreases as median incumbent vote share increases while the predicted probability of inventing only gradually slopes downhill as median incumbent vote share increases.116

. I account for the issue described above in a supplemental analysis by considering all thresholds of the same type (e.g., required versus voluntary) that fall within the same decile or bandwidth to be classified as the Page 115 →same policy feature. This means that all instances of a state requiring that a utility generate anywhere between 20% and 29.9% of electricity sold to consumers from renewable sources are considered to be examples of adopting the same exact policy feature. This also means that all instances of a state requiring that a utility generate anywhere between 10% and 19.9% of electricity sold to consumers from renewable sources are considered to be examples of the same policy feature, and it means that all instances of a state giving a utility the voluntary option of generating anywhere between 10% and 19.9% of electricity sold to consumers from renewable sources are considered to be examples of the same policy feature. Using the decile or bandwidth definition of invention with respect to thresholds alleviates the marginal invention problem by assuming that features of the same type within the same bandwidth are functionally if not exactly similar. Regression estimates shown in table 4 show that the substantive results are unchanged.117

. Page 116 →I include several supplemental analyses for robustness that can be seen in the appendix of the book. First, I account for whether allowing for policy to be adopted via direct ballot initiative changes results; as seen in table A5, results are unchanged. In table A6, I use mean incumbent vote share rather than median incumbent vote share, and I find that the inclusion of this variable does not affect substantive results.118 In table A7, I include legislative term limits in the analysis and find results unchanged.

. Both Illinois and Indiana adopted renewable portfolio standard policy regimes, and both did it through legislative action. However, the content of each state’s regime has been quite different. Illinois has invented repeatedly, and in the span from 2007 to 2011, chose four novel policy Page 117 →features.119 Indiana, on the other hand, has been a consummate borrower and adopted a renewable portfolio standard regime in 2011 that almost consists entirely of borrowed policy features.120 The factors uncovered quantitatively can help articulate why the Illinois and Indiana legislatures embraced different levels of novelty in their RPS regimes. First, the Illinois legislature is quite liberal in ideological comportment and has had ideological scores ranging from 92.3 to 86.8 during the 2007–2011 interval (recall that 50 is the most ideologically neutral or moderate score in Berry et al.’s scale). The liberal nature of Illinois’s legislature presumably helped give its legislators greater motivation and willingness to seek out and adopt novel policy features in the state’s RPS program and made more ambitious policy-making conceivable. In contrast, Indiana’s legislature has been one of the most conservative in the nation and garnered a score of 3.8 (out of 100) in 2011. Generally composed of legislators who are presumably not predisposed to ambitious renewable energy, it is perhaps not surprising that the Indiana legislature largely shied away from inventing in crafting its RPS regime.

. Chapter 6. While the fifth chapter contains a theoretical and empirical exposition about legislative RPS invention, the sixth chapter does the same Page 13 →regarding public utilities commission–led RPS invention. Public utilities commissions are bound by mission statements upholding a central objective of fairness in regulatory behavior and conceivably attempt to maintain an appearance of being neutral arbiters in decision-making. This desire to conform to the appearance of being a neutral arbiter arguably influences when public utilities commissions choose to invent RPS policy. RPSs potentially raise costs for entrenched or long-standing electric utility companies by pushing these companies to change their electricity sourcing or procurement.12 Novel RPS policy arguably carries greater uncertainty for these firms than borrowing, where the experiences of companies in other states can inform firm observers. As surveys of firms have shown that they generally abhor uncertainty (Lagerberg 2015; Baker and Raskolnikov 2017), it is not a leap to surmise that entrenched electric utility companies dislike RPS invention. Public utilities commissions ostensibly know about the feelings of entrenched electric utility companies and strategically invent when these companies are weak to minimize chances of company-led pushback. This condition is more likely to obtain when states have undergone electricity sector deregulation, as entrenched electric utility companies have lost monopolistic or near-monopolistic power and are plausibly more likely to accommodate regulatory wishes to preserve the position (essentially holding on to control over distribution) that they still retain.

. Both Illinois and Indiana adopted renewable portfolio standard policy regimes, and both did it through legislative action. However, the content of each state’s regime has been quite different. Illinois has invented repeatedly, and in the span from 2007 to 2011, chose four novel policy Page 117 →features.119 Indiana, on the other hand, has been a consummate borrower and adopted a renewable portfolio standard regime in 2011 that almost consists entirely of borrowed policy features.120 The factors uncovered quantitatively can help articulate why the Illinois and Indiana legislatures embraced different levels of novelty in their RPS regimes. First, the Illinois legislature is quite liberal in ideological comportment and has had ideological scores ranging from 92.3 to 86.8 during the 2007–2011 interval (recall that 50 is the most ideologically neutral or moderate score in Berry et al.’s scale). The liberal nature of Illinois’s legislature presumably helped give its legislators greater motivation and willingness to seek out and adopt novel policy features in the state’s RPS program and made more ambitious policy-making conceivable. In contrast, Indiana’s legislature has been one of the most conservative in the nation and garnered a score of 3.8 (out of 100) in 2011. Generally composed of legislators who are presumably not predisposed to ambitious renewable energy, it is perhaps not surprising that the Indiana legislature largely shied away from inventing in crafting its RPS regime.

. In terms of electoral vulnerability, Indiana legislators were definitely more vulnerable than their Illinois counterparts in the sense that the median incumbent vote share in the most recent election for the Indiana legislature was 64.72% while it was in the range of 100 to 73.73% for the Illinois legislature in the 2007–11 period.121 Although Indiana’s median incumbent vote share may seem high, it is lower than the state’s own median value for this variable (67.46) and lower than the overall 50 state median value for this variable (71.65). Given legislators’ overwhelming focus on working toward reelection and their adeptness at securing it (Mayhew 1974), it is possible that the Indiana legislators may have interpreted recent incumbent electoral margins (e.g., being more vulnerable than what is typically the norm in their own state or across states) as a sign of weakness and consequently sought out greater observability in Page 118 →their policy adoption choices, leading to an RPS regime consisting almost chiefly of borrowed policy features. In sum, Indiana’s dearth of invention is due in part to its legislative ideology—and particularly its conservatism, given that clean energy is within the liberal ideological worldview—and its borrowing is due in part to having a level of electoral vulnerability that is greater than typical within-state or across-state levels.122

. In terms of electoral vulnerability, Indiana legislators were definitely more vulnerable than their Illinois counterparts in the sense that the median incumbent vote share in the most recent election for the Indiana legislature was 64.72% while it was in the range of 100 to 73.73% for the Illinois legislature in the 2007–11 period.121 Although Indiana’s median incumbent vote share may seem high, it is lower than the state’s own median value for this variable (67.46) and lower than the overall 50 state median value for this variable (71.65). Given legislators’ overwhelming focus on working toward reelection and their adeptness at securing it (Mayhew 1974), it is possible that the Indiana legislators may have interpreted recent incumbent electoral margins (e.g., being more vulnerable than what is typically the norm in their own state or across states) as a sign of weakness and consequently sought out greater observability in Page 118 →their policy adoption choices, leading to an RPS regime consisting almost chiefly of borrowed policy features. In sum, Indiana’s dearth of invention is due in part to its legislative ideology—and particularly its conservatism, given that clean energy is within the liberal ideological worldview—and its borrowing is due in part to having a level of electoral vulnerability that is greater than typical within-state or across-state levels.122

. Ideology’s role in fostering legislative invention is also noteworthy because other factors that at first glance would appear to influence legislative invention do not. Far from causing legislators to seek out novel policy solutions, electoral vulnerability in the RPS area has no noticeable connection to legislative invention and actually encourages borrowing. And while it is in certain respects desirable that electoral vulnerability does not influence invention, the lack of a link between electoral vulnerability and invention is potentially problematic for public policy if the scale of a policy challenge—such as mitigating the consequences of climate change or devising a more effective way to regulate the burgeoning market for recreational drugs—is so large that it requires novel policy-making.123

. Another limit concerns the possibility that results may be different in a policy domain that is considered to be a more classical exemplar of “moral” policy-making than is RPS (Mooney and Lee 1995, 1999). While climate change (along with the policies devised to deal with it) is increasingly being viewed as a moral issue with overtones of right and wrong (Adger et al. 2017), it takes a greater leap of faith to say that RPS policies, perhaps especially given their complexity and the perceived indirect nature between RPS policy-making and carbon production, inspire feelings of morality among the public.124 Indeed, one recent study indicates that the general public’s concerns about RPS policies are primarily financial rather than moral (Stokes and Warshaw 2017). I do not expect the finding linking government ideology and legislative invention to differ substantially Page 121 →if more classically moral policy-making (such as a policy domain dealing with environmental justice and the eradication of health disparities among individuals owing to exposure to toxins) were analyzed, since legislatures that are more ideologically extreme should display a greater willingness than moderate peers to seek out and adopt novel policy features from policy domains that conform to that ideological worldview.

. While it seems reasonable and defensible to claim that state regulatory agencies are key actors in the invention of policy, the “policy innovation” literature in political science is primarily legislature-centric (e.g., Kousser 2005; Shipan and Volden 2006; Boushey 2010), and appears to have sidestepped the investigation of how regulatory agencies invent novel policy. Page 124 →The legislative focus not only encompasses seminal work on diffusion (Berry and Berry 1990) but also includes earlier work on policy reinvention (Glick and Hays 1991; Hays 1996) and recent work on policy reinvention in the renewable energy policy space (Carley, Nicholson-Crotty, and Miller 2017).125 It is easy to understand why scholars would give more attention to legislatures. First, legislative action arguably represents where the high drama of policy-making occurs, suggesting that accounts of legislative activity may be more compelling to broad audiences. Second, legislatures are the main institutional actors tasked with adopting law, which could impart a bias toward studying legislative activity. And third, state legislative acts—which are typically published as session laws—may be systematically easier to locate compared to state regulatory decisions.

. Whatever the reason for the lack of focus on regulatory agencies, ignoring an investigation of the conditions under which regulatory agencies invent policy would produce a glaring gap in our knowledge of how the creation of novel policy occurs across the states. This is not only because regulatory agencies adopt so many policies (Boushey and McGrath 2017) but also because they are in a position to adopt so many policies: regulatory agencies serve as mediators between the entities that they regulate and the public and therefore have considerable power to craft policy pertaining to how entities interact with the public.126 Given that regulatory bodies play such a prominent role in organizing modern life that modern government itself has been referred to as “the regulatory state,” understanding when regulatory agencies invent policy sheds light on a crucial pathway (regulatory agency decision-making) through which new policies can be introduced to the system of policies generated as a result of experimentation across the states (Glaeser and Shleifer 2003).

. I argue that this is the case because deregulation disrupts the ability of entrenched electric utility companies to challenge attempts by public utilities commissions to adopt novel regulation.127 Inventing represents unprecedented adoption—not just in Walker’s (1969) sense of being new to the state adopting it but also being new across the system of all states—and entrenched electric utility companies might find unprecedented regulation to be particularly nettlesome because they are uncertain about how such regulation will affect their finances and cannot use firm experiences in other states to make inferences about financial impacts (as could conceivably be done in borrowing). Therefore, public utilities commissions operating under sector deregulation (where entrenched electric utility companies are less well positioned to challenge policy development by the commissions) may be more likely to invent novel regulation compared to commissions operating under sector regulation (where entrenched electric utility companies are in a stronger position). This finding comports with Page 126 →the view dating back to the Weberian ideal (Constans 1958; Walton 2005; Gualmini 2008) and reiterated in literature on regulation (West 2005; Dal Bo 2006; Braithwaite 2008; Kettl 2008; Carrigan and Coglianese 2011) that regulatory agencies act as self-preservationists. Part of acting out of self-preservation also entails that regulatory agencies are strategic about when they seek to enact novel regulation that could stoke opposition from entrenched electric utility companies: regulatory agencies adopt such regulation when they think that entrenched electric utility companies are less likely to challenge them, and this condition is more likely to obtain under sector deregulation as opposed to regulation.128

. I argue that this is the case because deregulation disrupts the ability of entrenched electric utility companies to challenge attempts by public utilities commissions to adopt novel regulation.127 Inventing represents unprecedented adoption—not just in Walker’s (1969) sense of being new to the state adopting it but also being new across the system of all states—and entrenched electric utility companies might find unprecedented regulation to be particularly nettlesome because they are uncertain about how such regulation will affect their finances and cannot use firm experiences in other states to make inferences about financial impacts (as could conceivably be done in borrowing). Therefore, public utilities commissions operating under sector deregulation (where entrenched electric utility companies are less well positioned to challenge policy development by the commissions) may be more likely to invent novel regulation compared to commissions operating under sector regulation (where entrenched electric utility companies are in a stronger position). This finding comports with Page 126 →the view dating back to the Weberian ideal (Constans 1958; Walton 2005; Gualmini 2008) and reiterated in literature on regulation (West 2005; Dal Bo 2006; Braithwaite 2008; Kettl 2008; Carrigan and Coglianese 2011) that regulatory agencies act as self-preservationists. Part of acting out of self-preservation also entails that regulatory agencies are strategic about when they seek to enact novel regulation that could stoke opposition from entrenched electric utility companies: regulatory agencies adopt such regulation when they think that entrenched electric utility companies are less likely to challenge them, and this condition is more likely to obtain under sector deregulation as opposed to regulation.128

. State public utilities commissions emerged in the early twentieth century as an institutional answer to the issue of ensuring that electric utility companies retain financial viability without gouging consumers (Troesken 2006).129 When electric utility companies first emerged and constructed electricity generation and distribution infrastructure to supply electricity to consumers, oversight of the companies was practically nonexistent; municipal governments had entered into “franchise contracts” with electric utility companies but lacked authority to compel electric utility companies to treat consumers fairly, owing to the belief among state governments that electricity service regulation was not an “essential” function of municipal government (Gormley 1983; Troesken 2006). A consequence of this lack of oversight was that electric utility companies faced few constraints on the prices they could charge consumers and charged consumers exorbitantly high prices for electricity service. A solution to this problem was to allow municipal governments to regulate the activity of electric utility companies, and several states—including California, Florida, and Ohio—enacted laws permitting municipal governments to directly regulate the activity of electric utility companies operating within municipal boundaries.

. RPSs, in contrast, are levied on electric utility companies. In turn, electric utility companies have a largely adversarial relationship with the public, which chiefly interacts with these companies while paying bills or dealing with electricity outages, and which has low levels of trust concerning how fairly it is treated by these companies (Consumer Reports 2018).13 The fact that RPSs are imposed on parties—electric utility companies—withPage 20 → which the public does not have especially cordial relations, combined with the fact that costs borne by the public due to RPSs are largely indirect (meaning that they are folded into electricity bills instead of being standalone items), might account for why RPSs have taken hold across the bulk of the United States while policies like carbon taxes have not.

. The creation of state public utilities commissions was a compromise Page 128 →solution aimed at balancing the demands of electric utility companies and consumers (Troesken 2006; Parinandi and Hitt 2018). Indeed, in some states (such as Illinois, as relayed in Troesken 2006), electric utility companies themselves clamored for state-level regulation, hoping that placing regulatory authority in the hands of multimember state-level commissions would reduce opportunities for pandering compared to when such regulatory authority is wholly in the hands of municipal governments.130 While electric utility companies recognized that they would face some amount of regulation from state-level commissions, the companies believed that they would receive fairer treatment at the hands of state commissions rather than municipal governments and the regulation of electric utility companies by state-level public utilities commission has persisted into the twenty-first century (Knittel 2006; Neufeld 2008).

. Today, state public utilities commissions possess regulatory authority over electric utility companies operating in their states, retain considerable power over the prices that are charged by electric utility companies to consumers—by approving prices directly in regulated states, by setting price caps (which were in place for the bulk of the later part of the time period of this study) and still approving distribution prices in deregulated states, and by controlling access to electricity retail markets in their states.131 State public utilities commissions typically set policy with the goals of balancing electric utility company and consumer price demands while also ensuring access to “safe and reliable electricity” as well as preparing for “the future of the electricity system” in terms of readying electricity infrastructure to meet anticipated changes in demand and the energy mix used to derive Page 129 →supply (Nanavati and Gundlach 2016, 6–7). State public utilities commissions are also considered widely to be experts in the area of electricity regulation and enjoy latitude to set their own policies regarding the regulation, maintenance, and upgrade of electricity service provision (Byrnett and Shea 2019).

. Even though state public utilities commissions have latitude to set policy regarding electricity service provision, it is important to note that commissions are expected to chiefly serve as neutral arbiters of electric utility company and consumer interests and are further expected not to overly disadvantage companies or consumers in commission decision-making (White 2018). This expectation of neutrality—which is important so that public utilities commissions have legitimacy with both the companies they regulate and consumers at large—arguably may account for the rareness of regulatory policy feature adoption concerning RPS (since adopting an RPS policy feature opens up commissions to charges of being biased against electric utility companies). It may also account for why public utilities commissions are more likely to invent novel RPS policy when companies are in a less fortuitous position to challenge them.132 Many state public utilities commissions have utilized their regulatory authority and adopted RPS policy features—inventing several of these by making their states the first to adopt them—on the pretext of providing efficient access to electricity and planning for possible disruptions to the supply of energy.133 In table 5 below, I provide a list of states that adopted RPS policy features via public utilities commission rulemaking along with the percentage of a state public utilities commission’s adoptions that are examples of inventing.

. Even though state public utilities commissions have latitude to set policy regarding electricity service provision, it is important to note that commissions are expected to chiefly serve as neutral arbiters of electric utility company and consumer interests and are further expected not to overly disadvantage companies or consumers in commission decision-making (White 2018). This expectation of neutrality—which is important so that public utilities commissions have legitimacy with both the companies they regulate and consumers at large—arguably may account for the rareness of regulatory policy feature adoption concerning RPS (since adopting an RPS policy feature opens up commissions to charges of being biased against electric utility companies). It may also account for why public utilities commissions are more likely to invent novel RPS policy when companies are in a less fortuitous position to challenge them.132 Many state public utilities commissions have utilized their regulatory authority and adopted RPS policy features—inventing several of these by making their states the first to adopt them—on the pretext of providing efficient access to electricity and planning for possible disruptions to the supply of energy.133 In table 5 below, I provide a list of states that adopted RPS policy features via public utilities commission rulemaking along with the percentage of a state public utilities commission’s adoptions that are examples of inventing.

. The table displays significant breadth across the states in terms of regulatory adoption, with at least one state in each region of the United States adopting a RPS policy feature through public utilities commission decision. The table also shows that most of these states’ public utilities commissions invented and borrowed when adopting their respective RPS policy Page 130 →features. Similar to the strategy used with respect to legislatures in the last chapter, here we can compare how variables relate to regulatory inventing and borrowing and ultimately uncover factors that have unique resonance in explaining when regulatory agencies invent.134 Before doing this and specifically before articulating why sector deregulation corresponds uniquely with regulatory inventing, I review what sector deregulation is and what it has entailed for state public utilities commissions.

. Under deregulation, the electric utility companies that previously had a vertically integrated monopoly over the electricity supply chain lost this control. Specifically, entrenched electric utility companies—the companies that previously exercised monopolistic control under a regulated sector—lost control over much of their generating infrastructure (sometimes through forced divestment, as is discussed in Kwoka et al. 2010) and had to compete with new entrants that could sell electricity directly to final users (Borenstein and Bushnell 2015). While the entrenched electric utility companies retain full control over distribution infrastructure (distribution refers to delivering electricity to the final user), they have to allow other firms to have access to that distribution infrastructure. This combination of losing control of much of their generation capacity along with needing to accommodate other producers in their distribution networks is what exposed entrenched electric utility companies to retail competition over electricity but not the distribution of it (Borenstein and Bushnell 2015). In fact, under deregulation, a final user might pay two monthly bills for their electricity usage: one to a new entrant firm for the electricity itself, and one to an entrenched electric utility company for distributing the electricity to the final user.135 In the event where an entrenched electric utility company does not have enough generating capacity on hand to meet the demands of consumers—which has occurred commonly in several deregulated states in cases where entrenched electric utility companies lost much of their generating capacity through divestment but only lost small numbers of consumers to other firms—the entrenched electric utility company will Page 132 →need to procure or purchase the demanded electricity from other firms and distribute it over its own network to its consumers.

