Research on state resistance

Given its emerging importance, state resistance to federal policy has been the focal point of a number of recent studies. This research emphasises the importance of societal interests and government officials in conditioning state resistance. In 2002, for example, Congress enacted HAVA in response to the 2000 presidential election crisis. One issue that HAVA did not resolve was the provision of paper records in electronic voting (Palazzolo et al. Reference Palazzolo, Moscardelli, Patrick and Rubin2008). Given HAVA’s silence on VVPR, some state governments enacted VVPR provisions of their own (Palazzolo et al. Reference Palazzolo, Moscardelli, Patrick and Rubin2008). Research demonstrates that the adoption of VVPR provisions is associated with such factors as Democratic control of state legislatures and the presence of VVPR advocacy organisations (Palazzolo et al. Reference Palazzolo, Moscardelli, Patrick and Rubin2008).

In 2005, with the support of the 9/11 Commission, Congress enacted an emergency supplemental appropriations bill that included the REAL ID Act (Regan and Deering Reference Regan and Deering2009). REAL ID required state-issued drivers’ licenses and identification cards to include a number of standard features, such as front-facing photographs (Homeland Security New Wire 2013) and security measures to prevent tampering, duplication and counterfeiting (National Conference of State Legislatures 2005). Research demonstrates that states with larger numbers of interest groups per capita were more likely to resist REAL ID (Regan and Deering Reference Regan and Deering2009).

Research on state resistance also highlights factors separate from societal interests and government officials as determinants of state resistance. In 2002, President Bush signed into law the NCLB. NCLB required states to foster accountability through instruments such as testing, school choice and the certification of teachers.Footnote ¹ States that failed to implement NCLB measures were faced with the loss of federal funding (Shelly Reference Shelly2008). Research demonstrates the importance of overarching demographic characteristics, such as state wealth, as correlates of resistance to NCLB mandates (Shelly Reference Shelly2008).

Taken together, existing research suggests that the influence, or lack thereof, of societal interests and government officials over state resistance varies substantially across policy-making contexts. As a means of advancing understanding of this variation, this article conceptualises state resistance through the notion that characteristics of public policies affect the politics of government decision making (Lowi Reference Lowi1964). Specifically, the salience and complexity of policies are hypothesised to affect the political factors that condition state resistance to federal mandates (Gormley Reference Gormley1986). Grounded in this framework, the article develops a unified approach to the study of state resistance that establishes general expectations and, concomitantly, informs measurement decisions and analytical specifications across instances of resistance. This approach makes it possible to interpret variation in state behaviour as a reflection of characteristics of broad classes of public policy.

Salience, complexity and the politics of public policy

Originally developed in the context of regulatory politics, the salience-complexity framework has proven its utility in generating insight into political participation and policy outcomes more generally (Meier and McFarlane Reference Meier and McFarlane1993; Eshbaugh-Soha Reference Eshbaugh-Soha2006; Nicholson-Crotty Reference Nicholson-Crotty2009). Salient issues are conceptualised as those that affect large numbers of people “in a significant way” (Gormley Reference Gormley1986, 598). For an issue to be considered salient, it must be both broad in scope and intense in conflict, as in the case of threats to necessities, cultural values and economic well-being (Gormley Reference Gormley1986). Examples of typically salient issues include energy, environment, employment and health and safety, with issues such as transportation, finance, commerce and communications ordinarily not generating intense, widespread attention (Gormley Reference Gormley1986).

Complex issues are conceptualised as those raising factual questions that “cannot be answered by generalists or laypersons” (Gormley Reference Gormley1986, 598). Complexity pertains to the technical aspects of issues, rather than administrative or other sources of uncertainties and difficulties (Gormley Reference Gormley1986). The presence of monopolies and emergence of new technologies are common indicators of complexity (Gormley Reference Gormley1986). Issues in policy areas such as energy, finance, environment and transportation are often high in complexity, with issues in housing, employment and consumer protection typically characterised as relatively straightforward from a technical point of view (Gormley Reference Gormley1986).

Both salience and complexity can vary across issues in given areas of public policy. Some health care issues, such as nursing home inspections, are less complex than other aspects of health care provision, insurance and regulation (Gormley Reference Gormley1986). In addition, the salience and complexity of issues can vary across time periods, as when new problems emerge, demographic conditions change and issues are redefined (Gormley Reference Gormley1986). Such variations imply that specific issues in particular time periods are relevant units of classification for the application of the salience-complexity typology.

The classification of salience and complexity can be approached in two basic ways. The first approach is argumentation. Such an approach was utilised when the typology was first articulated, with dozens of regulatory issues impressionistically placed into salience and complexity categories (Gormley Reference Gormley1986). Subsequent research also has relied heavily on argumentation as a means of classification (Meier and McFarlane Reference Meier and McFarlane1993; Eshbaugh-Soha Reference Eshbaugh-Soha2006; Nicholson-Crotty Reference Nicholson-Crotty2009). In many instances, such arguments are relatively straightforward, as when the abortion issue is classified as high in salience and low in complexity (Meier and McFarlane Reference Meier and McFarlane1993).