. In figure 8, I show a map of the continental United States in which states are grouped according to whether or not they have deregulated their electricity sectors by the end of this study.136 While deregulation seemed like the wave of the future in the 1990s, it has largely stopped diffusing across the states and is concentrated in the Northeast, the West Coast, Texas, and the Great Lakes region (Delmas et al. 2007; Borenstein and Bushnell 2015). California’s experience with severe electricity blackouts in the early 2000s concerned officials in some other states about the reliability of electricity service under deregulation (Borenstein and Bushnell 2015). Nonetheless, a sizeable number of states have pursued and retained deregulated electricity sectors as of the end point of this study. Several of these states (e.g., Texas, New York, Illinois, Pennsylvania, and Ohio) are among the most populous in the United States, suggesting that many Americans receive electricity from deregulated sectors. Moreover, deregulation is Page 133 →probably here to stay given that many states are nearing their third decades under deregulation. While the jury is out in terms of how deregulation has impacted electricity prices (Borenstein and Bushnell 2015), it has arguably influenced how public utilities commissions approach inventing. Understanding how regulatory RPS invention activity differs across states with deregulated versus regulated sectors is consequently worth exploring since it sheds light on how the regulatory environment faced by public utilities commissions can influence policy adoption behavior.

. Entrenched electric utility companies have generally taken a dim view of RPS programs (Stokes 2020). A key reason why is that these companies incur financial costs in complying with RPS programs. This simply means that entrenched electric utility companies will incur some financial cost to switch from fossil-fuel-based sources to renewable sources to meet an RPS mandate. Scholars (Kim et al. 2016; Greenstone and Nath 2019) have shown that renewable energy is more expensive to produce than fossil-fuel-based energy.137 Given that the price of electricity is an important aspect of energy policy to consumers (Besley and Coate 2003), and given that a major element of the popularity of RPS programs among consumers is that the programs are imposed on electric utility companies rather than directly on consumers (Äklin and Urpelainen 2018), such as an individual Page 134 →carbon tax would be, entrenched electric utility companies will likely face some financial impact in complying with RPS obligations.138 Moreover, entrenched electric utility companies in both regulated and deregulated settings face the financial impacts (Kim et al. 2016): while entrenched electric utility companies in regulated settings may have to upgrade their own infrastructure to comply with RPS obligations, entrenched electric utility companies in deregulated settings (to the degree that they no longer own generation facilities) may have to negotiate costly procurement deals with new suppliers to obtain the renewable energy needed to meet RPS obligations.

. Entrenched electric utility companies have generally taken a dim view of RPS programs (Stokes 2020). A key reason why is that these companies incur financial costs in complying with RPS programs. This simply means that entrenched electric utility companies will incur some financial cost to switch from fossil-fuel-based sources to renewable sources to meet an RPS mandate. Scholars (Kim et al. 2016; Greenstone and Nath 2019) have shown that renewable energy is more expensive to produce than fossil-fuel-based energy.137 Given that the price of electricity is an important aspect of energy policy to consumers (Besley and Coate 2003), and given that a major element of the popularity of RPS programs among consumers is that the programs are imposed on electric utility companies rather than directly on consumers (Äklin and Urpelainen 2018), such as an individual Page 134 →carbon tax would be, entrenched electric utility companies will likely face some financial impact in complying with RPS obligations.138 Moreover, entrenched electric utility companies in both regulated and deregulated settings face the financial impacts (Kim et al. 2016): while entrenched electric utility companies in regulated settings may have to upgrade their own infrastructure to comply with RPS obligations, entrenched electric utility companies in deregulated settings (to the degree that they no longer own generation facilities) may have to negotiate costly procurement deals with new suppliers to obtain the renewable energy needed to meet RPS obligations.

. In contrast, with novel policy, the entrenched electric utility company has no other firm-specific reference points it can use to guide its planning and is arguably largely on its own in terms of anticipating how the policy could affect it. This impacts the company by raising uncertainty about whether its chosen strategy could backfire upon adoption of the policy. The company, for example, could advocate for a certain social benefits Page 136 →level but would be doing this without being able to benchmark its strategy against the experiences of firms in other states and might fear that it could incur unanticipated losses as a result. Similarly, the company could employ cost-cutting measures but would be doing this without being able to benchmark its strategy against the experiences of firms in other states and also might fear that it could incur unanticipated losses as a result.139 Ultimately, not being able to benchmark plans against the experiences of peers might increase feelings of uncertainty among company executives and lead to a situation where executives believe that they are unable to effectively manage or adapt to the policy under consideration for adoption. This feeling of a loss of control can generate opposition among company executives toward invention, as executives believe that their companies are being made the objects of regulatory experimentation. Company executives’ dislike of uncertainty (and by extension, of novel regulatory attempts that have not even been tried elsewhere) is not conjecture. A 2015 survey of global business executives by the firm Grant Thornton revealed that executives prefer higher taxes to greater uncertainty in regulation (Lagerberg 2015); this finding is corroborated by academic research (Baker and Raskolnikov 2017) showing that companies generally oppose uncertainty.

. Even though RPSs have found acceptance in much of the United States, it is worth noting that RPS policy adoption has been a state rather than federal phenomenon. To be sure, there has been a federal-level attempt to enact a policy similar to an RPS: the Obama administration’s Clean Power Plan functioned like an RPS in that states were expected to achieve clean energy targets through regulating electricity generation and procurement.14 However, RPS-like policies have failed to gather steam at the federal level, and the states have taken the leading role in advancing these policies (Parinandi 2020). Federal inaction has occurred not only because the U.S. Senate amplifies the interests of rural states with fossil-fuel-related industries (e.g., Alabama or Alaska), which may generally be skeptical of environmental regulation at the expense of urban states with less dependence on fossil fuel-related production (e.g., California or Rhode Island), which may be more receptive to environmental regulation (Warf 2008). It has also occurred because the individual states (and many of the members of Congress who represent these states) arguably desired flexibility to craft their energy and environmental policies in ways that they themselves saw fit (Peterson 1995; and Karch 2007).

. There is a second reason why entrenched electric utility companies in deregulated states may be less willing to push back against novel regulatory invention, and it boils down to the idea that entrenched electric utility companies in deregulated states have already lost monopoly power and adopt a conciliatory approach vis-à-vis regulators to preserve the advantages that they still retain. Losing much of its generation assets, for example, may make a formerly monopolistic electric utility company engage in loss prevention to reduce other threats to its dominance. For instance, a formerlyPage 138 → monopolistic electric utility company may still want to be declared the default electricity provider within its service area.140 Furthermore, the same entrenched electric utility company may want to preserve its control over the distribution of electricity. Given its desire to forestall future losses to its position, the entrenched electric utility company might act in a conciliatory manner toward regulators to build a positive reputation among those regulators. One manifestation of this conciliatory approach could be reduced opposition toward novel regulatory policy-making.

. I include the price of energy (measured in 2011 dollars per million British Thermal Units) to account for the possibility that energy pricing trends could spur regulatory policy adoption. I also include a state’s average annual daily solar radiation level (measured by the National Renewable Energy Laboratory and in units of kilowatt-hours per square meter per day); I include this variable since some of the states (e.g., Arizona and New Mexico) that have been prolific regulatory adopters of RPS policy features have high solar energy potential, which could have spurred policy feature adoption.141 As in the previous chapter, other notable controls include a state’s per capita income; a state’s percentage of energy that is produced from fossil fuel sources; and a state’s citizen ideology (where, like before, the ideology of a state’s citizens is measured on a 0–100 scale where 0 is perfectly conservative and 100 is perfectly liberal).142

. I include the price of energy (measured in 2011 dollars per million British Thermal Units) to account for the possibility that energy pricing trends could spur regulatory policy adoption. I also include a state’s average annual daily solar radiation level (measured by the National Renewable Energy Laboratory and in units of kilowatt-hours per square meter per day); I include this variable since some of the states (e.g., Arizona and New Mexico) that have been prolific regulatory adopters of RPS policy features have high solar energy potential, which could have spurred policy feature adoption.141 As in the previous chapter, other notable controls include a state’s per capita income; a state’s percentage of energy that is produced from fossil fuel sources; and a state’s citizen ideology (where, like before, the ideology of a state’s citizens is measured on a 0–100 scale where 0 is perfectly conservative and 100 is perfectly liberal).142

. Page 144 →I also include four variables for both the regulatory invention and regulatory borrowing models that capture how conditions across the other institutions of a state’s government might influence regulatory decision-making. Unified Democratic government is a binary variable that could influence when regulators feel more comfortable adopting RPS policy features. Government ideology comes from Berry et al. (1998) and captures whether regulators are operating in conservative versus liberal arena (where conservatism to liberalism is measured on a 100-point scale, where 0 denotes pure conservatism while 100 denotes pure liberalism).143 Legislative term limits is a binary variable receiving a value of 1 if a state imposes term limits on its legislators (it is possible that term limits reduce legislative capacity and thereby make regulatory policy adoption more likely).144 Finally, I include legislative median incumbent vote share from the most recent election (analogous to what was included in the previous chapter) to capture the possibility that regulators could be influenced by the anti-incumbent mood displayed toward legislators.145 As in the previous chapter, I utilize separate logistic regressions for each dependent variable and cluster standard errors at the state level to reflect the idea that decisions within a state are correlated.146

. Page 144 →I also include four variables for both the regulatory invention and regulatory borrowing models that capture how conditions across the other institutions of a state’s government might influence regulatory decision-making. Unified Democratic government is a binary variable that could influence when regulators feel more comfortable adopting RPS policy features. Government ideology comes from Berry et al. (1998) and captures whether regulators are operating in conservative versus liberal arena (where conservatism to liberalism is measured on a 100-point scale, where 0 denotes pure conservatism while 100 denotes pure liberalism).143 Legislative term limits is a binary variable receiving a value of 1 if a state imposes term limits on its legislators (it is possible that term limits reduce legislative capacity and thereby make regulatory policy adoption more likely).144 Finally, I include legislative median incumbent vote share from the most recent election (analogous to what was included in the previous chapter) to capture the possibility that regulators could be influenced by the anti-incumbent mood displayed toward legislators.145 As in the previous chapter, I utilize separate logistic regressions for each dependent variable and cluster standard errors at the state level to reflect the idea that decisions within a state are correlated.146

. Page 144 →I also include four variables for both the regulatory invention and regulatory borrowing models that capture how conditions across the other institutions of a state’s government might influence regulatory decision-making. Unified Democratic government is a binary variable that could influence when regulators feel more comfortable adopting RPS policy features. Government ideology comes from Berry et al. (1998) and captures whether regulators are operating in conservative versus liberal arena (where conservatism to liberalism is measured on a 100-point scale, where 0 denotes pure conservatism while 100 denotes pure liberalism).143 Legislative term limits is a binary variable receiving a value of 1 if a state imposes term limits on its legislators (it is possible that term limits reduce legislative capacity and thereby make regulatory policy adoption more likely).144 Finally, I include legislative median incumbent vote share from the most recent election (analogous to what was included in the previous chapter) to capture the possibility that regulators could be influenced by the anti-incumbent mood displayed toward legislators.145 As in the previous chapter, I utilize separate logistic regressions for each dependent variable and cluster standard errors at the state level to reflect the idea that decisions within a state are correlated.146

. Page 144 →I also include four variables for both the regulatory invention and regulatory borrowing models that capture how conditions across the other institutions of a state’s government might influence regulatory decision-making. Unified Democratic government is a binary variable that could influence when regulators feel more comfortable adopting RPS policy features. Government ideology comes from Berry et al. (1998) and captures whether regulators are operating in conservative versus liberal arena (where conservatism to liberalism is measured on a 100-point scale, where 0 denotes pure conservatism while 100 denotes pure liberalism).143 Legislative term limits is a binary variable receiving a value of 1 if a state imposes term limits on its legislators (it is possible that term limits reduce legislative capacity and thereby make regulatory policy adoption more likely).144 Finally, I include legislative median incumbent vote share from the most recent election (analogous to what was included in the previous chapter) to capture the possibility that regulators could be influenced by the anti-incumbent mood displayed toward legislators.145 As in the previous chapter, I utilize separate logistic regressions for each dependent variable and cluster standard errors at the state level to reflect the idea that decisions within a state are correlated.146

. Moving to some of the controls from table 6, results regarding the Page 146 →elected variable are worth discussing. Recall that the elected variable captures whether commissioners are selected by voters or through gubernatorial appointment subject to legislative confirmation and oversight. The positive association between election and regulatory invention is fascinating since it suggests that being free from gubernatorial and legislative oversight gives commissioners the ability to take policy-making proposals directly to voters, a path that is arguably closed to appointees, who potentially may be castigated by governors and legislative leaders sympathetic to electric utility company interests. At the same time, the even larger association between the elected variable and regulatory borrowing suggests that, like their legislative counterparts, elected public utilities commissioners may also emphasize the observable aspects of policy insofar as they can credibly tell voters that they are trying to replicate successes seen elsewhere (Kogan, Lavertu, and Peskowitz 2016). Another implication that can be drawn out from the elected variable is that appointees are less likely to either invent or borrow. A takeaway from this is that we should perhaps not expect invention to come disproportionately from appointees.147

. Page 147 →Another result worth discussing concerns the government ideology variable. Recall that this variable captures the broader governmental ideological environment in which regulators operate, as the variable is derived from legislator behavior. The negative and statistically significant relationship between government ideology and regulatory invention warrants explanation, and here, I believe this finding is an artifact of sizeable regulatory invention activity that took place in the state of Arizona. During the 1983–2011 time span of this study and based on mean government ideology scores, Arizona has been one of the most conservative states in the United States (with a mean government ideology score of 27.89, where 0 denotes pure conservatism while 100 denotes pure liberalism).148 Despite the state’s conservative profile, Arizona’s public utilities commission—officially called the Arizona Corporation Commission—invented with respect to adopting several RPS policy features and did so out of an attempt to capitalize on the state’s ample solar energy potential.149 This is confirmed by the Arizona Corporation Commission’s decision in 1996 to adopt an RPS program dedicated wholly to developing the state’s solar electricity industry (Arizona Corporation Commission 1996).150

. Page 147 →Another result worth discussing concerns the government ideology variable. Recall that this variable captures the broader governmental ideological environment in which regulators operate, as the variable is derived from legislator behavior. The negative and statistically significant relationship between government ideology and regulatory invention warrants explanation, and here, I believe this finding is an artifact of sizeable regulatory invention activity that took place in the state of Arizona. During the 1983–2011 time span of this study and based on mean government ideology scores, Arizona has been one of the most conservative states in the United States (with a mean government ideology score of 27.89, where 0 denotes pure conservatism while 100 denotes pure liberalism).148 Despite the state’s conservative profile, Arizona’s public utilities commission—officially called the Arizona Corporation Commission—invented with respect to adopting several RPS policy features and did so out of an attempt to capitalize on the state’s ample solar energy potential.149 This is confirmed by the Arizona Corporation Commission’s decision in 1996 to adopt an RPS program dedicated wholly to developing the state’s solar electricity industry (Arizona Corporation Commission 1996).150

. The state RPS adoption story is one that comports with other areas of policy experimentation (e.g., lottery adoption as in Berry and Berry 1990, or antismoking adoption as in Shipan and Volden 2008) where a few jurisdictions serve as initial adopters while a greater number of jurisdictions are later adopters who aid in the diffusion of the policy (Rogers 1962; Walker 1969; Gray 1973). What this study offers beyond those accounts is both a systematic distinction between invention and borrowing at the subpolicy or policy feature level that recognizes the fact that later adopters could still invent at the subpolicy level (thereby permitting us to observe invention and borrowing regardless of when a state adopted an RPS policy writ large) and the existence of institutional variation vis-à-vis legislatures and public utilities commissions that lets us study the factors influencing invention and borrowing in each of these two institutional actors.15

. Page 147 →Another result worth discussing concerns the government ideology variable. Recall that this variable captures the broader governmental ideological environment in which regulators operate, as the variable is derived from legislator behavior. The negative and statistically significant relationship between government ideology and regulatory invention warrants explanation, and here, I believe this finding is an artifact of sizeable regulatory invention activity that took place in the state of Arizona. During the 1983–2011 time span of this study and based on mean government ideology scores, Arizona has been one of the most conservative states in the United States (with a mean government ideology score of 27.89, where 0 denotes pure conservatism while 100 denotes pure liberalism).148 Despite the state’s conservative profile, Arizona’s public utilities commission—officially called the Arizona Corporation Commission—invented with respect to adopting several RPS policy features and did so out of an attempt to capitalize on the state’s ample solar energy potential.149 This is confirmed by the Arizona Corporation Commission’s decision in 1996 to adopt an RPS program dedicated wholly to developing the state’s solar electricity industry (Arizona Corporation Commission 1996).150

. Dropping observations corresponding to Arizona and re-estimating the analysis removes the association between government ideology and regulatory invention, providing a measure of support for the idea that the finding linking conservative government ideology to regulatory invention was driven partially by invention on the part of Arizona.151 The possibilityPage 148 → that the finding linking government ideology to regulatory invention is uninformative—not only because government ideology is constructed from legislator behavior and does not map neatly onto regulatory behavior, but also because the result could be driven by regulators in a conservative state (Arizona) inventing to take advantage of solar energy potential—raises the issue of how much we should even expect the ideology of public utilities commissioners (to the extent that this can actually be measured) to play an influential role in commissioner decision-making. The nature of the work that public utilities commissions perform—managing the relationship between electric utility companies and rate-paying consumers to make sure that the interests of both sides are represented fairly in decision-making—militates against ideology playing an overt role in public utilities commission decision-making. The Arizona Corporation Commission, for example, publicly states that a central goal in its oversight of utilities operations involves trying “to balance the consumers’ interest in affordable and reliable utility service with the utility’s interest in earning a fair profit” (Arizona Corporation Commission 2020).

. Similar to the previous chapter, I also show results where similar rates and thresholds have been combined; these results, shown in table 7, substantively mirror the results in table 6.152 I include other robustness checks Page 151 →in the analysis. In table A11 of the appendix, I include a variable accounting for whether a state allows for policy to be adopted via direct ballot initiative and find that substantive results do not change. Similarly, in table A12 of the appendix, I substitute mean incumbent vote share for median incumbent vote share and find that substantive results do not change.153 In tables A13 and A14, I respectively drop combinations and drop rates or thresholds from my analysis. In each of these analyses, the deregulation variable is no longer statistically significant, although it retains a larger influence with respect to invention compared to borrowing when rates Page 152 →or thresholds are dropped from the analysis. One thing to bear in mind is that dropping combinations results in an extremely low number of cases of regulatory invention (16), representing a more than 50% reduction in the number of cases of regulatory invention (36) appearing in the main regulatory invention dataset.

. Similar to the previous chapter, I also show results where similar rates and thresholds have been combined; these results, shown in table 7, substantively mirror the results in table 6.152 I include other robustness checks Page 151 →in the analysis. In table A11 of the appendix, I include a variable accounting for whether a state allows for policy to be adopted via direct ballot initiative and find that substantive results do not change. Similarly, in table A12 of the appendix, I substitute mean incumbent vote share for median incumbent vote share and find that substantive results do not change.153 In tables A13 and A14, I respectively drop combinations and drop rates or thresholds from my analysis. In each of these analyses, the deregulation variable is no longer statistically significant, although it retains a larger influence with respect to invention compared to borrowing when rates Page 152 →or thresholds are dropped from the analysis. One thing to bear in mind is that dropping combinations results in an extremely low number of cases of regulatory invention (16), representing a more than 50% reduction in the number of cases of regulatory invention (36) appearing in the main regulatory invention dataset.

. On the other hand, if a different observer believes that regulatory agencies have an obligation to use their managerial “on-the-ground” position to craft myriad novel solutions to a whole host of pressing social challenges, that different observer might find disappointment, not only in the scarcity of regulatory invention but also in the revelation (based on the sector regulation versus deregulation finding here) that invention may not take place in large areas that could benefit normatively from it.154 The behavior of state public utilities commissioners regarding RPS policy feature adoption suggests that those desiring greater invention perhaps should not look to regulatory rulemaking to serve as a substitute for legislative decision-making. In the next chapter, I move on to case studies that further explore the dynamics discussed in this and the previous chapter.

. Illinois and Indiana are not only neighboring states located in the Midwest but share a host of similar attributes. Both are states whose economies were heavily dependent on manufacturing (Matejka 1999; Turner 2016); both states were also impacted by deindustrialization in the late twentieth century (Matejka 1999; Turner 2016), arguably creating an opportunity to invest in technologies that could create employment prospects (renewablePage 155 → energy technology has often been billed in this manner). Both states are also appropriately described as agricultural titans: over the twenty-year span between 2000 and 2020, for example, Illinois and Indiana both ranked among the top five corn producing states in the United States (Crop Prophet 2021),155 and large corn yields could create an opportunity for cellulosic-based renewable energy. At the same time, both states are important producers of coal: a National Mining Association brief from 2014, for example, lists Illinois and Indiana among the top 10 leading coal producers (National Mining Association 2014), and the existence of a powerful coal industry could militate against renewable energy investment.156 In sum, the neighbors’ parallel economic profiles and positions with respect to agriculture and industry suggest that their RPS development patterns might be similar to one another.

. Illinois and Indiana are not only neighboring states located in the Midwest but share a host of similar attributes. Both are states whose economies were heavily dependent on manufacturing (Matejka 1999; Turner 2016); both states were also impacted by deindustrialization in the late twentieth century (Matejka 1999; Turner 2016), arguably creating an opportunity to invest in technologies that could create employment prospects (renewablePage 155 → energy technology has often been billed in this manner). Both states are also appropriately described as agricultural titans: over the twenty-year span between 2000 and 2020, for example, Illinois and Indiana both ranked among the top five corn producing states in the United States (Crop Prophet 2021),155 and large corn yields could create an opportunity for cellulosic-based renewable energy. At the same time, both states are important producers of coal: a National Mining Association brief from 2014, for example, lists Illinois and Indiana among the top 10 leading coal producers (National Mining Association 2014), and the existence of a powerful coal industry could militate against renewable energy investment.156 In sum, the neighbors’ parallel economic profiles and positions with respect to agriculture and industry suggest that their RPS development patterns might be similar to one another.

. Both groups offer ratings of state legislators in their respective states, and the ratings are instructive about how far legislators might be willing to go (including adopting novel and untested policy) in order to make their states clean energy leaders. Ratings range from 0 to 100, where 100 denotes full agreement with either of the two advocacy organizations. Page 157 →If we compare the ratings of the Illinois Environmental Council of Illinois lawmakers who served in a sponsorship or cosponsorship capacity on Illinois’s marquee RPS bill adopted in 2007 with the ratings of the Hoosier Environmental Council of Indiana lawmakers who served in a similar capacity on Indiana’s adopted 2011 RPS bill, we find wide disparities in the level of endorsement given to the subsets of lawmakers from each state.157 While the median endorsement rating given to the Illinois legislators by the Illinois Environmental Council was 90, the median endorsement given to the Indiana legislators by the Hoosier Environmental Council was a paltry 25.158

. Both groups offer ratings of state legislators in their respective states, and the ratings are instructive about how far legislators might be willing to go (including adopting novel and untested policy) in order to make their states clean energy leaders. Ratings range from 0 to 100, where 100 denotes full agreement with either of the two advocacy organizations. Page 157 →If we compare the ratings of the Illinois Environmental Council of Illinois lawmakers who served in a sponsorship or cosponsorship capacity on Illinois’s marquee RPS bill adopted in 2007 with the ratings of the Hoosier Environmental Council of Indiana lawmakers who served in a similar capacity on Indiana’s adopted 2011 RPS bill, we find wide disparities in the level of endorsement given to the subsets of lawmakers from each state.157 While the median endorsement rating given to the Illinois legislators by the Illinois Environmental Council was 90, the median endorsement given to the Indiana legislators by the Hoosier Environmental Council was a paltry 25.158

. The difference in endorsement level not only suggests that the preferences of the groups of lawmakers shepherding RPS bills in each state are probably different—if they were not different, one would expect the ratings given to each state’s group of legislators by advocacy groups with overlapping platforms to be similar—but also suggests that what each state’s legislators want out of RPS legislation will be different. The Illinois legislators are more likely to be in line with and attuned to the objective of making their state a clean energy leader and may embrace novel RPS policy-making out of a belief that doing so is a worthwhile endeavor (in other words, they recognize that making Illinois a leader in clean energy requires the state to lead on renewable policy-making). Judging by the ratings given to them by the Hoosier Environmental Council, Indiana lawmakers on the other hand do not appear to aspire to make their state a national clean energy leader and consequently may be less receptive toward advancing novel RPS policy as a way to build the state’s renewable energy Page 158 →reputation, preferring instead to adopt existing policies that imply that the state supports some renewable energy investment but does not want to jeopardize its fossil-fuel-driven industry.159

. By the mid-to-late 2000s, RPSs began gaining traction among a wider group of states and were adopted more broadly across the United States. A large swath of states stretching from the mid-Atlantic to the nation’s interior, including Pennsylvania (2004), Michigan (2008), Ohio (2008), and West Virginia (2009), adopted their own RPS programs (North Carolina Clean Energy Technology Center 2016). The four states above are noteworthy because they did not merely craft their programs around bolstering conventional (e.g., “clean” with respect to carbon emissions) renewable sources but also included decidedly fossil-fuel-based sources (e.g., “coal mine methane” with respect to Pennsylvania’s RPS) within their RPS programs (e.g., Pennsylvania Legislature 2004). Another later adopter, North Carolina (2007), made concessions to its own locally dominant industries by mandating that, in its RPS program, some electricity be procured using swine and poultry waste (North Carolina Legislature 2007). In fact, there are dozens and dozens of different kinds of sources and technologies (such as microturbines as well as combined heat and power) that states could incorporate into their RPS programs. Other later adopters, such as Indiana (2011), adopted RPS programs but made them voluntary rather than mandatory in orientation (Indiana General Assembly 2011). State RPSs can also differ in terms of how quickly they aspire to reach their target levels.16

. The emphasis by the Indiana lawmakers on careful deliberation also may corroborate the link between electoral vulnerability and borrowing. In the year before Indiana adopted its RPS in 2011, the state’s legislators faced a somewhat skeptical electoral environment, with the median vote share earned by an incumbent in state legislative races being 64.72%, representing a nearly 20 point decrease from a couple of elections prior.160 If Indiana legislators believed that they faced potential skepticism from voters—and the recent drop in vote share suggests that this could have been the case—then those legislators may have tried to project themselves as meticulous analysts who chose wisely from a menu of existing options during the policy adoption process. In the case of Indiana, a state whose population is conservative with a nonetheless sizeable group of moderates who may be drawn to renewable energy due to issues of energy self-sufficiency and environmental conservation, lawmakers would ostensibly try to craft the Page 161 →state’s RPS to include traditional renewable sources while also protecting the state’s nonrenewable energy supply. Indiana lawmakers did just that and constructed an RPS program that was almost entirely borrowed: traditional renewable energy key features, including allowing for wind and solar thermal electric to be renewable, had first been adopted in other states decades before (initially in Iowa and then elsewhere; Iowa legislature 1983), but other less common features—including a voluntary 10% retail standard—also already had been adopted by other states (for example, this already appeared in both North and South Dakota; see House Bill 1506 of the Legislative Assembly of North Dakota in 2007; House Bill 1123 of the South Dakota legislature in 2008). Even Indiana’s signature decision to include nuclear energy within its program had already been pioneered three years before in neighboring Ohio (see Substitute Senate Bill 221 of the Ohio General Assembly in 2008).

. The authors of a recent study identified and tracked the adoption of 29 distinct anti-abortion policies across the U.S. states for a nearly forty-year period starting with the national legalization of abortion in the U.S. Supreme Court’s 1973’s Roe v. Wade decision.161 Taking their data on state-level adoptions of anti-abortion policy, and categorizing and transforming their data to identify cases of invention and borrowing as was discussed in chapter 3 and implemented in chapters 5 and 6, we can ascertain whether increased conservatism predicts invention and whether electoral vulnerability predicts borrowing in the area of anti-abortion policy. The anti-abortion policy adoption data in Kreitzer and Boehmke (2016) come from state legislative decision-making. This means that we are able to provide a conservative analog to the legislative analysis conducted in chapter 5. The conservative analog offers a nice extension to the legislative chapter utilizing RPS—which, given that state legislatures have been the institutional actors most likely to adopt RPS policy, is one of the central analyses of the book. Given that Kreitzer and Boehmke identify legislative policy action and that the leading repository of abortion-related policy devotes its coverage of policy developments to tracking state anti-abortion legislation (Guttmacher Institute 2020), I do not include a conservative analog to the regulatory analysis in chapter 6. In the conclusion of the book, I offer Page 174 →thoughts about how a regulatory analysis could be conducted in a conservative policy area and how the analytical framework employed in this book could be expanded across a host of policy areas.162

. The authors of a recent study identified and tracked the adoption of 29 distinct anti-abortion policies across the U.S. states for a nearly forty-year period starting with the national legalization of abortion in the U.S. Supreme Court’s 1973’s Roe v. Wade decision.161 Taking their data on state-level adoptions of anti-abortion policy, and categorizing and transforming their data to identify cases of invention and borrowing as was discussed in chapter 3 and implemented in chapters 5 and 6, we can ascertain whether increased conservatism predicts invention and whether electoral vulnerability predicts borrowing in the area of anti-abortion policy. The anti-abortion policy adoption data in Kreitzer and Boehmke (2016) come from state legislative decision-making. This means that we are able to provide a conservative analog to the legislative analysis conducted in chapter 5. The conservative analog offers a nice extension to the legislative chapter utilizing RPS—which, given that state legislatures have been the institutional actors most likely to adopt RPS policy, is one of the central analyses of the book. Given that Kreitzer and Boehmke identify legislative policy action and that the leading repository of abortion-related policy devotes its coverage of policy developments to tracking state anti-abortion legislation (Guttmacher Institute 2020), I do not include a conservative analog to the regulatory analysis in chapter 6. In the conclusion of the book, I offer Page 174 →thoughts about how a regulatory analysis could be conducted in a conservative policy area and how the analytical framework employed in this book could be expanded across a host of policy areas.162

. Since I utilize data from other scholars to perform the analyses in this chapter, I first discuss the nature of the Kreitzer and Boehmke data and then expound upon how I transformed this data for the purposes of examining invention and borrowing in anti-abortion policy.163 Kreitzer and Boehmke gathered data on the state-level adoption of any of 29 anti-abortion policies and structured their data so that it could be investigated using pooled event history analysis (Boehmke 2009). Their identification of so many distinct anti-abortion policy choices lets me scrutinize when state legislatures invent and borrow within the anti-abortion policy space.

. I transform Kreitzer and Boehmke’s data using the same procedure and definitions discussed in chapter 3. Invention occurs when a state is the Page 176 →first U.S. state to adopt a particular anti-abortion policy—meaning one of the distinct anti-abortion policies identified by Kreitzer and Boehmke—or if a state adopts that policy by the next year after the first state adopted the policy. Borrowing occurs if a state adopts a policy any time after the year in which the initial adopting state adopted the policy.164 Similar to what was discussed in chapter 3 and implemented in chapters 5 and 6 of this book and elsewhere (Parinandi 2020), the risks of inventing and borrowing a particular anti-abortion policy are disjointed to reflect the idea that inventing and borrowing have mutually exclusive definitions and that policymakers cannot simultaneously entertain the prospect of inventing and borrowing with respect to the same policy.165 Also similar to the discussion in chapter 3 of this book and implementation in chapters 5 and 6 of this book and elsewhere (Parinandi 2020), states gain the opportunity to invent with respect to any of the anti-abortion policies in the year 1973, which is the year corresponding to when Kreitzer and Boehmke start their dataset. The year 1973 witnessed the Supreme Court’s famous Roe v. Wade decision legalizing abortion and also witnessed a slew of state policy-making attempting to limit abortion access within the confines of the Roe v. Wade decision. Allowing for states to have the opportunity to invent the anti-abortion policies starting in 1973 reflects the idea that the Roe v. Wade case ushered in the possibility that states could devise novel ways to limit abortion access given the decision in Roe. Analogously, based on my conceptualization of borrowing, I wait until two years after a first state invented a policy before allowing states to borrow with respect to that same policy. The last year in which I analyze data is 2012, which is also the last year included in Kreitzer and Boehmke’s models.

. I transform Kreitzer and Boehmke’s data using the same procedure and definitions discussed in chapter 3. Invention occurs when a state is the Page 176 →first U.S. state to adopt a particular anti-abortion policy—meaning one of the distinct anti-abortion policies identified by Kreitzer and Boehmke—or if a state adopts that policy by the next year after the first state adopted the policy. Borrowing occurs if a state adopts a policy any time after the year in which the initial adopting state adopted the policy.164 Similar to what was discussed in chapter 3 and implemented in chapters 5 and 6 of this book and elsewhere (Parinandi 2020), the risks of inventing and borrowing a particular anti-abortion policy are disjointed to reflect the idea that inventing and borrowing have mutually exclusive definitions and that policymakers cannot simultaneously entertain the prospect of inventing and borrowing with respect to the same policy.165 Also similar to the discussion in chapter 3 of this book and implementation in chapters 5 and 6 of this book and elsewhere (Parinandi 2020), states gain the opportunity to invent with respect to any of the anti-abortion policies in the year 1973, which is the year corresponding to when Kreitzer and Boehmke start their dataset. The year 1973 witnessed the Supreme Court’s famous Roe v. Wade decision legalizing abortion and also witnessed a slew of state policy-making attempting to limit abortion access within the confines of the Roe v. Wade decision. Allowing for states to have the opportunity to invent the anti-abortion policies starting in 1973 reflects the idea that the Roe v. Wade case ushered in the possibility that states could devise novel ways to limit abortion access given the decision in Roe. Analogously, based on my conceptualization of borrowing, I wait until two years after a first state invented a policy before allowing states to borrow with respect to that same policy. The last year in which I analyze data is 2012, which is also the last year included in Kreitzer and Boehmke’s models.

. Categorizing and transforming the Kreitzer and Boehmke data yields two dependent variables of interest: legislative invention and legislative borrowing.Page 177 → There are 104 positive occurrences of the legislative inventing variable (with 18,782 opportunities to invent) and 568 positive occurrences of the legislative borrowing variable (with 26,831 opportunities to borrow).166 The unit of analysis here is anti-abortion policy-state-year. As with chapter 5, my key independent variables for both dependent variables are government ideology and median incumbent vote share.167 In the models pertaining to both dependent variables, I include all of the right-hand-side variables utilized in the Kreitzer and Boehmke article. Variables include measurements of “public preferences” toward abortion (including state-level conservative attitudes toward abortion [Norrander] as well as state-level religious adherence rates);168 party-related variables including whether a state has a unified Democratic legislature, the amount of a state legislature consisting of female Democrats, and whether a state has a Democratic governor; and a variable related to the whether it is difficult to launch initiatives in a state (Kreitzer and Boehmke 2016, 134).

. Categorizing and transforming the Kreitzer and Boehmke data yields two dependent variables of interest: legislative invention and legislative borrowing.Page 177 → There are 104 positive occurrences of the legislative inventing variable (with 18,782 opportunities to invent) and 568 positive occurrences of the legislative borrowing variable (with 26,831 opportunities to borrow).166 The unit of analysis here is anti-abortion policy-state-year. As with chapter 5, my key independent variables for both dependent variables are government ideology and median incumbent vote share.167 In the models pertaining to both dependent variables, I include all of the right-hand-side variables utilized in the Kreitzer and Boehmke article. Variables include measurements of “public preferences” toward abortion (including state-level conservative attitudes toward abortion [Norrander] as well as state-level religious adherence rates);168 party-related variables including whether a state has a unified Democratic legislature, the amount of a state legislature consisting of female Democrats, and whether a state has a Democratic governor; and a variable related to the whether it is difficult to launch initiatives in a state (Kreitzer and Boehmke 2016, 134).

. Categorizing and transforming the Kreitzer and Boehmke data yields two dependent variables of interest: legislative invention and legislative borrowing.Page 177 → There are 104 positive occurrences of the legislative inventing variable (with 18,782 opportunities to invent) and 568 positive occurrences of the legislative borrowing variable (with 26,831 opportunities to borrow).166 The unit of analysis here is anti-abortion policy-state-year. As with chapter 5, my key independent variables for both dependent variables are government ideology and median incumbent vote share.167 In the models pertaining to both dependent variables, I include all of the right-hand-side variables utilized in the Kreitzer and Boehmke article. Variables include measurements of “public preferences” toward abortion (including state-level conservative attitudes toward abortion [Norrander] as well as state-level religious adherence rates);168 party-related variables including whether a state has a unified Democratic legislature, the amount of a state legislature consisting of female Democrats, and whether a state has a Democratic governor; and a variable related to the whether it is difficult to launch initiatives in a state (Kreitzer and Boehmke 2016, 134).

. Other variables included in their analysis are the median income of a state and the population of a state. Kreitzer and Boehmke also include variables corresponding to the influence of time: these include a binary variable capturing whether or not a state is making a decision after the Webster v. Reproductive Health Services case of 1989, a linear year variable, and a nonlinear year squared variable.169 The authors lastly include a neighbors variablePage 178 → capturing the amount of geographically adjacent states that adopted the same anti-abortion policy. As with earlier chapters in the book, I utilize logistic regression with standard errors clustered by state to investigate legislative inventing and borrowing.

. Figure 2 largely corroborates the diffusion outline expressed in the past few pages. While the starting point of Iowa’s action in 1983 makes sense given that this was the first instance of a state adopting a prototypical RPS program, I would like to elaborate about why the year 2011 makes a worthwhile end point. First, choosing the year 2011 provides observers with nearly three decades’ worth of data on RPS policy-making. Included Page 24 →in this nearly thirty-year span is the adoption and (if applicable) amendment experiences of 37 states, which collectively make up 74% of all states in the United States. The broadness of the states adopting RPSs—coming from almost every region of the country, including urban and rural states, and including states with varying industrial makeups—is beneficial to minimize a situation where research findings are simply an artifact of a few states experimenting with the issue area (Parinandi, Langehennig, and Trautmann 2020).17

. In table 8, I display empirical results pertaining to legislative inventing and borrowing. Model specifications 1 and 2 correspond to using the variables from Kreitzer and Boehmke as controls but adding government ideology and median incumbent vote share. Model specifications 3 and 4 contain the same variables as in specifications 1 and 2, but here I add additional controls employed in earlier chapters of the book. These additional controls include Berry et al.’s measure of citizen ideology, the fraction of anti-abortion policy adoptions that have occurred in ideologically similar states to a given state, the fraction of prior invention in anti-abortion policy that has occurred in a state (for the invention specification), and the fraction of prior borrowing in anti-abortion policy that has occurred in a state (for the borrowing specification). Also for the borrowing specification, I include a featureyear variable capturing the number of years that have elapsed since a given anti-abortion policy was first adopted across the states.170

. A quick glance at table 8171 reveals that the government ideology variable relates negatively and is statistically significant with respect to invention in both invention model specifications but lacks such a statistically significant Page 179 →relationship with respect to borrowing in both borrowing model specifications. Figure 10 displays the relationship between government ideology with respect to both dependent variables. The figure utilizes model specifications 3 and 4 (which include more control variables) and sets binary control variables to their most frequently occurring values.

. Figure 11 corroborates the directionality of the trends unearthed in table 8. Rising median incumbent vote share (denoting reduced electoral vulnerability) makes anti-abortion invention less likely but increases the probability of anti-abortion borrowing. The divergent directions of slope here are different than in the case of RPS legislation, when the slope of median incumbent vote share traveled in the same direction for invention and borrowing.172 More generally, returning to the tables displaying Page 183 →multivariate regression results and comparing the electoral vulnerability findings from table 3 in chapter 5 to table 8 in this chapter—and specifically pointing to the observation that the median incumbent vote share variable relates significantly to borrowing but not invention for RPS while the median incumbent vote share variable relates significantly to invention but not to borrowing for anti-abortion policy—suggests that the role of electoral vulnerability in fostering invention and borrowing might differ by policy type. I turn next to explaining why results were consistent for government ideology across RPS and anti-abortion policies and why they were different for electoral vulnerability across RPS and anti-abortion policies.

. Looking at anti-abortion policy and comparing the statistical association between government conservatism (as seen by lower levels of the government ideology variable) and invention versus government conservatism and borrowing in table 8, we see the same dynamic at play as in the RPS case in the sense that government ideology (except here, it is conservatism and not liberalism) has a more pronounced statistical association with invention than it does with borrowing. As was the case with RPS, here, there is anecdotal evidence suggesting that a wider ideological array of states embrace anti-abortion borrowing compared to the ideological array of states embracing anti-abortion invention. If we compare anti-abortion invention and borrowing activity among states that have government ideology scores in the highest one-third of the government ideology score (implying that these states are quite liberal), we see that 21.4% instances of borrowing occur in this grouping of states compared to 15.3% of instances of invention.173 If we compare anti-abortion invention and borrowing activity among the most liberal states in the United States—those states with government ideology scores between 90 and 100—we see that the percentage of instances of inventing that occur among this group of very liberal states is only 0.9% while the percentage of instances of borrowing occurring among this group is 2.9%.174 This suggests that anti-abortion Page 185 →borrowing is more accessible to an ideologically wider group of states than is anti-abortion inventing.