Salience and complexity also have been established through quantitative indicators. Researchers have measured salience as the amount of coverage issues have received in media outlets (Gerber and Teske Reference Gerber and Teske2000; Ringquist et al. Reference Ringquist, Worsham and Eisner2003; Nicholson-Crotty Reference Nicholson-Crotty2009). Complexity has been operationalised as the length of laws and regulations (Gerber and Teske Reference Gerber and Teske2000). Although such quantitative measures are different in orientation from impressionistic classifications, research has demonstrated a high level of association between the two approaches (Gerber and Teske Reference Gerber and Teske2000).

Classifying state resistance

Both argumentation and quantitative measurement are utilised as a means of classifying the four episodes of state resistance considered in the article. Three of the episodes have been the subject of existing studies of state opposition (Palazzolo et al. Reference Palazzolo, Moscardelli, Patrick and Rubin2008; Shelly Reference Shelly2008; Regan and Deering Reference Regan and Deering2009). The aim in reconsidering NCLB, REAL ID and HAVA is to analyse these instances through the unifying lens of the salience-complexity framework.

The fourth episode consists of state resistance to the mandates of ACA. Signed into law by President Obama in 2010, ACA addressed a number of issues related to health care coverage and its associated costs (Stolberg and Pear Reference Stolberg and Pear2010). Key provisions included the issuance of health insurance without regard to preexisting conditions, a mandate that individuals not covered by government or employer-sponsored plans purchase approved insurance and the expansion of Medicaid eligibility to millions of families with incomes in the vicinity of the poverty level (Kaiser Family Foundation 2012).

In a number of respects, ACA is one of the most significant social policies enacted in recent decades. According to both supporters and opponents, ACA is a defining accomplishment of the Obama administration (Zelizer Reference Zelizer2013). ACA has been the subject of a landmark Supreme Court decision in which the justices upheld the constitutionality of the act’s individual mandate (Negrin and De Vogue Reference Negrin and De Vogue2012). By any metric, ACA meets the standard of an issue that is both broad in scope and intense in conflict. As it affects a large number of Americans “in a significant way” (Gormley Reference Gormley1986, 598), ACA is an archetypal example of an issue that is high in salience.

ACA also serves as a benchmark for classification as high in complexity. ACA is 906 pages in length and requires the promulgation of dozens, if not hundreds, of regulations by the Department of Health and Human Services and other federal agencies.Footnote ² ACA is complex from a technical standpoint, as it raises economic and medical questions that cannot be readily resolved by individuals without specialised training (Gruber Reference Gruber2011).

By a number of metrics, it is straightforward to classify NCLB as high in salience. NCLB was discussed in 52 New York Times’ articles, whereas neither REAL ID nor HAVA were mentioned in more than 10 articles.Footnote ³ NCLB continues to be viewed as one of the major legacies of the Bush administration, as evidenced by the fact that media outlets across the country discussed NCLB when covering the dedication, in April 2013, of the George W. Bush Presidential Library and Museum.Footnote ⁴ Given this sustained attention, NCLB is categorised as high in salience.

By contrast, REAL ID is classified as low in salience. REAL ID was mentioned far fewer times in the New York Times than NCLB. A search of the website of the Gallup organisation reveals large numbers of surveys addressing ACA and NCLB. Not a single result is returned, by contrast, for the term “REAL ID”. Although the establishment of a national identification card is a concern for civil liberties advocates (Regan and Deering Reference Regan and Deering2009), it is not likely that large numbers of ordinary citizens were aware of the passage of REAL ID.

Despite the seminal nature of the 2000 presidential election crisis, HAVA attracted little coverage in the New York Times. The provision of paper records in electronic voting is a rather specific manifestation of election reform that is not likely to resonate deeply with large numbers of ordinary citizens. For example, public hearings on a VVPR pilot project in Georgia produced “very limited” participation (Office of the Secretary of State, Georgia 2007, 24). Given the article’s focus on state activism on the particular issue of VVPR, HAVA is categorised as low in salience.

HAVA also is classified as low in complexity. The foremost challenges in replacing punch card voting systems are not technical, but rather administrative in orientation. Such challenges are illustrated by the law’s main objectives:

NCLB also is classified as low in complexity. In the original application of the salience-complexity framework (Gormley Reference Gormley1986), education regulation was categorised, as a general matter, as low in complexity. NCLB’s overarching aim is to reduce the achievement gap between white and minority students through instruments such as choice and accountability (Dillon Reference Dillon2009). This aim is to be specifically achieved by implementing initiatives designed to foster parental choice, increase the quality of teachers, provide language instruction for students of limited English proficiency and improve accountability through regular administration of standardised testing.Footnote ⁶ Although administratively complex, such issues can be readily understood by ordinary citizens, the vast majority of whom have spent many years of their lives in the education system as students and parents.