. Looking at anti-abortion policy and comparing the statistical association between government conservatism (as seen by lower levels of the government ideology variable) and invention versus government conservatism and borrowing in table 8, we see the same dynamic at play as in the RPS case in the sense that government ideology (except here, it is conservatism and not liberalism) has a more pronounced statistical association with invention than it does with borrowing. As was the case with RPS, here, there is anecdotal evidence suggesting that a wider ideological array of states embrace anti-abortion borrowing compared to the ideological array of states embracing anti-abortion invention. If we compare anti-abortion invention and borrowing activity among states that have government ideology scores in the highest one-third of the government ideology score (implying that these states are quite liberal), we see that 21.4% instances of borrowing occur in this grouping of states compared to 15.3% of instances of invention.173 If we compare anti-abortion invention and borrowing activity among the most liberal states in the United States—those states with government ideology scores between 90 and 100—we see that the percentage of instances of inventing that occur among this group of very liberal states is only 0.9% while the percentage of instances of borrowing occurring among this group is 2.9%.174 This suggests that anti-abortion Page 185 →borrowing is more accessible to an ideologically wider group of states than is anti-abortion inventing.

. Moral policies (Mooney and Lee 1999) are viewed within the nexus of “good versus evil” and tend to resonate with voters insofar as voters want lawmakers to take a stand on the side of good against evil. Along the same vein, moral issues may elicit starkly dichotomous opinions and deemphasize context in favor of black and white thinking. For instance, an opponent of abortion may look at abortion as a purely binary (right versus wrong) issue and dismiss context related to the issue. That same binary worldview may also cause an abortion opponent to downweight the importance of the track record as a skepticism-reducing tool. This is because the abortion opponent intrinsically believes that adopting policy making abortion harder to obtain (the 29 anti-abortion policies documented by Kreitzer and Boehmke all make abortion harder to obtain) is valuable and does not need to be persuaded by the experiences of other states. In other words, needing the experiences of other states to confirm one’s support for a policy suggests that a person is somewhat undecided Page 186 →about that policy. If that person subscribes to a worldview where abortion is evil and should be made more difficult to obtain (as many adherents in the anti-abortion movement do), then that person is unlikely to find other states’ experiences to be of much use in influencing their support for anti-abortion policy.175

. At the same time, the abortion opponent may be especially receptive to the signal that lawmakers are making extraordinary commitments to advancing the cause of restricting abortion. One way lawmakers can send a signal to abortion opponents that they are making extraordinary commitments is by adopting novel anti-abortion policy. Recall from chapter 5 that I discussed how North Dakota’s legislature made that state the very first in the country to ban abortion when a fetus is shown to have some type of genetic abnormality (Eligon and Eckholm 2013). Commenting about the legislation, recall that Paul Maloney of North Dakota Right to Life (an anti-abortion organization) remarked that “we have the kind of legislative body that would pass these kinds of pieces of legislation” (quoted in Eligon and Eckholm 2013: A13). The comment from Maloney potentially reflects a view that North Dakota’s legislature has the courage to adopt anti-abortion policies that other states had previously not countenanced. And taken more broadly, the comment suggests that abortion opponents may look favorably upon unprecedented attempts to advance their cause.176

. Lawmakers, for their part, generally know what their constituents want and most likely would know if anti-abortion constituents are receptive to invention as a signal of issue commitment (Fenno 1978). If lawmakers regard anti-abortion opponents to be an integral part of their voter coalition, they may be more likely to attempt to invent anti-abortion policy as a way to cultivate anti-abortion bona fides with anti-abortion-minded constituents.177 Given that such lawmakers may view invention as a method to Page 187 →demonstrate their bona fides to anti-abortion constituents, increased electoral vulnerability may make these lawmakers more likely to signal their commitment to anti-abortion voters through invention. And insofar as invention is a stronger way to signal commitment to anti-abortion voters than is borrowing—which is plausible if we interpret comments from the North Dakota Right to Life organization to mean that this organization is especially proud that the North Dakota legislature has the courage to adopt unprecedented and novel anti-abortion policy—then it is possible that increased electoral vulnerability could exert greater influence on anti-abortion invention compared to anti-abortion borrowing.178

. Lawmakers, for their part, generally know what their constituents want and most likely would know if anti-abortion constituents are receptive to invention as a signal of issue commitment (Fenno 1978). If lawmakers regard anti-abortion opponents to be an integral part of their voter coalition, they may be more likely to attempt to invent anti-abortion policy as a way to cultivate anti-abortion bona fides with anti-abortion-minded constituents.177 Given that such lawmakers may view invention as a method to Page 187 →demonstrate their bona fides to anti-abortion constituents, increased electoral vulnerability may make these lawmakers more likely to signal their commitment to anti-abortion voters through invention. And insofar as invention is a stronger way to signal commitment to anti-abortion voters than is borrowing—which is plausible if we interpret comments from the North Dakota Right to Life organization to mean that this organization is especially proud that the North Dakota legislature has the courage to adopt unprecedented and novel anti-abortion policy—then it is possible that increased electoral vulnerability could exert greater influence on anti-abortion invention compared to anti-abortion borrowing.178

. I bring up the above example because it shows that RPS policy has a greater diversity of purpose than does anti-abortion policy. This is to say, RPS policy elicits a wider range of objectives about what that policy will programmatically do compared to anti-abortion policy. This wider range of objectives itself matters because it suggests that proponents of RPS policy are less binary in their thinking about RPS than are anti-abortion proponents regarding abortion. The idea here is that binary thinking about an issue—and remember that a hallmark of a moral issue is that it tends to encourage binary thinking about that issue—may lead to narrow views about the kinds of objectives that should be pursued when policy is crafted to deal with that issue. In contrast, if thinking about an issue is more diverse, there may be a concomitant diversity of objectives pursued in policy-making dealing with that issue.179 Given that the RPS area exhibits a greater diversity of objectives than does anti-abortion policy, and given that less diversity of objectives may be associated with binary conceptualization of an issue (which again is more likely in moral policy), then the designation of an area as moral may be more appropriate with anti-abortion policy rather than with RPS. There is public opinion data that supports the idea that abortion is more widely regarded as a moral issue compared to renewable energy. Research from Hess et al. (2016) reveals that the attitudes of Americans are relatively open regarding renewable energy; at the same time, work from Bowman and Marisco (2014) argues that the attitudes of Americans concerning abortion have largely not changed since the Roe v. Wade decision. The stickiness of attitudes regarding abortionPage 189 → combined with the more flexible and open-ended nature of attitudes regarding renewable energy suggests that the public may be more receptive to learning about renewable energy policy-making and consequently more swayable by evidence, thereby giving added credence to the claim that observability matters more for renewable energy compared to abortion.

. A second reason for considering the year 2011 to serve as the study’s end point is that RPS programs had largely concluded diffusing across the system of U.S. states by that year. If 2004 and 2007 represent the years when the most states initially adopted RPS programs, 2011 is perhaps characterized appropriately as a year by which RPS programs were approaching their limit of cross-state U.S. expansion. Only one state that had never before adopted any kind of RPS program as of 2011 adopted one between that year and the present time.18 The other nonadopting states as of 2011—Alabama, Alaska, Arkansas, Florida, Georgia, Idaho, Kentucky, Louisiana, Mississippi, Nebraska, Tennessee, and Wyoming—had still never adopted an RPS program as of 2021.

. Highlighting how anti-abortion may be considered to be more of a moral issue area than RPS allows us to return to the discussion of track record and how it might matter differentially with regard to how electoral vulnerability influences adoption in RPS versus anti-abortion policy. In a less moral issue area such as RPS, the public may lack concrete opinions about whether policy-making is right or wrong, which opens up the possibility that the public is persuadable using evidence. Given that this is a policy area where the public is receptive to evidence, and given that increased electoral vulnerability might create a view among lawmakers that the public is skeptical, increased electoral vulnerability may increase the likelihood of borrowing, as lawmakers hope to assuage the public by claiming that they are replicating experiments tried elsewhere. In a more moral issue area such as anti-abortion, however, the public may be more likely to have ex-ante opinions of right versus wrong and be less amenable to persuasion through evidence. Given that this is a policy area where the public is less receptive to evidence, the way that lawmakers might deal with electoral vulnerability is not to invoke evidence but rather to double down on moral appeals in hopes of convincing moral-issue-inclined members of the public that the lawmakers are committed to advancing the moral cause by adopting unprecedented policy that moves the moral cause forward.180

. There are some clarification issues I would like to address. First, the observation that lawmakers operating under increased electoral vulnerability are more likely to invent to signal commitment to anti-abortion voters does not mean that such behavior is entirely absent with respect to RPS policy. In table 3 in chapter 5, the median incumbent vote share is negatively related to RPS invention, suggesting that there is some effort to demonstrate commitment to constituents that especially desire an RPS. Page 190 →However, the significance of the median incumbent vote share variable with respect to anti-abortion invention but not RPS invention suggests that electoral vulnerability explains anti-abortion invention better than it does RPS invention. One reason why this might be the case, to use the logic I have put forth in this chapter, is that the relative unanimity of objectives surrounding anti-abortion policy makes it more likely that abortion opponents will believe that an unprecedented policy will move their cause forward (with the requisite verdict, of course, that lawmakers are displaying courage in a novel way). Less unanimity of objective in RPS suggests that lawmakers would get less of a bump from dedicated environmentalists with respect to RPS than lawmakers would receive from dedicated abortion opponents with respect to anti-abortion policy, which in turn might explain why the median incumbent vote share variable is linked to anti-abortion invention but not RPS invention.181

. One potential avenue for future inquiry corresponds with applying this framework to other moral policy areas, including liberal manifestations of abortion policy.182 Another extension involves decomposing this analysis across different stages of the legislative process; I have offered a macro-level Page 193 →analysis here, but a micro-level legislator-centered investigation (where the individual legislator rather than the state legislature is the unit of analysis) can pinpoint how this process occurs at the individual level. An added benefit of the individual analysis is that it would allow us to ascertain how the influences of ideology and electoral vulnerability drive legislator behavior across the various stages of the legislative process (e.g., sponsorship, cosponsorship, committee deliberation, and floor voting). While I analyze invention and borrowing through adoption, adoption is preceded by other important legislative actions, and studying these earlier actions can shed light on the chronology of invention and borrowing. A third possibility motivated by the abortion case involves pushing on the concept that morality policy is interpreted through the binary lens of right and wrong. Although abortion may plausibly still be perceived in this binary manner, other policies that were traditionally considered to be moral (for example, gambling) have become widespread and may have ostensibly lost their moral connotation. It would be fascinating to evaluate whether the electoral dynamics surrounding an issue like gambling change as the issue loses its moral connotation.183 Taken alongside the rest of the book, the findings from this chapter can help augment a future agenda on invention and borrowing. In the book’s conclusion, I elaborate on that agenda in greater detail.

. One potential avenue for future inquiry corresponds with applying this framework to other moral policy areas, including liberal manifestations of abortion policy.182 Another extension involves decomposing this analysis across different stages of the legislative process; I have offered a macro-level Page 193 →analysis here, but a micro-level legislator-centered investigation (where the individual legislator rather than the state legislature is the unit of analysis) can pinpoint how this process occurs at the individual level. An added benefit of the individual analysis is that it would allow us to ascertain how the influences of ideology and electoral vulnerability drive legislator behavior across the various stages of the legislative process (e.g., sponsorship, cosponsorship, committee deliberation, and floor voting). While I analyze invention and borrowing through adoption, adoption is preceded by other important legislative actions, and studying these earlier actions can shed light on the chronology of invention and borrowing. A third possibility motivated by the abortion case involves pushing on the concept that morality policy is interpreted through the binary lens of right and wrong. Although abortion may plausibly still be perceived in this binary manner, other policies that were traditionally considered to be moral (for example, gambling) have become widespread and may have ostensibly lost their moral connotation. It would be fascinating to evaluate whether the electoral dynamics surrounding an issue like gambling change as the issue loses its moral connotation.183 Taken alongside the rest of the book, the findings from this chapter can help augment a future agenda on invention and borrowing. In the book’s conclusion, I elaborate on that agenda in greater detail.

. One of the mysteries about novel policy adoption centers on the extent to which ideology influences legislatures (the chief lawmaking branch) to invent policy: Is novel policy-making the province of committed ideologues who believe in the worth of advancing policy into proverbially uncharted territory, or is there a less clear role for ideology? The analysis of RPS adoption, a left-leaning area (Potrafke 2010), provides a window into addressing this question from the liberal direction.184 My finding linking increased liberalism to increased RPS invention suggests that at least for renewable energy policy, it is the ideologues (in this case, liberals) who push for policy to be advanced into uncharted territory and who are more likely to believe that such action is worthwhile even though analogues or templates do not exist for doing so. My finding showing the lack of a clear connection between increased liberalism and RPS borrowing suggest that an ideologically diverse grouping of state legislatures is willing to embrace a policy once it has been tested by a pioneering state.

. These findings have substantial implications for how we understand novel policy adoption in renewable energy policy. Assuming that observers want more invention and pioneering policy-making in the area of renewables, then ideological extremism in the liberal direction is beneficial insofar as it leads to novel policies entering the menu of renewable portfolio policies employed across the system of states. The fact that ideology is less influential in explaining borrowing is also beneficial in that RPS policy Page 198 →adoption is not limited to ideologues: less liberal states are willing to adopt RPS policy as long as it has been tried elsewhere, and this augurs well for the spread or diffusion of RPS policy-making across an ideologically heterogeneous breadth of states.185 To sum up the findings harkening back to Brandeis’s terminology, in the area of RPS, “courageous” states are those characterized by ideological liberalism while other states benefit from the courage of the former group. This courage is presumably a normative good if one supports RPS policy, but it is a normative bad—note that recklessness is the antonym of courage—if one opposes such policy.

. Insofar as policymakers, scholars, and the public will continue to look to the individual states to serve as substitutes for federal-level dysfunction, it is crucial for us to learn more about the broad inventive capacity of the American states. Given the widespread feeling that federal dysfunction shows no sign of abating any time soon, the time to shed more light on Brandeis’s prophetic opinion is now.186 In the pages that follow, I elaborate on how spillover projects that either use or are based on the framework developed here can deepen our knowledge of state policy invention and act as a springboard for future practitioners and scholars of American politics, public policy, and state politics and policy to conceptualize and weigh novel state-based policy solutions. The same spillover projects could also ultimately help usher in a new generation of state policy-making related research.

. In the book, I studied RPS policy adoption for a number of solid reasons: it deals with an important issue where federal action has largely been lacking; it features state-level policy adoption, allowing for the analysis of invention put forth in this book; and it features variation in invention by institution, allowing for the analysis of legislative and regulatory invention. The gathering, coding, and processing of the RPS data was a painstaking endeavor Page 203 →and took close to half a decade to accomplish. I was able to integrate the anti-abortion data from Kreitzer and Boehmke (2016) only because this data was already in subpolicy form, had already been identified as being legislative in nature (allowing me to use this data to analyze legislative invention), and was seen as emblematic of the states devising their own solutions given a lack of federal clarity (allowing me to continue along the path of analyzing how the states invent when they take policy-making into their hands).187

. One may wish to extend the framework devised here across a greater diversity of policy areas, and such an extension is a worthwhile effort. A group of prominent diffusion scholars recently compiled a massive dataset called the “State Policy Innovation and Diffusion,” or “SPID” dataset (Boehmke et al. 2018). SPID encompasses adoption data on 728 policies adopted by the states over more than a century of time. Applying my framework to SPID would help to establish general patterns (across a tremendous diversity of areas) of invention. However, there are several processing steps that need to be accomplished before such an analysis can occur. Adoption data in SPID that is gathered at the policy level would have to be gathered at the policy feature or subpolicy level so that we could identify invention and borrowing within each of the policies comprising the SPID dataset. Then, in order to evaluate legislative or regulatory invention using this massive dataset, lawmaking documents for each instance of adoption for the 728 policies would have to be gathered so we can identify what institution is behind a given adoption. Information about which institutions are responsible for the adoptions in SPID are not currently available, would be labor intensive to acquire, and would represent a major breakthrough that would facilitate the use of my framework with the SPID data. If one wants to extend the number of areas but preserve the focus on state invention given a lack of federal intervention, one would need to find and take out policies from SPID that featured federal intervention. Finally, after these steps have been completed, instances of invention and borrowing have been identified and coded as such, and pooled event history data structures have been created for legislative and regulatory invention, the invention framework I advance here can be used in conjunction with Page 204 →SPID.188 Extending the framework developed here through using SPID represents a potentially large advance for the field of state politics and is a natural follow-up to this project.

. A major motivator for this project and, frankly, for the use of RPS as an area of study, is that state governments can take matters into their own hands given federal inaction or dysfunction and adopt novel solutions to pressing challenges. RPS is arguably not the only area where the federal government has left it to the states to devise their own solutions to challenges. Historically, one could look at state policy experimentation in the 1950s and 1960s prior to the passage of the federal Clean Air Act or the creation of the Environmental Protection Agency (this state experimentation served as a major inspiration for federal clean air policy-making). In the present, one could look at the spate of state policy experimentation with respect to recreational marijuana legalization, which has arguably featured a lack of federal intervention insofar as the federal government has chosen not to enforce its laws against recreational marijuana use in states that have undergone legalization. State-level experimentation with stay-at-home or lockdown policy-making in the early stages of the COVID-19 pandemic also potentially fits the bill, as the federal government neglected to coordinate a policy response across the states (Kettl 2020). Finally, persistent federal dysfunction (Binder 2015) may make the occurrence of letting states handle their own policy-making more likely. Thus, the topic of investigating state policy-making when the federal government does not intervene is not limited to RPS but extends to other pressing areas where the states represent the forefront of policy advancement within the American federal system.189

. Although it is possible that dynamics similar to those driving invention and borrowing in the adoption of policy could also play a role in accounting for hypothetical invention and borrowing in the abolition or termination of policy, the field of political science has typically isolated the study of adoption (which is ensconced firmly within literature on how Page 25 →adopted policies spread or diffuse as typified by Walker 1969; Gray 1973; Berry and Berry 1990; Shipan and Volden 2006; and Boushey 2010) from the much less well-established study of policy abolition or termination (one example of this nascent literature can be found in Volden 2015). This monograph lies within the larger tradition of studying adoption and seeks to investigate how inventing occurs across different institutional actors in the course of adoption using a policy area that has diffused across a broad number of states. The revelations in this monograph are meant primarily to inform future work on adoption and could potentially inform our knowledge about how invention occurs in non-RPS related policy-making that may be diffusing across the U.S. states. In advice intended to guide future scholarship in the conclusion of the book, I discuss how the central framework utilized in this study could also be applied to policy abolition and also discuss how scholars can begin the arduous yet important task of extending the framework across a diverse range of policy areas. In the meanwhile, I encourage readers interested in learning more about the recent freezing and repeal of RPS policy to consult the pathbreaking work of Stokes (2020).19

. The extension just discussed dealt with using the framework that I have devised to investigate state invention in policy areas where the federal government actively intervenes in state policy-making. However, my project can also stimulate research into how inventions by state governments influence later federal-level policy-making. Work by Bulman-Pozen (2014) and Orren and Skowronek (2017) suggests that state-level inventions inform federal policy-making, and findings from this project can be harnessed to further explore this phenomenon.190 Federal-level attempts to meaningfully intervene in setting renewable energy development imperatives in state-specific electricity sectors, whether in the form of the Clean Power Plan or some version of the Green New Deal, have failed to be successfully adopted. Should they ultimately be adopted in some form—which is possible given Democratic control of Congress and the arrival of the Biden administration—we would have a golden opportunity to study how state inventions impact later federal policy-making. If some version of the Clean Power Plan or Green New Deal were to be adopted, we could identify feature-level components of this policy-making, trace whether any of these feature-level components were previously invented by state governments as part of their RPS policy-making, and then evaluate which states are the ones that are most likely to be followed by the federal government. This extension both complements the work of Bulman-Pozen and Orren and Skowronek and addresses the timely and important issue of how state-level experimentation can eventually shape national policy.

. Broader inquiries may presumably be necessary so long as the federal government continues to exhibit dysfunction and inadequacy: I have Page 216 →detailed when and under what conditions invention has occurred with respect to state renewable portfolio adoption, and the results of this exploration can be used by interested observers to pinpoint where we should be focusing our attention if we want to detect instances of novel renewable-energy-related policy-making.191 Furthermore, the institution-specific focus may help observers decide that some paths may or may not be worth pursuing: the potential that public utilities commissions, for example, are so concerned with pushback from electric utility companies may give pause to those who think that public utilities commissions are the natural outlet from which renewable energy policy should be promulgated. Studies of a similar nature may be necessary to help guide observers in areas of state policy-making ranging from immigration to firearm safety to autonomous vehicles. These studies provide clues as to how the federal system can address a host of unanswered challenges but also move the study of adoption toward a recognition that there are different kinds of adoption that are worth exploring in their own right. In this book, I have developed the language as well as the conceptual and empirical tools to make that recognition clear, and I have outlined a suite of ideas that could advance our knowledge about the inventive capacity of the states and interrogate Brandeis’s insight into the foreseeable future.

. Cutting across the ideological spectrum, journalists also invoke state leadership on policy-making as an antidote to federal inaction or even disagreement with the stated federal position on policy. An author in Vox, a news and analysis website, suggested that policy-making autonomy at the state level could be marshalled to dilute the Trump administration’s immigration agenda (Gerken 2017). Commenting from a decidedly dissimilar ideological perspective in the newsmagazine Reason, a different author claimed that Obamacare stifled state-level autonomy and thus “undermines a core principle of American federalism” (Staley 2010). The ability of the concept of state-level policy-making autonomy to gain adherents from different ideological dispositions speaks to the existence of a widespread faith in the power of the states to solve pressing challenges and is arguably an antidote to the paralysis of federal polarization insofar as the individual state governments are more likely to be ideologically homogenous compared to the federal government.2 This faith is mirrored in differences in reported public sentiment in the competence of federal versus state governments to handle problems. Plotting the percentage of respondents in 10 years of Gallup polls who signal that they have a “great deal” of trust in the ability of the federal government to handle domestic problems alongside the percentage of respondents who signal that they have a “great deal” of trust in the ability of state governments to handle state problems gives a rough measure of how some of the public ranks the relative competence of federal versus state governments.3 In figure 1, I provide this plotted comparison using Gallup poll results collected in the 2010–2020 interval (Gallup 2020).

. Figure 3 displays the familiar “S-curve” that is ubiquitous in studies of policy diffusion (Rogers 1962; and Boushey 2010). According to this curve, as time elapses, a greater number of states are likely to adopt a given program, such as RPS. While many previous studies of diffusion (for example, Rogers 1962) focus on how a program may drift through the population of states based on the S-curve, I am less concerned with the S-curve here and instead seek to evaluate why a state would invent within a program (at the subpolicy level) even if that state is adopting a program that has already been adopted in other states. Even a late program adopting laggard (such as West Virginia in the RPS space) can invent, and I want to explain that invention.20

. Page 26 →In spite of the fact that RPSs have largely stopped diffusing to new states and that some states have made attempts to dismantle their RPS infrastructure, RPSs are arguably the most widespread and durable policies falling under the aegis of clean energy to have been adopted in the United States (Allison and Parinandi 2020). Furthermore, even though the federal government was unable to implement the Clean Power Plan that it devised in the mid-2010s (recall that the federal government had a laissez-faire attitude toward the issue of fostering renewable energy development in the nation’s electricity network until the attempted promulgation of the Clean Power Plan),21 the federal plan overwhelmingly borrowed ideas from state RPS programs (United States Environmental Protection Agency 2015). Insofar as the federal government might revisit implementation of the Clean Power Plan in the years ahead, we can rest assured that state-led RPS experimentation will form much of the bedrock of any future federal clean energy electricity mandate. Evaluating the novel policy creation that took place when RPSs became the preferred renewable energy development instrument across the U.S. states thus might help us make sense of the origination of policies that could comprise a future federal electricity Page 27 →agenda.22 In the next section of this chapter, I move on to discussing how that different institutional actors played a role in crafting state RPS policy.

. Page 26 →In spite of the fact that RPSs have largely stopped diffusing to new states and that some states have made attempts to dismantle their RPS infrastructure, RPSs are arguably the most widespread and durable policies falling under the aegis of clean energy to have been adopted in the United States (Allison and Parinandi 2020). Furthermore, even though the federal government was unable to implement the Clean Power Plan that it devised in the mid-2010s (recall that the federal government had a laissez-faire attitude toward the issue of fostering renewable energy development in the nation’s electricity network until the attempted promulgation of the Clean Power Plan),21 the federal plan overwhelmingly borrowed ideas from state RPS programs (United States Environmental Protection Agency 2015). Insofar as the federal government might revisit implementation of the Clean Power Plan in the years ahead, we can rest assured that state-led RPS experimentation will form much of the bedrock of any future federal clean energy electricity mandate. Evaluating the novel policy creation that took place when RPSs became the preferred renewable energy development instrument across the U.S. states thus might help us make sense of the origination of policies that could comprise a future federal electricity Page 27 →agenda.22 In the next section of this chapter, I move on to discussing how that different institutional actors played a role in crafting state RPS policy.

. RPS programs operate via the electricity sector, and one noteworthy aspect about this sector is that different institutional actors have historically had the ability to adopt policy governing this sector. By virtue of being designated as the chief lawmaking branch and having a concomitant responsibility to use the lawmaking process to cater to the interests of their constituents (Mayhew 1974; Fenno 1978; Binder 1999; Kousser 2005; Volden and Wiseman 2014), legislatures have naturally taken a leading role in crafting policy about an issue as important as renewable energy. And yet legislatures are not the only institutional actors that can claim a mandate to directly enact policy regulating a given state’s electricity sector and specifying how firms operating within that sector should behave. Public utilities regulatory commissions were predominantly set up across the U.S. states in the early twentieth century (Troesken 2006). These commissions were largely set up with the blessing of electric utility companies, which feared that they were being fleeced by municipal governments and thought that state-level regulation would be more even-handed (Troesken 2006).23

. Consequently, based on being able to mediate the participation of electric utility firms in a state’s electricity sector, state public utilities commissions have possessed the jurisdictional authority to adopt RPSs. The RPS policy space, then, merits attention insofar as multiple institutional actors had the opportunity to adopt RPSs; the RPS adoption experiences of the U.S. states confirm institutional heterogeneity in RPS involvement, as some states’ RPS programs were crafted wholly by legislatures, other states’ programs were crafted wholly by public utilities commissions, and still other states’ programs were crafted by a mix of legislative and public Page 29 →utilities commission activity.24 In figure 4, I display the breakdown of RPS policy feature adoptions by state based on the type of institutional actor doing the adopting.

. Figure 425 reveals the distribution of the different institutional adopters of RPS policy features across the U.S. states. One immediate takeaway is the relative ubiquity of legislative adoption, with most states adopting most of their policy features legislatively. This probably makes sense given that the legislative branch is considered the chief policy crafting branch and given that energy-related issues represent a potentially important area of legislative policy-making. The preponderance of legislative adoption comports with the traditional literature on state policy adoption (e.g., Berry and Berry 1990; Boushey 2010; Boehmke and Skinner 2012; Parinandi, Langehennig, and Trautmann 2020) that characterizes adoption as a primarily legislative endeavor. What this book adds to that literature is a systematic way to identify novel adoption (invention), distinguish it from already-tested adoption (borrowing), evaluate what makes legislatures more likely to invent policy, and discern whether the factors influencing invention are similar to those influencing borrowing.26 To do this effectively, I identify the components of an RPS program that could be adopted (I call these policy features), and I then determine whether these instances of policy feature adoption are either invention or borrowing. Concrete examples of policy feature adoption include sources that a state considers to be renewable (such as solar thermal energy) as well as the final target rate of an RPS (say, requiring 30% of a utility’s procured electricity to be from renewable sources).

. Figure 425 reveals the distribution of the different institutional adopters of RPS policy features across the U.S. states. One immediate takeaway is the relative ubiquity of legislative adoption, with most states adopting most of their policy features legislatively. This probably makes sense given that the legislative branch is considered the chief policy crafting branch and given that energy-related issues represent a potentially important area of legislative policy-making. The preponderance of legislative adoption comports with the traditional literature on state policy adoption (e.g., Berry and Berry 1990; Boushey 2010; Boehmke and Skinner 2012; Parinandi, Langehennig, and Trautmann 2020) that characterizes adoption as a primarily legislative endeavor. What this book adds to that literature is a systematic way to identify novel adoption (invention), distinguish it from already-tested adoption (borrowing), evaluate what makes legislatures more likely to invent policy, and discern whether the factors influencing invention are similar to those influencing borrowing.26 To do this effectively, I identify the components of an RPS program that could be adopted (I call these policy features), and I then determine whether these instances of policy feature adoption are either invention or borrowing. Concrete examples of policy feature adoption include sources that a state considers to be renewable (such as solar thermal energy) as well as the final target rate of an RPS (say, requiring 30% of a utility’s procured electricity to be from renewable sources).

. The point that both state legislatures and public utilities commissions had the opportunity to adopt RPS policies is worth reiterating. Of course, some state public utilities commissions are more attuned to influence from their respective state legislatures than are other public utilities commissions. In my close reading of state legislative and regulatory RPS documents, however, I do not find systematic evidence of state legislatures detailing how state public utilities commissions should approach the adoption of RPS policy nor do I find evidence about how much RPS policy crafting state legislatures ask public utilities to take up.27 What I observe is that RPS represents an area where multiple institutions had and took advantage of the opportunity to adopt and invent policy, allowing us to better study invention. Nonetheless, in chapter 6 of the book, which contains the quantitative empirical analysis of public-utilities-commission-led RPS policy feature adoption, I account for how much a state’s legislature could influence regulatory agency workings by including variables capturing whether a state’s legislature has term limits (under the assumption that Page 32 →term limits weaken the ability of legislatures to learn about regulatory matters) and the resource endowment or professionalism of a state’s legislature (legislatures with higher endowments may be better able to monitor regulatory developments). In the same chapter, I also include a variable capturing whether or not public utilities commissioners are appointed (and thus subject to legislative confirmation and monitoring).

. While one could speculate why ballot-initiative-led RPS policy feature invention was so rare and also speculate how this pathway of invention may be more common in other areas (recreational marijuana legalization may be one such area worth exploring), I focus in this book on unpacking the determinants of invention in the overwhelmingly most common institutional pathways of invention in the RPS space and save discussion of how ballot-initiative-led invention could be investigated for the book’s conclusion.28 In the book’s appendix, I include a variable corresponding to whether a state allows for laws to be adopted via direct ballot initiative in my analyses; adding this variable does not change empirical results discussed in chapters 5, 6, and 8.

. What is the point behind adopting all these policies? At a very general level and without getting into the classic debate about whether governments should adopt so many policies, the simple answer is that state governments adopt policies to satisfy and implement the demands of their constituents. This is true for both legislatures and bureaucratic agencies. Legislatures adopt policies that they think will best cater to the wishes of the median voter in a legislative election.29 And bureaucratic agencies Page 35 →adopt policies that they think will best allow them to fulfill their executive responsibilities while minimizing gubernatorial and legislative pushback or in the occasional case where agency members are elected, cater to the wishes of the median voter in an agency election.30

. Cutting across the ideological spectrum, journalists also invoke state leadership on policy-making as an antidote to federal inaction or even disagreement with the stated federal position on policy. An author in Vox, a news and analysis website, suggested that policy-making autonomy at the state level could be marshalled to dilute the Trump administration’s immigration agenda (Gerken 2017). Commenting from a decidedly dissimilar ideological perspective in the newsmagazine Reason, a different author claimed that Obamacare stifled state-level autonomy and thus “undermines a core principle of American federalism” (Staley 2010). The ability of the concept of state-level policy-making autonomy to gain adherents from different ideological dispositions speaks to the existence of a widespread faith in the power of the states to solve pressing challenges and is arguably an antidote to the paralysis of federal polarization insofar as the individual state governments are more likely to be ideologically homogenous compared to the federal government.2 This faith is mirrored in differences in reported public sentiment in the competence of federal versus state governments to handle problems. Plotting the percentage of respondents in 10 years of Gallup polls who signal that they have a “great deal” of trust in the ability of the federal government to handle domestic problems alongside the percentage of respondents who signal that they have a “great deal” of trust in the ability of state governments to handle state problems gives a rough measure of how some of the public ranks the relative competence of federal versus state governments.3 In figure 1, I provide this plotted comparison using Gallup poll results collected in the 2010–2020 interval (Gallup 2020).

. What is the point behind adopting all these policies? At a very general level and without getting into the classic debate about whether governments should adopt so many policies, the simple answer is that state governments adopt policies to satisfy and implement the demands of their constituents. This is true for both legislatures and bureaucratic agencies. Legislatures adopt policies that they think will best cater to the wishes of the median voter in a legislative election.29 And bureaucratic agencies Page 35 →adopt policies that they think will best allow them to fulfill their executive responsibilities while minimizing gubernatorial and legislative pushback or in the occasional case where agency members are elected, cater to the wishes of the median voter in an agency election.30

. Policy adoption rates by the states show little sign of slowing down and may actually increase given gridlock and polarization in the federal government (McCann, Shipan, and Volden 2015). In deciding to adopt policy, state governments must grapple with the issue of adopting one of two types of policy: borrowing existing policy solutions from other states or inventing their own particular policy solutions. Invention implies “customization” and gives the adopting state government the benefit of selecting a policy that is uniquely tailored to address that state’s problems (Karch 2007).31 Borrowing, on the other hand, implies copying and gives the adopting state government the benefit of selecting a policy that has already been adopted and tried in other states. While invention is potentially risky for the adopting state government in the sense that the adopted policy has never been tried and may produce unintended negative consequences, borrowing is potentially risky for the adopting state government in the sense that the adopted policy may not be a good fit for the borrowing state and could be a “square peg in a round hole.” For the adopting state government, the strengths and drawbacks of invention and borrowing are essentially mirror images of each other.

. The Kansas tax experiment is an example of invention that carries a great deal of notoriety, but there are more mundane examples of state policy invention. For example, Michigan has long had the highest truck gross weight limit of any state and allows trucks with gross weights of up to 164,000 pounds to traverse its roads (Oosting 2014).32 Michigan could have chosen a more common weight limit but opted for a unique (and uniquely high) one. Michigan’s policy invention may have brought unique benefits to the state: for example, it is possible that domestic auto production has not declined in the state as much as it would have absent the invention since automakers can reduce the number of trips between production facilities due to a higher weight allowance and lower their own costs. It is also possible that the same logic (firms lowering transaction costs by taking advantage of high weight restrictions) might link the higher weight to the consolidation of Michigan’s position as a key conduit in American-Canadian international trade. However, Michigan’s policy invention has also brought about a disadvantage—which may have been avoided had the state adopted the federal weight standard—in the form of a transportation infrastructure that is deteriorating in part due to the stress placed on roads by heavy trucks.33

. The Kansas tax experiment is an example of invention that carries a great deal of notoriety, but there are more mundane examples of state policy invention. For example, Michigan has long had the highest truck gross weight limit of any state and allows trucks with gross weights of up to 164,000 pounds to traverse its roads (Oosting 2014).32 Michigan could have chosen a more common weight limit but opted for a unique (and uniquely high) one. Michigan’s policy invention may have brought unique benefits to the state: for example, it is possible that domestic auto production has not declined in the state as much as it would have absent the invention since automakers can reduce the number of trips between production facilities due to a higher weight allowance and lower their own costs. It is also possible that the same logic (firms lowering transaction costs by taking advantage of high weight restrictions) might link the higher weight to the consolidation of Michigan’s position as a key conduit in American-Canadian international trade. However, Michigan’s policy invention has also brought about a disadvantage—which may have been avoided had the state adopted the federal weight standard—in the form of a transportation infrastructure that is deteriorating in part due to the stress placed on roads by heavy trucks.33

. Page 37 →Just as there are many examples of state governments inventing during policy adoption, there are also ample examples of state governments borrowing during policy adoption. For example, states have been shown to copy other states in adopting lotteries in order to prevent their residents from traveling to other nearby states to purchase lottery tickets (Berry and Berry 1990; Baybeck, Berry, and Siegel 2011; Shipan and Volden 2012). States not only borrow for competitive reasons but also for cooperative ones. America’s ongoing opioid epidemic has affected every state and created a need for coordinated and streamlined policy-making in order to shut down pill mills and disrupt drug distribution networks. The need for interstate coordination in this example has provided the states with a golden opportunity to borrow best practices from one another in an attempt not only to utilize the most effective anti-opioid policy tools but also to work together to mitigate a severe and widespread public policy challenge.34

. A policy regime is simply the set of all policy features adopted by the Page 39 →state government to meet the programmatic goal motivating the state’s adoption of a policy. In the working example of waterway cleanliness, a policy regime is the set of all policy features that a state government has adopted within its policy aimed at ensuring clean waterways. Continuing from the working example, this means that the state’s policy regime for ensuring clean waterways is the combination of all of the policy features—how a state government defines the cleanliness of its waterways; what businesses will be inspected; how often businesses will be inspected; and how businesses will be punished for polluting the waterways—that the state selected in its macro-level policy adoption (or policy regime) geared toward cleaning its waterways. A state government adopts a policy regime when it initially adopts any set of policy features that taken together encompass how the state will accomplish the goal motivating its policy adoption (or in the parlance of diffusion studies, a state government adopts a policy regime when it adopts a policy). However, the state government can change its policy regime over time by adopting new policy features or dropping existing policy features from the combination of policy features it has chosen to operationalize its policy. Two states have identical policy regimes if they have each adopted the exact same set of policy features:35 if not, then the states have different policy regimes.

. I now use my terminology of policy features, policy regimes, and policy domains to show how invention can systematically be distinguished from borrowing, and the method that I display here will be used for a theoretical investigation of invention and borrowing that takes place in chapters 4, 5, 6, 7, and 8. The policy domain that I use to display my method and explore causes of invention and borrowing is that of the renewable portfolio standard (RPS), and it is worth recapping why this is such a good policy domain to study. First, from an environmental perspective, state-level RPSs are the preeminent carbon mitigation tool used in the United States, meaning that if we can understand what caused the states to invent and borrow with respect to RPS, we may be able to also understand what caused the states to invent and borrow with respect to other green energy policy domains (Rabe 2004, 2007; Carley, Nicholson-Crotty, and Miller 2017; Stokes and Warshaw 2017). And second, from a purely political science and public policy-driven perspective, the RPS policy domain is ideal to analyze for two reasons. First, the federal government has essentially left RPS policy-making entirely in the hands of the states. And while this may seem like a problem at first glance, the lack of federal intervention in RPS policy-making is highly desirable for our purposes since it allows us to assess authentic state-driven explanations of invention and borrowing; through the classic use of sticks and carrots, after all, the federal government may distort state incentives to develop original policy (Peterson 1995). This suggests that analyzing state-level invention and borrowing in a policy domain where the federal government intervenes substantially (such as welfare provision or transportation policy) can create false impressions of what influences state-level invention and borrowing absent federal intervention.36 The second desirable reason for studying the RPS policy Page 42 →domain is that different kinds of institutions across the states (namely, legislatures and public utilities commissions) have taken the lead in crafting and adopting RPSs, and we can use the policy adoption experiences of each kind of institution to understand how the drivers of invention and borrowing might differ between legislatures and regulatory agencies.

. Renewable portfolio standards are a set of policy features that state governments use to encourage (and in most cases, mandate) electric utility companies to procure and distribute electricity from renewable sources of energy (Rabe 2007). Electric utility companies typically deliver electricity from producers to end-use consumers,37 and state RPS policy regimes (or the set of policy features that each respective state adopts to operationalize its own RPS program) promote renewable energy use by placing the onus on electric utility companies to supply their consumers with electricity derived from renewable raw materials. RPS policy regimes have emerged as the preferred tool that state policymakers use in trying to spur renewable energy consumption: an RPS policy regime is more viable politically than alternatives that impose direct and observable costs on end-use energy consumers (a carbon tax is an example of an alternative that Page 43 →imposes direct and observable costs on end-use energy consumers), and an RPS policy regime typically imposes direct costs on actors—electric utility companies—with whom the vast majority of end-use energy consumers have an oppositional relationship (Besley and Coate 2003; Holland, Hughes, and Knittel 2009).38

. Renewable portfolio standards are a set of policy features that state governments use to encourage (and in most cases, mandate) electric utility companies to procure and distribute electricity from renewable sources of energy (Rabe 2007). Electric utility companies typically deliver electricity from producers to end-use consumers,37 and state RPS policy regimes (or the set of policy features that each respective state adopts to operationalize its own RPS program) promote renewable energy use by placing the onus on electric utility companies to supply their consumers with electricity derived from renewable raw materials. RPS policy regimes have emerged as the preferred tool that state policymakers use in trying to spur renewable energy consumption: an RPS policy regime is more viable politically than alternatives that impose direct and observable costs on end-use energy consumers (a carbon tax is an example of an alternative that Page 43 →imposes direct and observable costs on end-use energy consumers), and an RPS policy regime typically imposes direct costs on actors—electric utility companies—with whom the vast majority of end-use energy consumers have an oppositional relationship (Besley and Coate 2003; Holland, Hughes, and Knittel 2009).38

. RPS policy regimes are also viable politically because they are not purely associated (in the eyes of the public and some policymakers) with climate change (Rabe 2004, 2007). Policy alternatives that promote renewable energy use, like fossil fuel taxes or cap and trade systems, have directly been labeled as anti-fossil-fuel policies and face opposition from climate change skeptics; RPS policy regimes, on the other hand, have been lauded as an effective way to reduce dependence on foreign petroleum while creating jobs and have even won support from policymakers in several politically conservative states where the domestic fossil fuel industry represents an important constituency (Rabe 2004, 2007).39

. As figure 1 reveals, a higher number of respondents believe that state governments exhibit a great deal of competence compared to the federal Page 4 →government. This gives heft to the notion that much of the public believes in letting states take the initiative in solving problems based on a belief in greater state-level competence. Public belief in state-level policy-making complements a famous statement made by Justice Louis Brandeis in the 1932 case New State Ice Company v. Liebmann. Commenting about one of the virtues of American federalism in his dissenting opinion, Brandeis argued that “it is one of the happy incidents of the federal system that a single courageous State may, if its citizens choose, serve as a laboratory; and try novel social and economic experiments without risk to the rest of the country” (285 U.S. 262). Meant to articulate one of the benefits of allowing for state autonomy in policy-making, Brandeis’s comment has become canonical advice invoked to justify state control over policy-making. From welfare to tax policy to agricultural policy, the thinking goes, letting states take the lead in spearheading policy-making can engender systemwide (e.g., across the 50 U.S. states) benefits in terms of creating unique solutions to problems that may be faced across several states. To use an analogy from gardening, giving the states policy autonomy allows for different kinds of flowers to be grown in 50 different garden plots, which presumably increases the diversity of cultivation compared to if only one Page 5 →garden plot (the federal or national one) were utilized.4 By this logic and given federal governmental inaction with respect to green energy policy, one might expect that state-level leadership in this area could precipitate an explosion of experimentation in the states, with the possibility that some state-developed policies could diffuse and serve as the basis for similar policy-making in other states and even the federal government (Rabe 2004, 2007, 2008; Volden 2006). In short, the states could illustrate paths forward where the federal government has failed to do so.