By contrast, REAL ID, as an application of emerging technologies (Gormley Reference Gormley1986), is classified as high in complexity. Key provisions of REAL ID require the use of specialised knowledge to address technical issues associated with the creation of enhanced drivers’ licenses and identification cards.Footnote ⁷ For example, data stored on identification cards must be recorded in a machine-readable format. In addition, identification cards must incorporate security features designed to prevent fraudulent uses of encoded information.Footnote ⁸ Overall, REAL ID requires the resolution of factual questions that “cannot be answered by generalists or laypersons” (Gormley Reference Gormley1986, 598).

Expectations for political influence

As summarised in Figure 1, the four episodes of state resistance each occupy a distinctive location in the salience-complexity typology. According to the typology, each combination of salience and complexity is associated with expectations regarding the societal interests and government officials who influence policy making (Gormley Reference Gormley1986). Issues that are high in salience and high in complexity, such as ACA, are characterised by influence on the part of citizens and elected officials (Gormley Reference Gormley1986). In addition, the technical complexities associated with such issues provide specialised interest groups with opportunities to influence policy making (Gormley Reference Gormley1986).

Figure 1 Salience, complexity and political influence.

Citizens and elected officials are equipped to influence issues, such as NCLB, that are high in salience and low in complexity (Gormley Reference Gormley1986). Although specialised interest groups possess strong preferences regarding policy outcomes in such areas, these preferences are constrained by the imperative to respond to politicians and public opinion (Gormley Reference Gormley1986). By contrast, issues that are low in salience and high in complexity, such as REAL ID, are characterised by professionalised decision making on the part of specialised interest groups and their allies in government (Gormley Reference Gormley1986).

Finally, issues that are low in salience and low in complexity, such as HAVA, typically experience little mobilisation of any sort (Gormley Reference Gormley1986). Citizens, elected officials and specialised interest groups are, for the most part, not actively involved in the policy-making process for such issues (Gormley Reference Gormley1986). The end result is decision making independent of such political considerations (Gormley Reference Gormley1986).

Operationalising state resistance

Consistent with state resistance as a general matter (Gormley Reference Gormley2006; Regan and Deering Reference Regan and Deering2009), opposition to federal policy in the context of ACA, NCLB, HAVA and REAL ID has been manifested through a variety of instruments. This article focuses on legislative and executive resistance as the primary outcomes of interest. In recent years, the passage of legislation and signing of executive orders have been common means of opposing federal policy (Derthick Reference Derthick2001; Krane Reference Krane2007; Posner Reference Posner2007; Dinan Reference Dinan2008), as well as the focal point of research on state resistance (Palazzolo et al. Reference Palazzolo, Moscardelli, Patrick and Rubin2008; Shelly Reference Shelly2008; Regan and Deering Reference Regan and Deering2009). As legal actions of state governments, legislative and executive enactments are distinct in character from alternative forms of state resistance, such as lobbying efforts and committee hearings.

State resistance to NCLB and REAL ID is measured as passage of a statute or resolution in opposition to the respective federal mandates. Both statutes and resolutions are legal enactments, with statutes placing states in direct defiance of federal law and resolutions petitioning the federal government for policy modification (Shelly Reference Shelly2008). Eight states have enacted statutes or resolutions in response to NCLB, whereas 21 states have utilised such approaches to oppose the mandates of REAL ID.Footnote ⁹ State resistance to ACA is operationalised as the enactment of statutes or constitutional amendments opting out of ACA or challenging mandatory ACA provisions. A total of 21 states have opposed ACA through such instruments.Footnote ¹⁰ In the context of HAVA, state action consists of the enactment of VVPR policy in the absence of federal legislation. A total of 29 states have engaged in this form of opposition, either through the passage of legislation or signing of an executive order.Footnote ¹¹

In all four cases, state resistance manifested relatively quickly. The majority of resistance to NCLB occurred during 2003 and 2004 (Shelly Reference Shelly2008). Similarly, state challenges to REAL ID were all enacted within an 18-month period (Regan and Deering Reference Regan and Deering2009). Statutes and constitutional amendments in opposition to ACA occurred between 2010 and 2013.Footnote ¹² Finally, the adoption of VVPR policy was concentrated in the 2004–2007 period (Palazzolo et al. Reference Palazzolo, Moscardelli, Patrick and Rubin2008).