. An RPS policy regime consists of a set of policy features that dictate how an electric utility company should provide renewable-energy-derived electricity to consumers. Each state RPS policy regime typically includes four types or groups of policy features that specify how an electric utility company should provide renewable energy-derived electricity to consumers: (1) policy features belonging to the first group specify the exact energy sources or technological processes that are considered to be “renewable” (such as electricity derived from “hydroelectric” or “wind” sources) for the purposes of meeting the requirements of a state’s RPS program; and (2) policy features belonging to the second group deal with the final target rate (which can either be mandatory or voluntary) that a state imposes on electric utility companies to be in compliance with its RPS program: for example, many states specify that electric utility companies procure some amount of electricity sold to end-use consumers from renewable sources, and each unique procurement target amount combined with whether that target is mandatory or voluntary represents its own policy feature. (3) Policy features belonging to the third group deal with whether a state requires that part of its RPS be met using a specific energy source or technological process: New Hampshire, for example, requires that electric utility companies operating in that state must meet part of their RPS obligations by procuring electricity from biomass sources.40 And (4) policy features belonging to the fourth group deal with whether a state allows electric utility companies to trade credits to meet RPS requirements. I list all of the policy features utilized in this book in the appendix. Furthermore, the policy features were identified by renewable energy policy experts who assembled the Database of State Incentives for Renewables and Efficiency, or DSIRE. DSIRE is an initiative of the North Carolina Clean Energy Technology Center at North Carolina State University. Experts at DSIRE have compiled and maintain overview pages of each state’s RPS program, and these overview pages list which policy features have been included in each state’s RPS program.

. Looking at state policy-making at the level of policy feature adoption gives the analyst a more accurate picture of the true uniqueness of a state’s RPS policy regime compared to looking at state policy-making at the more macro level of RPS policy regime or program adoption. Imagine that I compare the RPS policy regimes of Michigan and Ohio but ignore adoption at the granular level of policy features and only look at whether Michigan and Ohio adopted RPS policy regimes.41 The RPS policy regimes of Page 46 →the two states would appear identical even though they differ substantially in terms of what sources and processes are considered to be “renewable” and whether electric utility companies must meet carve-outs or technology minimums. Focusing instead on policy feature rather than policy regime adoption would allow us to better discern the diversity of RPS policy-making within and across states.

. The mistaken conclusion mentioned above can occur with respect to the RPS policy domain. If I looked at the initial adoption of RPS policy regimes but did not consider the policy-feature-specific content of those regimes, I may conclude that West Virginia, the third to last state to adopt an RPS policy regime in the time span of this book, is a laggard that has entirely borrowed the policy-making choices of states that established their RPS policy regimes earlier. However, looking at the adoption of policy features reveals West Virginia to be a big inventor that has adopted several policy features, mainly related to the incorporation of coal and fossil fuel sources and technologies within the RPS landscape, which were previously not adopted by any state.42

. A key hallmark of this book is that I define invention and borrowing at the level of policy features and also discuss how to identify invention and borrowing in the adopted policy features associated with state RPS policy regimes. Although I utilize my coding process with respect to the policy domain of RPS, I emphasize that this process is extendable to policy Page 49 →domains besides RPS. The one caveat, however, is that the analyst must have knowledge about what policy features are actually associated with a policy domain.43 As I have mentioned, I use policy features that have been identified by analysts at DSIRE. Utilizing the DSIRE database is valuable insofar as it gives me confidence that I am taking the correct policy features into consideration in my analysis. I am aware, however, that the DSIRE database is not perfect in listing policy features associated with each state’s RPS program. Therefore, I supplement information gathered from the DSIRE database with state RPS program fact sheets provided by the Union of Concerned Scientists. Two different states may use different terminology to refer to the same policy feature in their respective RPS policy regimes, and I consequently devise and utilize a method of identifying and merging synonymous policy feature names (e.g., different names that actually describe the same policy feature) to avoid the situation where counts of invention and borrowing are incorrect because synonymous policy features are mistakenly treated as different and distinct policy features. I utilize the synonyms identification and merging process prior to quantitative analysis and describe this process later in this chapter.

. My coding strategy centers on two key assumptions. First, I assume that invention and borrowing are distinct and separable types of policy adoption available to policymakers. The distinction between invention and borrowing, I argue, is one of relative timing and boils down to whether a specific policy feature that is being adopted by state i has already been adopted by at least one other state j. If state i adopts a specific policy feature that has never been adopted by another state, then I claim that state i’s policymakers are inventing; on the other hand, if state i adopts a specific policy feature that has already been adopted by another state, then I claim that state i’s policymakers are borrowing. Invention and borrowing differ because borrowing provides adopting policymakers with a visible track record while invention does not. Borrowing policymakers in state i in some sense know what to expect from adopting a policy feature that has already been adopted in state j because the policymakers in state i can observe what happened when state j adopted the same policy feature.44 Inventing Page 50 →policymakers in state i know comparatively less about what to expect when adopting a novel policy feature because they do not have access to some other state’s track record with that same policy feature. Following Volden, Ting, and Carpenter (2008) and Cai and Treisman (2009), I assume that the existence of a track record generally makes borrowing less risky than inventing, and I use the idea that borrowing offers adopting policymakers a visible track record while inventing does not to give invention and borrowing different empirical forms.45 Importantly, a track record can be visible even accounting for the granularity of analyzing policy at the level of a policy feature: for instance, policymakers considering whether to borrow would be able to see the impacts of another state designating “wind” as a renewable resource and then could decide whether to follow suit.

. My coding strategy centers on two key assumptions. First, I assume that invention and borrowing are distinct and separable types of policy adoption available to policymakers. The distinction between invention and borrowing, I argue, is one of relative timing and boils down to whether a specific policy feature that is being adopted by state i has already been adopted by at least one other state j. If state i adopts a specific policy feature that has never been adopted by another state, then I claim that state i’s policymakers are inventing; on the other hand, if state i adopts a specific policy feature that has already been adopted by another state, then I claim that state i’s policymakers are borrowing. Invention and borrowing differ because borrowing provides adopting policymakers with a visible track record while invention does not. Borrowing policymakers in state i in some sense know what to expect from adopting a policy feature that has already been adopted in state j because the policymakers in state i can observe what happened when state j adopted the same policy feature.44 Inventing Page 50 →policymakers in state i know comparatively less about what to expect when adopting a novel policy feature because they do not have access to some other state’s track record with that same policy feature. Following Volden, Ting, and Carpenter (2008) and Cai and Treisman (2009), I assume that the existence of a track record generally makes borrowing less risky than inventing, and I use the idea that borrowing offers adopting policymakers a visible track record while inventing does not to give invention and borrowing different empirical forms.45 Importantly, a track record can be visible even accounting for the granularity of analyzing policy at the level of a policy feature: for instance, policymakers considering whether to borrow would be able to see the impacts of another state designating “wind” as a renewable resource and then could decide whether to follow suit.

. Two questions arise: Is invention so easily distinguishable from borrowing? Isn’t the distinction between invention and borrowing more nuanced than I suggest here? I acknowledge these concerns. First, not only do I code each state’s individual policy feature as an example of invention or borrowing but also code each state’s combination of policy features (or each state’s policy regime) as an example of invention or borrowing. I am able to identify the specific set of policy features that make up each state’s RPS policy regime because I already identified the individual policy features that have been adopted by each state. Looking at each state’s policy regime is a reflection of the complexity and nuance of state policy choices. A given state i (Indiana, for example) could adopt a policy regime that consists entirely of individual policy features that have already been adopted in other states, but if Indiana is adopting individual policy features that have never before been combined together, then we would want to acknowledge Indiana’s invention of combining individual policy features in a novel and untested manner.46

. My second main assumption deals with how I empirically distinguish invention from borrowing: namely, I code the adoption of a specific policy feature as an example of invention if state i either (a) adopts that policy feature before any other state has adopted the policy feature or (b) adopts that policy feature by the next calendar year after another state j became the very first state to adopt the policy feature. I build the “by the next calendar year” grace period into the definition of invention to account for the situation where state i essentially adopts a policy feature at roughly the same time as state j (states i and j adopt a feature at different times but are close enough in their adoption that we cannot say definitively that one is the inventor while the other is not). The “by the next calendar year” grace period builds a cushion for invention that effectively pushes borrowing on a given policy feature adoption out until two years after the calendar year in which the first state adopted the same policy feature. For example, if Ohio is the first state to adopt a particular policy feature in 2008, any other state’s adoption of that same policy feature that occurs from 2010 onward would be borrowing. Having this amount of time to distinguish invention from borrowing (e.g., a policy feature adoption counts as borrowing once two years have passed since the calendar year in which that feature was first adopted by any of the U.S. states) is arguably more generous than work by Beck, Gleditsch, and Beardsley (2006), Swank (2006), and Volden, Ting, and Carpenter (2008), who argue that one year is sufficientPage 52 → for one state to observe another state’s policy-making. While the distinction is not perfect (some states could ostensibly be myopic borrowers), it conforms to the fairly straightforward assumption that later adopters could be informed by first movers if they chose to do so. I therefore code the adoption of a specific policy feature as an example of borrowing if state i adopts a policy feature at any time once two years have passed since the calendar year in which state j was the very first state among any state to adopt the policy feature (so, if state j is the first state to adopt a policy feature in the year 2007, state i’s adoption of that same feature counts as borrowing if it occurs in the year 2009 or later).47

. My second example illustrates why the grace period is necessary. In November 1997, Massachusetts invented by becoming the first state to include energy from waves as an eligible source or technological process within its RPS program with the passage of Chapter 164 of the Acts of 1997. Neighboring Connecticut followed suit in April 1998 with the passage of House Bill 5005 and also included wave energy as an eligible source within its own RPS program. Although about five months separated the Page 53 →actions of Massachusetts and Connecticut, it is difficult to say conclusively that Connecticut borrowed from Massachusetts. Connecticut policymakers may have been unable to establish an RPS and include wave energy within it in late 1997 due to having shorter legislative sessions than peers in Massachusetts. Furthermore, the five-month gap in adoption between both states arguably may not have given Connecticut enough time to actually benefit from a visible track record from Massachusetts’s adoption. Categorizing Connecticut’s adoption as an example of invention is an acknowledgment of the idea that a second state that adopts soon after the original state is inventing in its own right.48

. In table 1, I display the number of adoptions that occur by state along with a breakdown of how many of those adoptions represent invention or borrowing.49 The counts by state include adopting individual policy features as well as combining features together within a regime. These choices occur at the same time (when a state is deciding whether to adopt an RPS policy), refer to the same question faced by policymakers (“What shall we include in our policy?”), and are combined together in analyses throughout this book.50

. The idea that the states can go it alone and find novel policy solutions is alluring and oft-mentioned in policy circles.5 Moreover, I am highly confident that states adopt novel policies: a 2014 article from the Washington Post revealed that state legislatures ushered in over 24,000 laws that year (Wilson 2014); some of the components of these 24,000 laws were almost certainly unique to the states adopting them and not borrowed from other states’ policy-making.6 However, even though policymakers, pundits, and the public believe in the capacity of the states to deliver novel solutions to pressing problems, and even though instances exist of where states adopt novel solutions, we know very little about what factors motivate states to adopt novel policy in the first place. Political science scholarship on policy adoption, which is dominated by work on the U.S. states (Mooney 2020), has conceptualized experimentation in state policy adoption through the lens of a state adopting a policy that is new to it regardless of whether that policy is actually new across the system of all states (Mooney 2020; and Walker 1969). This choice of how to conceptualize experimentation in Page 6 →state policy adoption, along with a heavy emphasis being placed on the diffusion or spread of policy from state to state (Walker 1969; Gray 1973; Berry and Berry 1990; Shipan and Volden 2006; Karch 2007; Shipan and Volden 2008; Boushey 2010; Parinandi et al. 2020), means that scholarship has largely skipped the issue of when states are likely to adopt policy that has never been tested in other states.7

. In table 1, I display the number of adoptions that occur by state along with a breakdown of how many of those adoptions represent invention or borrowing.49 The counts by state include adopting individual policy features as well as combining features together within a regime. These choices occur at the same time (when a state is deciding whether to adopt an RPS policy), refer to the same question faced by policymakers (“What shall we include in our policy?”), and are combined together in analyses throughout this book.50

. A glance at table 151 reveals that some states invent more than others. Connecticut invented in 15 of its 22 adoptions (or 68%) and Nevada invented in 16 of its 26 adoptions (or 61%). Other states, such as already-mentionedPage 56 → Indiana (which invented in 1 of its 20 adoptions, or 5%) or Montana (which invented in 2 of its 12 adoptions, or 16%), invented much less than other states and tend to borrow during adoption. There is ample variation across states and (in some cases) within states in terms of where invention and borrowing occurred, and this variation aids us in determining why states invent or borrow. However, before delving into theoretical explanations in subsequent chapters, there are two remaining tasks to be completed in this chapter. First, dates of policy adoption—needed to establish whether a given adoption is an example of invention or borrowing, are not listed in the DSIRE database. Consequently, I needed to gather a vast corpus of official documentation across the states that I could use to “time-stamp” each adoption of each state and thereby relate adoptions of the same policy feature across states to one another in order to code instances of invention and borrowing.52 And second, I needed to take the output from that coding process and create master pooled event history datasets that can be utilized to evaluate the determinants of invention and borrowing. I devote the next two sections to pursuing each of these respective tasks.

. A glance at table 151 reveals that some states invent more than others. Connecticut invented in 15 of its 22 adoptions (or 68%) and Nevada invented in 16 of its 26 adoptions (or 61%). Other states, such as already-mentionedPage 56 → Indiana (which invented in 1 of its 20 adoptions, or 5%) or Montana (which invented in 2 of its 12 adoptions, or 16%), invented much less than other states and tend to borrow during adoption. There is ample variation across states and (in some cases) within states in terms of where invention and borrowing occurred, and this variation aids us in determining why states invent or borrow. However, before delving into theoretical explanations in subsequent chapters, there are two remaining tasks to be completed in this chapter. First, dates of policy adoption—needed to establish whether a given adoption is an example of invention or borrowing, are not listed in the DSIRE database. Consequently, I needed to gather a vast corpus of official documentation across the states that I could use to “time-stamp” each adoption of each state and thereby relate adoptions of the same policy feature across states to one another in order to code instances of invention and borrowing.52 And second, I needed to take the output from that coding process and create master pooled event history datasets that can be utilized to evaluate the determinants of invention and borrowing. I devote the next two sections to pursuing each of these respective tasks.

. The first step is to gather the names of all individual RPS policy features listed in DSIRE’s state-specific RPS pages.53 While the full list of features is available in the appendix (in table A1), I have already described several examples of individual policy features. Making wind energy an allowable source or technological process is an example of an individual policy feature. And so too was the adoption of wave energy as an allowable source or technological process by Massachusetts and Connecticut. Another example of an individual policy feature is California’s 2011 adoption in Senate Bill X1–2 of a requirement that electric providers procure 33% of electricity sold to retail consumers from renewable sources or technological processes.

. Having identified a master list of policy features that appear in the RPS policy domain as well as a list of official documents in which states made changes to their RPS programs, I can determine when states adopted their respective policy features—by searching through each state’s list of official documents to determine when that state first mentions including a policy feature in its RPS program—and then code adoptions across states as instances of invention or borrowing based on the coding rule explained Page 59 →earlier where a state invents if it adopts a specific policy feature before any other state has adopted that policy feature or adopts the policy feature by the next calendar year after another state became the first state to adopt the policy feature (borrowing occurs in all other adoption cases).54

. Page 60 →Incidentally, Maryland’s adoption of the three policy features described in the previous paragraph all count as borrowing; this is the case since Maryland adopted “wind” energy as an eligible RPS source roughly 21 years after Iowa first did the same, adopted “tidal” energy as an eligible RPS source roughly seven years after Maine first did the same, and adopted “wave” energy as an eligible RPS source over six years after Massachusetts first did the same. Maryland’s adoption of the fourth policy feature—mandating that utilities provide a target rate of 20% of electricity sold to consumers from renewable sources—was not adopted with the passage of House Bill 1308 in 2004. Rather, Maryland initially chose a required final target rate specifying that 7.5% of electricity sold to consumers be derived from renewable energy (House Bill 1308, 2004).55 Maryland amended its RPS policy regime with the passage of House Bill 375 in 2008, and one area of amendment was an increase in the state’s final target rate. Specifically, Maryland adopted a new final target rate specifying that 20% of electricity sold to consumers be derived from renewable sources and included language to this effect, stating that “20%” of electricity will eventually need to be derived “from renewable sources” (House Bill 375, 2008). The Maryland legislature passed the final version of House Bill 375 on April 4, 2008; therefore, Maryland adopted the required 20% final target rate on April 4, 2008. Maryland’s adoption of the required 20% target rate based on retail sales was an example of borrowing, since it occurred over five years after California adopted a required 20% final target rate based on retail sales in 2002.

. And what about the process for identifying state policy regimes and coding each of them as instances of invention or borrowing? For each iteration of a state’s RPS policy-making, I look at the combination of individual RPS policy features that the state adopts and characterize the choice to adopt that combination of individual features as representing yet another policy adoption decision that the state makes when it adopts the individual features. I do this for all states and often repeatedly for the same state, since a state adopts a different policy regime when it amends its own RPS program and thereby adopts a different combination of individual policy features compared to what it had before.56 I ultimately am able to identify the distinct combinations of individual policy features that have been adopted within and across the states and can use the exact same coding process described earlier to code these adoptions Page 62 →as inventions or borrowings.57 I can then add these adoptions to the individual policy feature adoptions to evaluate explanations about why invention and borrowing occur.58

. And what about the process for identifying state policy regimes and coding each of them as instances of invention or borrowing? For each iteration of a state’s RPS policy-making, I look at the combination of individual RPS policy features that the state adopts and characterize the choice to adopt that combination of individual features as representing yet another policy adoption decision that the state makes when it adopts the individual features. I do this for all states and often repeatedly for the same state, since a state adopts a different policy regime when it amends its own RPS program and thereby adopts a different combination of individual policy features compared to what it had before.56 I ultimately am able to identify the distinct combinations of individual policy features that have been adopted within and across the states and can use the exact same coding process described earlier to code these adoptions Page 62 →as inventions or borrowings.57 I can then add these adoptions to the individual policy feature adoptions to evaluate explanations about why invention and borrowing occur.58

. And what about the process for identifying state policy regimes and coding each of them as instances of invention or borrowing? For each iteration of a state’s RPS policy-making, I look at the combination of individual RPS policy features that the state adopts and characterize the choice to adopt that combination of individual features as representing yet another policy adoption decision that the state makes when it adopts the individual features. I do this for all states and often repeatedly for the same state, since a state adopts a different policy regime when it amends its own RPS program and thereby adopts a different combination of individual policy features compared to what it had before.56 I ultimately am able to identify the distinct combinations of individual policy features that have been adopted within and across the states and can use the exact same coding process described earlier to code these adoptions Page 62 →as inventions or borrowings.57 I can then add these adoptions to the individual policy feature adoptions to evaluate explanations about why invention and borrowing occur.58

. I stated in the introductory chapter and briefly in this one that the states overwhelmingly adopted RPS programs via legislative action and public utilities commission rulemaking, with a small minority of states adopting RPS programs via ballot initiative. I would like to make explicit here that my coding process is agnostic to the type of governmental actor that adopts a particular policy feature or combination of policy features. A policy feature’s adoption (or the adoption of a given combination of policy features) registers as a case of invention or borrowing regardless of the actor doing the adopting: in short, a legislature in one state can borrow from a regulatory commission in another state and vice versa. My coding process yields cases where legislatures and regulatory commissions emerge as major inventors and also yields cases where legislatures, regulatory commissions, and citizens operating through the ballot predominantly borrow. I make my coding process agnostic to the type of actor adopting policy to reflect that any type of actor can potentially take on the risk of inventing and to acknowledge that any type of actor can observe and borrow from the policy choices made by other governments.59

. The idea that the states can go it alone and find novel policy solutions is alluring and oft-mentioned in policy circles.5 Moreover, I am highly confident that states adopt novel policies: a 2014 article from the Washington Post revealed that state legislatures ushered in over 24,000 laws that year (Wilson 2014); some of the components of these 24,000 laws were almost certainly unique to the states adopting them and not borrowed from other states’ policy-making.6 However, even though policymakers, pundits, and the public believe in the capacity of the states to deliver novel solutions to pressing problems, and even though instances exist of where states adopt novel solutions, we know very little about what factors motivate states to adopt novel policy in the first place. Political science scholarship on policy adoption, which is dominated by work on the U.S. states (Mooney 2020), has conceptualized experimentation in state policy adoption through the lens of a state adopting a policy that is new to it regardless of whether that policy is actually new across the system of all states (Mooney 2020; and Walker 1969). This choice of how to conceptualize experimentation in Page 6 →state policy adoption, along with a heavy emphasis being placed on the diffusion or spread of policy from state to state (Walker 1969; Gray 1973; Berry and Berry 1990; Shipan and Volden 2006; Karch 2007; Shipan and Volden 2008; Boushey 2010; Parinandi et al. 2020), means that scholarship has largely skipped the issue of when states are likely to adopt policy that has never been tested in other states.7

. The use of the coding process that I describe throughout this chapter results in the identification of cases of invention and borrowing. However, before an analyst can evaluate why invention and borrowing occur, the analyst must structure the cases of invention and borrowing in a way that allows for reasonable inference. Typically, quantitative social scientists have embraced event history modeling as the standard method of analyzing policy adoption across the American states (Berry and Berry 1990; Box-Steffensmeier and Jones 1997; Box-Steffensmeier and Jones 2004). Event history modeling comes from epidemiology, and with event history modeling, an analyst uses statistical regression to evaluate what makes some event of interest more or less likely to occur. While epidemiologists may be concerned with understanding the causes of heart attack occurrence or lung cancer onset, adoption and diffusion scholars are concerned with understanding what makes a state more or less likely to adopt a given policy.60

. Pooled event history modeling takes the principles of event history modeling one step further by creating separate risk sets for different events and pooling these risk sets together. Pooled event history analysis has already been used by adoption and diffusion scholars (e.g., Boushey 2016) to evaluate the factors that make the adoption of policies that fall under a common policy domain more likely.61 Here, I create a distinct risk set for each policy feature (and each combination of policy features) that appears in the project and pool these risk sets together. Since invention and borrowing are exclusive concepts (meaning that a state cannot simultaneously Page 64 →possess the opportunity to both invent and borrow the same policy feature), I then separate the pooled risk set into two pooled risk sets: one corresponding to invention and the other corresponding to borrowing. For each policy feature and each combination of policy features, I start the invention risk set for all states in 1983, when Iowa became the first state to adopt a prototypical RPS program.62 A state loses the opportunity to invent a given policy feature or combination of policy features once it has invented that feature or combination (e.g., it is the first state to adopt the feature or combination or has done so by the next calendar year after the first state) or if two years have elapsed since the year in which that feature or combination was first adopted by one of the 50 U.S. states.

. Pooled event history modeling takes the principles of event history modeling one step further by creating separate risk sets for different events and pooling these risk sets together. Pooled event history analysis has already been used by adoption and diffusion scholars (e.g., Boushey 2016) to evaluate the factors that make the adoption of policies that fall under a common policy domain more likely.61 Here, I create a distinct risk set for each policy feature (and each combination of policy features) that appears in the project and pool these risk sets together. Since invention and borrowing are exclusive concepts (meaning that a state cannot simultaneously Page 64 →possess the opportunity to both invent and borrow the same policy feature), I then separate the pooled risk set into two pooled risk sets: one corresponding to invention and the other corresponding to borrowing. For each policy feature and each combination of policy features, I start the invention risk set for all states in 1983, when Iowa became the first state to adopt a prototypical RPS program.62 A state loses the opportunity to invent a given policy feature or combination of policy features once it has invented that feature or combination (e.g., it is the first state to adopt the feature or combination or has done so by the next calendar year after the first state) or if two years have elapsed since the year in which that feature or combination was first adopted by one of the 50 U.S. states.

. Examples help once more to clarify. Iowa invented by being the first state to adopt “wind” energy as an eligible RPS source. Since Iowa did this when launching the first RPS in 1983, all states have the opportunity beginning in 1983 act in a similar manner. Iowa, of course, exits the inventing dataset for this policy feature following 1983 but other states remain in the inventing dataset until 1984, after which they drop out of this dataset (as it turns out, no other state followed suit by the end of 1984). From 1985 onward, all states except for Iowa have the opportunity to borrow by adopting “wind” as an eligible RPS source: Connecticut, for example, borrowed by following suit in 1998 and is removed from being at risk of adopting “wind” beginning in 1999; similarly, Texas borrowed by following suit in 1999 and is removed from being at risk of adoptingPage 65 → “wind” beginning in 2000. States that never establish an RPS policy regime or program (such as Idaho or Tennessee) remain at risk of borrowing by adopting “wind” as an eligible source through the end of the study in 2011.63

. However, the phrases “distributed generation” and “customer-sited” refer to the same concept: using local resources and facilities to generate electricity instead of transmitting it over long distances. This means that a “distributed generation” carve-out and a “customer-sited” carve-out refer to the same policy feature. The fact that these two differently named features are the same is important to acknowledge since we would be overcounting cases of invention and undercounting cases of borrowing if we ignored the synonymous nature of New York’s and Arizona’s decisions.64

. Prior to coding adoptions as cases of invention or borrowing and prior to constructing the master event history datasets, I therefore devised and utilized a process for identifying when the names of policy features are synonyms of one another.65 While details of this process are covered in the appendix (in list A3), I consulted a corpus of renewable energy summary fact sheets created by the United States Department of Energy, the National Renewable Energy Laboratory, the United States Environmental Protection Agency, the United States Energy Information Administration, and various industry associations to distinguish features from one another and determine when features are synonyms. My rule for identifying whether features are synonyms is to read descriptions from the fact sheets about features and then ask myself whether the descriptions for multiple features are describing the same process. If the answer to this question is “yes,” then the features describing the same process are combined or consolidated into one feature. The incorporation of this combined or consolidated feature into the invention and borrowing datasets follows all of the Page 67 →same rules used for features that do not have synonyms. Importantly, for combined or consolidated features, I use the date of the earlier synonym’s adoption to distinguish invention from borrowing.

. The coding process described here and the master datasets emanating from that process can be used to fill a void in our understanding of policy adoption and federalism and determine when and why invention and borrowing occurs among the American states. One particular opportunity that is opened up through this process is that we can evaluate how state legislatures—the chief institutions charged with crafting policy in the states—handle the task of inventing and borrowing policy; in chapter 5, I document how legislatures approach invention and borrowing in RPS policy-making. In order to pursue this endeavor, I extract adoptions from the master datasets that occurred through legislative action and also Page 68 →extract all nonadoptions from the master datasets to reflect the idea that the opportunity for a legislature in state i to borrow a policy feature is influenced by whether a public utilities commission in state j invented with respect to that policy feature.66 I additionally extract all nonadoptions to reflect the idea that a legislature in a given state loses the opportunity to adopt a policy feature or combination of policy features if another institution in that state has already adopted the same policy feature or combination of policy features.

. At their core, legislatures and regulatory agencies have different fundamental responsibilities, and these different responsibilities inform where we might expect RPS invention to predominantly occur. In the American political system, the legislative branch is chiefly tasked with making law. This central role is not only referenced in the United States Constitution (United States Constitution, Article I) but also extends to state legislatures, which share the same primary responsibility of proposing, crafting, and adopting legislation as their federal peer (Squire 2007, 2017). In short, legislatures function as policy adoption specialists, and this role, in turn, shapes the behavior of those who serve in legislatures. Legislators are likely to try to get preferred candidate policies adopted into law, are likely to look at policy adoption as one of preeminent ways through which society’s challenges can be solved, and are likely to view their own effectiveness through the lens of whether they were able to codify their preferred policy candidates into law (Volden and Wiseman 2014; Bucchianieri, Volden, and Wiseman 2020). Legislators therefore plausibly consider legislating to be an integral if not the integral part of their job.67

. Legislators’ own view about the nature of their job appears to have legitimacy across the broader political system. One aforementioned sign Page 71 →of legitimacy is the United States Constitution and, by extension, state constitutions, which express that legislatures retain the primary authority to adopt law. Another sign of legitimacy is the observation that the judiciary possesses the power of judicial review but generally has not claimed primacy over lawmaking itself. While the judiciary has exercised review since the early nineteenth century (Marbury v. Madison, 1803), it arguably has not parlayed the power of review into a demand for control over the job of legislating itself; the absence of such a demand suggests that broader legitimacy exists around the assertion that a prime responsibility of legislators is to make policy.68 Even the public, for all the reputed and sometimes humorous animosity directed toward the legislative process in surveys, seems to believe that the job of legislators is to make policy.69 The finding that public disapproval of legislators increases when legislative bodies suffer from gridlock (Flynn and Harbridge 2016) offers additional credence to the idea that the public has some expectation that legislators should try to adopt policy.

. Legislators’ own view about the nature of their job appears to have legitimacy across the broader political system. One aforementioned sign Page 71 →of legitimacy is the United States Constitution and, by extension, state constitutions, which express that legislatures retain the primary authority to adopt law. Another sign of legitimacy is the observation that the judiciary possesses the power of judicial review but generally has not claimed primacy over lawmaking itself. While the judiciary has exercised review since the early nineteenth century (Marbury v. Madison, 1803), it arguably has not parlayed the power of review into a demand for control over the job of legislating itself; the absence of such a demand suggests that broader legitimacy exists around the assertion that a prime responsibility of legislators is to make policy.68 Even the public, for all the reputed and sometimes humorous animosity directed toward the legislative process in surveys, seems to believe that the job of legislators is to make policy.69 The finding that public disapproval of legislators increases when legislative bodies suffer from gridlock (Flynn and Harbridge 2016) offers additional credence to the idea that the public has some expectation that legislators should try to adopt policy.

. The idea that the states can go it alone and find novel policy solutions is alluring and oft-mentioned in policy circles.5 Moreover, I am highly confident that states adopt novel policies: a 2014 article from the Washington Post revealed that state legislatures ushered in over 24,000 laws that year (Wilson 2014); some of the components of these 24,000 laws were almost certainly unique to the states adopting them and not borrowed from other states’ policy-making.6 However, even though policymakers, pundits, and the public believe in the capacity of the states to deliver novel solutions to pressing problems, and even though instances exist of where states adopt novel solutions, we know very little about what factors motivate states to adopt novel policy in the first place. Political science scholarship on policy adoption, which is dominated by work on the U.S. states (Mooney 2020), has conceptualized experimentation in state policy adoption through the lens of a state adopting a policy that is new to it regardless of whether that policy is actually new across the system of all states (Mooney 2020; and Walker 1969). This choice of how to conceptualize experimentation in Page 6 →state policy adoption, along with a heavy emphasis being placed on the diffusion or spread of policy from state to state (Walker 1969; Gray 1973; Berry and Berry 1990; Shipan and Volden 2006; Karch 2007; Shipan and Volden 2008; Boushey 2010; Parinandi et al. 2020), means that scholarship has largely skipped the issue of when states are likely to adopt policy that has never been tested in other states.7

. While a central part (and many would contend, the central part) of legislative job responsibility is to adopt policy, the situation is arguably different for regulatory agencies. Regulatory agencies exist as executive bodies and are primarily tasked with administering and enforcing policy. Administration refers to running or managing policy while enforcement refers to making sure that the various parties subject to contract stipulations within the policy are following those stipulations. Policy, according to the view that regulatory agencies should be pure executives, is something that is idealized as coming from outside the regulatory organization (canonically originating in the legislative branch) and that is then put into practice by impartial regulatory arbiters. Of course, this view neglects the reality that regulatory agencies do adopt policy. Regulatory agencies might adopt policies to govern their own internal affairs or human resource matters;70 Page 72 →they might adopt policies to make their executive responsibilities easier to accomplish;71 and they might make policies at the behest of the legislative branch.72 However, the view of regulatory agencies as being primarily executive in character creates an ideal type for how regulatory agencies should behave regarding policy-making and arguably limits their room to maneuver on policy adoption.73

. While a central part (and many would contend, the central part) of legislative job responsibility is to adopt policy, the situation is arguably different for regulatory agencies. Regulatory agencies exist as executive bodies and are primarily tasked with administering and enforcing policy. Administration refers to running or managing policy while enforcement refers to making sure that the various parties subject to contract stipulations within the policy are following those stipulations. Policy, according to the view that regulatory agencies should be pure executives, is something that is idealized as coming from outside the regulatory organization (canonically originating in the legislative branch) and that is then put into practice by impartial regulatory arbiters. Of course, this view neglects the reality that regulatory agencies do adopt policy. Regulatory agencies might adopt policies to govern their own internal affairs or human resource matters;70 Page 72 →they might adopt policies to make their executive responsibilities easier to accomplish;71 and they might make policies at the behest of the legislative branch.72 However, the view of regulatory agencies as being primarily executive in character creates an ideal type for how regulatory agencies should behave regarding policy-making and arguably limits their room to maneuver on policy adoption.73

. While a central part (and many would contend, the central part) of legislative job responsibility is to adopt policy, the situation is arguably different for regulatory agencies. Regulatory agencies exist as executive bodies and are primarily tasked with administering and enforcing policy. Administration refers to running or managing policy while enforcement refers to making sure that the various parties subject to contract stipulations within the policy are following those stipulations. Policy, according to the view that regulatory agencies should be pure executives, is something that is idealized as coming from outside the regulatory organization (canonically originating in the legislative branch) and that is then put into practice by impartial regulatory arbiters. Of course, this view neglects the reality that regulatory agencies do adopt policy. Regulatory agencies might adopt policies to govern their own internal affairs or human resource matters;70 Page 72 →they might adopt policies to make their executive responsibilities easier to accomplish;71 and they might make policies at the behest of the legislative branch.72 However, the view of regulatory agencies as being primarily executive in character creates an ideal type for how regulatory agencies should behave regarding policy-making and arguably limits their room to maneuver on policy adoption.73

. While a central part (and many would contend, the central part) of legislative job responsibility is to adopt policy, the situation is arguably different for regulatory agencies. Regulatory agencies exist as executive bodies and are primarily tasked with administering and enforcing policy. Administration refers to running or managing policy while enforcement refers to making sure that the various parties subject to contract stipulations within the policy are following those stipulations. Policy, according to the view that regulatory agencies should be pure executives, is something that is idealized as coming from outside the regulatory organization (canonically originating in the legislative branch) and that is then put into practice by impartial regulatory arbiters. Of course, this view neglects the reality that regulatory agencies do adopt policy. Regulatory agencies might adopt policies to govern their own internal affairs or human resource matters;70 Page 72 →they might adopt policies to make their executive responsibilities easier to accomplish;71 and they might make policies at the behest of the legislative branch.72 However, the view of regulatory agencies as being primarily executive in character creates an ideal type for how regulatory agencies should behave regarding policy-making and arguably limits their room to maneuver on policy adoption.73

. Combined with the view that it is primarily the job of the legislative branch to make policy, the ideal-type view of regulatory agencies as pure executives makes justifying policy adoption comparatively more difficult for the regulatory agencies vis-à-vis the legislative branch. While legislatures can simply invoke their authority to make laws on behalf of the public when adopting policy, regulatory agencies have to be ready to argue that their policy adoptions lie within their jurisdictional zones of authority. Even when this issue might seem clear cut—suppose hypothetically that a state public utilities commission utilizes its authority to maintain the safety and reliability of a state’s electricity grid by raising electricity prices to retrofit electricity infrastructure—regulatory agencies still may face pushback about overstepping their bounds. For instance, in the hypothetical example, electric utility companies, interest groups affiliated with those electric utility companies, angry legislators, and even ordinary consumersPage 73 → can take umbrage with the public utilities commission’s behavior and claim that the regulatory agency went beyond its purview. This same assortment of disaffected actors need not stop there and could use the claim of the regulatory agency acting outside of its purview to make the job of the agency much more onerous: these actors could utilize legislative and judicial hearings, for example, to consume the time and resources of regulators and sow doubt about the ability of regulators to act as impartial executors.74 The point here is to emphasize that regulatory agencies must exhibit caution when adopting policy in order to try to minimize accusations of overstepping their executive role.

. Trying to minimize accusations of overstepping bounds plausibly matters in explaining why state public utilities commissions have invented RPS policy less than state legislatures. Invention is by definition novel in the sense that it has not yet been tried across the states. Given that regulatory agencies seek to avoid appearing to overstep their bounds, they may shy away from adopting policy that as a result of being novel may elicit such an appearance.75 In contrast, as actors of the chief policy-making branch, legislatures are expected to seek out policy solutions and arguably (relative to their regulatory agency peers) need not shy away from inventing due to the fear that they will be accused of overstepping their prescribed governmental role, as legislatures are widely recognized to possess the authority to adopt policy and as invention falls under the archetype of adopting policy. This is not to say, of course, that legislatures are frequent inventors of policy: if the RPS policy area explored prominently in this book is any guide, legislatures invented 169 times but borrowed 357 times, implying that invention represented roughly 32% of legislative policy adoption. Invention does not have a different definition for legislatures (e.g., it is still Page 74 →characterized by its novelty), and the fact that candidates for invention have not been vetted across other states may partly account for why invention occurs less frequently than borrowing, even for legislatures. However, even accounting for the fact that invention is not especially common, the greater legitimacy given to the role of legislatures as policy adopters arguably creates increased permissiveness for legislatures to invent vis-à-vis regulatory agencies, which might help explain the higher incidence of RPS invention among legislatures compared to public utilities commissions.

. The same difference in view about the primary job responsibility of legislatures versus regulatory agencies—where legislatures adopt policy while regulatory agencies execute it—also potentially explains why legislatures have been more frequent borrowers of RPS policy than have public utilities commissions, as borrowing also represents policy adoption. At the same time, it is possible that borrowing may be less likely to precipitate accusations of regulatory agencies overstepping their bounds than inventing based on the idea that many in the public may be more accepting of regulatory agencies adopting tested and vetted policy compared to adopting novel and unprecedented policy, leading regulatory agencies to borrow more than they invent.76 In short, at least in the RPS policy area, the belief that it should mainly be the purview of legislatures to adopt policy along with the belief that it should mainly be the purview of regulators to execute policy may shed light on why legislatures have been more prolific policy adopters than regulatory agencies.77

. The same difference in view about the primary job responsibility of legislatures versus regulatory agencies—where legislatures adopt policy while regulatory agencies execute it—also potentially explains why legislatures have been more frequent borrowers of RPS policy than have public utilities commissions, as borrowing also represents policy adoption. At the same time, it is possible that borrowing may be less likely to precipitate accusations of regulatory agencies overstepping their bounds than inventing based on the idea that many in the public may be more accepting of regulatory agencies adopting tested and vetted policy compared to adopting novel and unprecedented policy, leading regulatory agencies to borrow more than they invent.76 In short, at least in the RPS policy area, the belief that it should mainly be the purview of legislatures to adopt policy along with the belief that it should mainly be the purview of regulators to execute policy may shed light on why legislatures have been more prolific policy adopters than regulatory agencies.77

. Having broad legitimacy to make policy, combined with being exposed to and espousing policy ideas comporting with one’s ideological affiliation, suggests that a legislator is likely to advocate for adopting policy that comports with their ideological affiliation. This not only makes sense at the micro (legislator-specific) but also at the macro (legislature-specific) level, suggesting that as a legislature takes on a particular ideology, that legislature should be more likely to advocate for adopting policy comporting with its particular ideology. Even though many states adopting RPS programs have had conservative bona fides, the fact that RPS sits within the universe of energy and environmental regulation suggests that legislatures should be more likely to advocate for the adoption of RPS policy as they become more liberal in character. On its face, this might not seem all that surprising, but distinguishing invention from borrowing allows us to discern whether the ideological comportment of a legislature has a differential relationship with respect to inventing and borrowing. In the next chapter, I argue and show using statistical analysis that increases in government liberalism more meaningfully account for RPS invention compared to RPS borrowing.78 I offer that this discrepancy may potentially be attributable to the idea that greater ideological extremism (here, in the liberal direction) makes legislatures more apt to ignore or overlook a lack of evidence when adopting policies that align with their ideological agenda, which would explain the link between government liberalism and RPS invention given that RPS is within the liberal paragon of policies. While increased government liberalism admittedly should also increase the probability that legislatures borrow RPS policy, the quality of already Page 77 →having been tested brings a wider ideological spectrum of legislatures on board with borrowing, thereby leading to a weakened association between government liberalism and borrowing. In the eighth chapter of the book, I show that this same pattern occurs with respect to legislatures’ level of conservatism when considering the conservative policy area of restrictions toward obtaining abortion.

. When deliberating about why the ideological orientation of a legislature has such influence on the adoption choices made by that legislature—increased liberalism in a legislature has greater statistical association with RPS invention than with RPS borrowing even though both forms of adoption are positively correlated with rising liberalism, and increased conservatism in a legislature has greater statistical association with anti-abortion invention than anti-abortion borrowing even though both forms of adoption are positively correlated with rising conservatism—it seems important to reiterate that “taking a view” in terms of staking one’s ideological ground is part and parcel of the task of legislating. Insofar as ideology informs the lens through which legislators perceive the world and identify policies to work toward their worldview, it is perhaps not surprising that ideological extremism should lead to an increased willingness to disregard evidence in policy adoption.79

. Ignoring the influences driving a state to go it alone and inaugurate its own novel policy solutions is problematic, especially since we invoke the states as a potential cure for federal dysfunction. If we view something as a cure but do not understand the motivations behind that cure, then we fail to properly respect the medicine we are prescribing and also fail to anticipate possible side effects that could emanate from that medicine. At a more granular level, investigating novel policy adoption entails doing justice to the part of Brandeis’s comment where he speaks about the uniqueness of a given state’s policy. If a key virtue within American federalism is that the individual states, facing federal inaction, have the opportunity to enact novel solutions to problems, and if a key expectation among proponents of American federalism is that some novel state-devised solutions will become common best practices utilized across the federation, then exploring the antecedents of novel state-level policy-making is the sine qua non to unpacking what many observers believe to be a major selling point of American federalism. What does state-level novel policy adoption, which I refer to throughout this book and elsewhere (Parinandi 2020) as invention, look like, and how can it be distinguished from state adoption of a policy that has already been adopted in another or other state(s), an action that I refer to as borrowing? Are the factors motivating state-level invention and borrowing different, and, if so, how are they different? And can we identify explanations that primarily or even uniquely account for when Page 7 →states inject novel policy into the shared cross-state library of attempted policy-making?8

. Another hallmark of the job of legislating is that those serving in legislatures are subject to electoral accountability and that legislative impressions of electoral danger may influence RPS invention and borrowing. Brandeis’s comment in New State Ice Company versus Liebmann (285 U.S. 262) about a “single courageous state” attempting “novel social and economic experiments without risk to the rest of the country” implies that the novel or untested nature of inventing carries some risk for any state government undertaking inventing.80 At the same time, Brandeis’s comment Page 78 →also implies that this same risk is less pronounced for borrowing based on the logic that state governments can observe the experimentation of other states with respect to borrowing.81

. Another hallmark of the job of legislating is that those serving in legislatures are subject to electoral accountability and that legislative impressions of electoral danger may influence RPS invention and borrowing. Brandeis’s comment in New State Ice Company versus Liebmann (285 U.S. 262) about a “single courageous state” attempting “novel social and economic experiments without risk to the rest of the country” implies that the novel or untested nature of inventing carries some risk for any state government undertaking inventing.80 At the same time, Brandeis’s comment Page 78 →also implies that this same risk is less pronounced for borrowing based on the logic that state governments can observe the experimentation of other states with respect to borrowing.81

. Since RPS is the central policy area studied in this book, I devote the space above to discussing linkages between legislative electoral vulnerability and RPS invention and borrowing. Recall though that in the eighth chapter of the book, I evaluate legislative inventing and borrowing with respect to the conservative area of anti-abortion policy. This evaluation includes interrogating how legislative electoral vulnerability affects anti-abortion inventing and borrowing, and I ultimately find that legislative electoral vulnerability in this area has a statistically meaningful relationship with invention (where greater vulnerability corresponds to a higher likelihood of invention) but not borrowing. In terms of explaining my rationale for this finding, I draw a distinction between policy areas that tend to be viewed within a strictly moral “right-versus-wrong” prism (such as abortion, as is documented in Mooney 1999 and Kreitzer and Boehmke 2016) and policy areas that may be less likely to be viewed as purely good or evil propositions (RPS may fit the bill here, as RPS policies appear to have a greater diversity of purpose than do anti-abortion policies, which all seek to restrict abortion access).82 Evidence may matter less in moral policy areas considering that voters might interpret such areas through intrinsic references to right versus wrong. At the same time, voters caring about such issues may look favorably upon invention, as adopting unprecedented policy (here, unprecedented anti-abortion policy) may be a signal to these voters about legislators’ commitment to their cause. As legislative electoral vulnerability increases, legislators may feel emboldened to invent moral policy to showcase their commitment to voters caring about moral policy. In contrast, where evidence is in play (e.g., policy areas that relative to abortion are less likely to be seen purely through the prism of right-versus-wrong, such as RPS), legislative electoral vulnerability drives an emphasis toward tested or vetted policy-making.

. Of course, leaving aside the two factors discussed here—lawmaker ideology and electoral vulnerability—it is possible that other factors could affect legislative RPS invention and borrowing such as the resource capacity that a legislature possesses or the per capita income of a given state. I control for these and other possibilities in the fuller legislative RPS analysis displayed in the next chapter.83 Having remarked about how important within legislature variation could impact RPS invention and borrowing, I next turn to exploring how the core job responsibility of regulatory agencies might shed light on when public utilities commissions invent RPS policy.

. In my discussion of regulatory agencies earlier in this chapter, I not only described a classic view positing that regulatory agencies should serve mainly as executors of policy, but I also parlayed this into the possibility that regulatory agencies might shy away from inventing due to the fear of being accused of veering from their mainly executive nature by not just merely adopting policy but adopting policy that is novel and has not been tested elsewhere. Going one step further, it is possible that the same logic might help us illuminate when public utilities commissions (an example of a regulatory agency) are more likely to invent with respect to RPS policy. One thing that stands out from the idealized view that regulators should Page 81 →strive to behave as fair policy arbiters is that regulators may generally exhibit conflict avoidance. Regulators may seek to limit facing scrutiny from other actors (e.g., regulated entities, legislators, and the court system) since such scrutiny can cast doubt on the ability of regulators to function as fair policy arbiters.84

. RPS policy generally imposes costs on electric utility companies—many of which are entrenched in the sense that they possess near-monopolistic control of electricity provision under sectoral regulation or had near-monopolistic control prior to deregulation—by asking them to change how they procure electricity, and RPS invention does so in a way where electric utility companies cannot turn to the experiences of companies in other states to get a sense of how the policy might impact them.85 Knowing that electric utility companies will generally dislike uncertainty in an area (RPS) where the electric utility companies are potentially being asked to substantially change their electricity procurement practices, public utilities Page 82 →commissions may be more likely to invent RPS policy when they think that electric utility companies will be less likely to challenge such policy-making, and I argue that this situation is more common when a state has a deregulated electricity sector.

. Prior to state electricity deregulation, entrenched electric utility companies had control over the procurement and distribution of electricity in their respective service areas. With electricity deregulation (Ka and Teske 2002; Borenstein and Bushnell 2015), these same entrenched electric utility companies lost monopolistic control over the procurement side of electricity—consumers could choose to purchase electricity from alternate electricity vendors—but retained control over the distribution or transmission of electricity over their grid infrastructure. Deregulation signifies that some control of electricity provision has been taken away from entrenched electric utility companies, and it may have the effect of influencing entrenched electric utility companies to be more pliant toward regulators out of a desire to preserve the control over electricity provision that they still possess. Public utilities commissions, for their part, are conceivably aware that deregulation imposes a demonstration effect on entrenched electric utility companies that reduces these companies’ willingness to challenge regulatory policy-making.86 In turn, this suggests that if public utilities commissions were deliberating about inventing with respect to RPS, they would be more likely to actually invent RPS policy when electric utility companies are less apt to challenge them, which is more likely under deregulation.

. I find some evidence in chapter 6 that RPS invention is more likely given state electricity sector deregulation but fail to find a statistically meaningful relationship between deregulation and RPS borrowing. This does not mean that electric utility companies somehow love RPS borrowing or that public utilities commissions think that electric utility companies will never challenge RPS borrowing. Indeed, I find a positive (though not statistically significant) association between deregulation and RPS borrowing, which gives a weak measure of support for the idea that deregulation increases the probability of RPS borrowing. The nonsignificant relationship, however, provides less cover for the notion that public utilities commissions rely Page 83 →on the weaker position of electric utility companies afforded by deregulation to increase RPS borrowing, and a reason for this might be that firms accommodate borrowing more than they do invention since borrowing arguably creates a history that firms can access to study potential policy impacts.87

. Ultimately, the finding linking electricity sector deregulation to increased regulatory RPS invention is promising in the sense that it gives us a clue as to where novel RPS policy-making by regulatory actors might occur but also limiting considering that a minority of states have deregulated electricity sectors.88 The same finding begs the question of whether a relationship exists between giving regulatory actors greater voice to speak to the public and discerning an increased chance of novel RPS policy adoption. While a majority of states deem that public utilities commission members are chosen by elite-level actors (usually gubernatorial appointment subject to legislative confirmation), a handful of states specify that public utilities commission members be chosen through statewide vote (Besley and Coate 2003; Parinandi and Hitt 2018). The public utilities commissions across both cases generally have similar responsibilities (for example, regulating commerce related to electricity, gas, and water usage within their respective states) and have similar stated expectations about adhering to impartial regulatory roles.89 However, in states where public Page 84 →utilities commissioners are selected through statewide vote, the ability of public utilities commissioners to frame regulatory issues as popular matters may provide the commissioners with some leeway to deviate from wanting to accommodate electric utility companies. That is to say, public utilities commissioners in states with direct election may be able to take their case directly to voters about why RPS policy-making is worthwhile and, in doing so, they may be able to create a cushion to downweight some of the worry about electric utility company pushback. This suggests that public utilities commissioners in states with direct election may be more likely to make RPS policy compared to appointed public utilities commissioners, who are arguably less able to take cases directly to the public.

. Ultimately, the finding linking electricity sector deregulation to increased regulatory RPS invention is promising in the sense that it gives us a clue as to where novel RPS policy-making by regulatory actors might occur but also limiting considering that a minority of states have deregulated electricity sectors.88 The same finding begs the question of whether a relationship exists between giving regulatory actors greater voice to speak to the public and discerning an increased chance of novel RPS policy adoption. While a majority of states deem that public utilities commission members are chosen by elite-level actors (usually gubernatorial appointment subject to legislative confirmation), a handful of states specify that public utilities commission members be chosen through statewide vote (Besley and Coate 2003; Parinandi and Hitt 2018). The public utilities commissions across both cases generally have similar responsibilities (for example, regulating commerce related to electricity, gas, and water usage within their respective states) and have similar stated expectations about adhering to impartial regulatory roles.89 However, in states where public Page 84 →utilities commissioners are selected through statewide vote, the ability of public utilities commissioners to frame regulatory issues as popular matters may provide the commissioners with some leeway to deviate from wanting to accommodate electric utility companies. That is to say, public utilities commissioners in states with direct election may be able to take their case directly to voters about why RPS policy-making is worthwhile and, in doing so, they may be able to create a cushion to downweight some of the worry about electric utility company pushback. This suggests that public utilities commissioners in states with direct election may be more likely to make RPS policy compared to appointed public utilities commissioners, who are arguably less able to take cases directly to the public.

. Working toward and deciphering answers to the questions above matters for any area where the states try to go it alone and devise their own policy solutions to problems, but it arguably matters urgently with respect to green or renewable energy policy. Due to challenges brought about by a warming climate, the United States faces a great imperative to build its renewable energy infrastructure (Allison and Parinandi 2020). Much of that imperative pertains to the country’s retail electricity sector, which has historically been a key source of fossil fuel emissions and has been primarily subject to state-level regulation.9 The individual states possess the authority to determine how their own retail electricity markets function, and the federal government has not been able to tell the individual states how they should accommodate renewable energy development within their respective retail electricity markets—the inability of the federal Obama-era Clean Power Plan to materialize is one example of the federal government’s failure to force state governments to commit to pursuing renewable energy development in designing their own retail electricity markets (United States Environmental Protection Agency 2015). Therefore, seeing how individual states choose to incorporate attempts to foster renewable energy development in their retail electricity markets and specifically seeing how individual states adopt novel policy to foster such renewable energy development provides a window into how the states are spearheading policy in such an important area absent strong federal direction. What are the characteristics of the states that take the lead and adopt novel policy pertaining to renewable energy in their retail electricity markets, and to which states should we look in hopes of finding invention today that may become tomorrow’s best practices? Do the factors influencing invention differ based on what actor within state government is doing the inventing? Page 8 →And, lastly, how can the invention experiences of states in this one important area potentially inform state attempts in other areas where federal government inaction has arguably resulted in individual states claiming much of the mantle of policy-making?

. Based on the idea that electric utility companies may generally be more amenable to borrowing than invention due to the tested nature of borrowing, it is possible that elected public utilities commissioners might find taking their case directly to the public to be especially valuable with respect to justifying novel RPS adoption. This would, if true, establish a key institutional pathway for regulatory RPS invention. As shown see in chapter 6, the benefit of direct election in terms of inducing regulatory RPS adoption is not exclusive to invention, as elected public utilities commissioners are more likely to invent as well as borrow RPS policy compared to their non-elected peers. Indeed, the likelihood of borrowing RPS policy given direct election is greater than the likelihood of inventing RPS policy, and this perhaps makes intuitive sense in light of the observations that borrowing may be more accepted by electric utility companies and that borrowing allows regulators to claim that they are pursuing results produced elsewhere.90 As much of the mandate for this book revolves around conceptually separating invention and borrowing and then determining whether various factors differentially influence invention and borrowing, I give less centrality to the direct election regulatory finding even though direct election increases RPS invention and RPS policy adoption more generally.

. More generally, the finding with respect to government ideology draws attention to the difficulty of identifying and measuring the belief systems of state public utilities commissions. Many existing measures of government ideology are heavily modeled on legislative behavior.91 And even if some public utilities commissioners take on partisan affiliations, many—such as those in Alaska, California, and Kentucky—are avowedly nonpartisan, meaning that using regulatory party affiliation as a measure of regulatory belief would create serious gaps in the data to be evaluated.92 Moreover, even when party affiliation is present and even if regulators have previous histories in other branches of government, the job of the regulatorPage 87 → itself can potentially make those affiliations and previous experiences less informative. In the book’s conclusion, I discuss how we might reliably measure public utilities commission ideology and use it within studies of energy policy moving forward.93

. More generally, the finding with respect to government ideology draws attention to the difficulty of identifying and measuring the belief systems of state public utilities commissions. Many existing measures of government ideology are heavily modeled on legislative behavior.91 And even if some public utilities commissioners take on partisan affiliations, many—such as those in Alaska, California, and Kentucky—are avowedly nonpartisan, meaning that using regulatory party affiliation as a measure of regulatory belief would create serious gaps in the data to be evaluated.92 Moreover, even when party affiliation is present and even if regulators have previous histories in other branches of government, the job of the regulatorPage 87 → itself can potentially make those affiliations and previous experiences less informative. In the book’s conclusion, I discuss how we might reliably measure public utilities commission ideology and use it within studies of energy policy moving forward.93

. More generally, the finding with respect to government ideology draws attention to the difficulty of identifying and measuring the belief systems of state public utilities commissions. Many existing measures of government ideology are heavily modeled on legislative behavior.91 And even if some public utilities commissioners take on partisan affiliations, many—such as those in Alaska, California, and Kentucky—are avowedly nonpartisan, meaning that using regulatory party affiliation as a measure of regulatory belief would create serious gaps in the data to be evaluated.92 Moreover, even when party affiliation is present and even if regulators have previous histories in other branches of government, the job of the regulatorPage 87 → itself can potentially make those affiliations and previous experiences less informative. In the book’s conclusion, I discuss how we might reliably measure public utilities commission ideology and use it within studies of energy policy moving forward.93

. In the American political system, the legislative branch is the main governmental institution tasked with crafting and adopting laws. The Founders intended for the legislative branch to be the most important branch, and with rare exceptions, federal politics in the late eighteenth and nineteenth centuries revolved around powerful legislators articulating their policy positions in the halls of the national Congress (Elkins and McKitrick 1995; Cox and McCubbins 2005; Tulis 2017). Even as the nineteenth century gave way to a twentieth (and so far, a twenty-first century) typified by the emergence of a powerful administrative state (Wilson 1887; Lewis 2012; Potter 2019), the legislative branch formally retains the role of writing law and tasks the executive branch with enforcement.94

. The role ascribed to the legislature in the above paragraph rings true for state legislatures. The American Congress was modeled on the workings of the 13 colonial legislatures, and the state legislatures in turn have been influenced by the workings of the American Congress (Squire 2017).95 Inasmuch as the national legislature has served as the principal institution in charge of crafting laws at the national level, so too have state legislatures served as the principal institutions in charge of crafting laws at the state level. And as the states have taken on greater policy responsibility in recent decades (Peterson Page 90 →1995), state legislatures have arguably needed to exercise their lawmaking responsibility more often and on matters of increasing complexity.

. The North Dakota legislature invented by banning abortion related to genetic abnormality, and North Dakota’s invention has since been borrowed (though unsuccessfully) by states such as Indiana and Ohio.96 North Dakota’s invention arguably represented a proverbial crossing of the Rubicon in the abortion policy domain, not only because it expanded the scope of conflict as predicted by classic political science theory (Schattschneider 1975), but also because it made a policy feature adoption that may have been considered (at least in some circles) as impossible become possible. In understanding what made the North Dakota legislature’s adoption of this policy feature possible, Paul Maloney of North Dakota Right to Life highlighted some theoretical possibilities with his comment that the “people of North Dakota . . . have the kind of legislative body that would pass these kinds of pieces of legislation.”

. The motivating examples of North Dakota and California display the promise of federalism as envisioned by Brandeis. Pioneering states adopt novel policies and in doing so create some policies that may be borrowed and co-opted by other states.97 Policymakers in individual states can “customize” and choose policies that are best supported by their respective constituencies (Karch 2007; Ostrom 2008). Since legislatures are the primary institutions within the states tasked with adopting laws, it is worth exploring and uncovering the factors that contribute to legislatures inventing when they adopt policy. Are the factors that contribute to legislatures inventing the same as those that contribute to legislatures borrowing existing policy from other states? And what explanations are the most compelling in terms of predicting legislative invention?

. Recall that what I refer to in this book as invention, or the adoption of a policy that is new to a system of actors by a member of that system of actors, builds on the work of Jack Walker (1969).98 Walker is interested in determining what makes state legislatures “innovate” and defined innovationPage 94 → as occurring whenever a state adopts “a program or policy which is new to the states adopting it, no matter how old the program may be or how many other states have adopted it” (Walker 1969, 881). Walker’s own endeavor was motivated heavily by the work of Everett Rogers (1962), who determined that ideas and technologies (what he also termed “innovations”) spread out or diffuse across a population or system of actors according to an S-shaped pattern. In the context of Rogers’s diffusion theory, Walker wants to know why a state legislature would make an adoption that is new to that state regardless of where the adoption lies along the S-curve. My goal for the book in the context of Rogers’s diffusion theory is to evaluate why a state would adopt a policy that has not yet been introduced along the S-curve for all 50 states; and an integral part of this goal centers on figuring out whether the factors predicting invention are the same as those that predict borrowing. My goal for this chapter in the context of Rogers’s diffusion theory is to evaluate why a state legislature would adopt a policy that has not yet been introduced along the S-curve for all 50 states.

. Scholars have begun to tie government ideology to the diffusion of policy. Ideology describes how systematic one’s beliefs are along a conservative-to-liberal spectrum and has emerged as a powerful predictor of whether Page 96 →policymakers support and try to advance the policy proposals that they do (Krehbiel 1998). Ideology was missing a role in Walker’s story on legislative policy innovation in the states; one reason why is presumably that scholars had not yet devised a way to systematically measure ideological orientation within and across the U.S. states. In the intervening decades, scholars have engineered reliable measures of ideological orientation that allow for comparisons to be made within and across the states, with the result being an explosion in the use of ideology as an explanatory variable for a host of political and policy-specific outcomes.99