Explanatory variables

To assess expectations of policy-making influence derived from the salience-complexity typology, measures of state-level characteristics of citizens, elected officials and specialised interest groups are used. As laid out in Figure 1, citizen characteristics are expected to influence government decision making on issues of high salience such as ACA and NCLB, but not in the context of low salience issues such as HAVA and REAL ID. According to this expectation, citizen characteristics are expected to influence high-salience policy making as a general matter, without regard to particular directional, substantive effects. Citizen characteristics are operationalised as the political ideology of the state’s population. This variable, Liberalism, is constructed such that larger values denote more liberal populations (Berry et al. Reference Blalock1998).Footnote ¹³

A similar expectation holds for elected officials. Specifically, characteristics of officeholders are hypothesised to influence, without regard to substantive direction, state resistance to federal policy on issues of high salience such as ACA and NCLB, but not low salience issues such as HAVA and REAL ID. Elected official characteristics are measured as the number of institutions of state government that are controlled by the Republican Party. This variable, Republican, ranges from 0 in states in which neither chamber of the state legislature nor the governorship is controlled by the Republican Party to 3 in states in which all three institutions are under Republican control.Footnote ¹⁴

According to the salience-complexity typology, characteristics of specialised interest groups are expected to influence policy making, without regard to direction, on issues of high complexity such as ACA and REAL ID, but not in the context of low complexity issues such as NCLB and HAVA. Characteristics of specialised interest groups are operationalised as the prevalence of interest groups active in state policy making (Gray and Lowery Reference Gray and Lowery1996). This variable, Interest Groups, which has been utilised in research on state resistance (Regan and Deering Reference Regan and Deering2009), is specifically measured as state population divided by the number of registered lobbying organisations.Footnote ¹⁵

In addition to being a function of factors derived from the salience-complexity framework, resistance to federal action is likely affected by the determinants of state policy making more generally. As a means of accounting for such determinants, measures of three types of state characteristics are incorporated into the analysis. These measures are not designed to assess specific expectations, but rather to ensure that the results regarding salience and complexity are robust to the inclusion of factors that are typically associated with state policy making.

The first category of state characteristics consists of population demographics – the percentage of the population composed of racial and ethnic minorities (Non-White) and percentage of the population that is 65 years of age and older (65 Years). The second category measures economic conditions – the per capita income of the population (Income) and percentage of the population living in poverty (Poverty).Footnote ¹⁶ The third category – the political capacity of the state’s government – incorporates into the analysis an indicator of legislative professionalism (Professionalism) (Squire Reference Squire2007).Footnote ¹⁷

Analysing the correlates of state resistance

The dependent variables are dichotomous indicators of state resistance to each of the federal policies under investigation. Given this construction, logistic regression is utilised to model the determinants of state opposition (Aldrich and Nelson Reference Aldrich and Nelson1984). The logit analysis is applied to cross-sectional data, with the state as the unit of observation.Footnote ¹⁸ The results of the logistic regressions are reported in Table 1.

Table 1 Determinants of state resistance

Notes: The numbers in the cells are parameter estimates from logistic regression equations. The numbers in parentheses are robust standard errors. All four equations were also estimated without Republican. This allows the analysis to include Nebraska, which has a unicameral, nonpartisan legislature. A number of multicollinearity assessments were conducted. Variance inflation factors, which were calculated for all four equations, indicate that multicollinearity is not a significant problem. Income and Poverty, however, are correlated at a fairly high level. All four of the equations were estimated with Income and Poverty dropped one at a time. In all alternative specifications, the results for variables testing state resistance hypotheses were not substantially affected. All hypothesis tests are two-tailed.

ACA=Affordable Care Act; NCLB=No Child Left Behind Act; HAVA=Help America Vote Act.

Statistically significant at *p<0.1; **p<0.05; ***p<0.01.

Overall, the results of the analysis are consistent with expectations for state resistance derived from the salience-complexity typology. In Table 1, the results highlighted in bold affirm expectations regarding the societal interests and government officials who influence decision making within particular classes of public policy. For example, the parameter estimates for Liberalism, Republican and Interest Groups are all statistically significant in the analysis of ACA resistance. These estimates confirm the expectation, presented in Figure 1, that citizens, elected officials and specialised interest groups all influence policy making in areas characterised by high levels of both salience and complexity.

The results also support the contrasting expectation that citizens, elected officials and specialised interest groups do not condition state resistance in areas that are low in both salience and complexity. This affirmation is demonstrated by the fact that none of the parameter estimates for Liberalism, Republican or Interest Groups are statistically significant in the analysis of HAVA resistance. Given that null expectations are being assessed in the context of these expectations, as well as some of the hypotheses regarding state resistance to NCLB and REAL ID, conventional standards must be adjusted as a means of ensuring appropriate evaluation. Rejecting the null hypothesis when the probability of committing a Type I false positive error is 5% stacks the deck in favour of failing to reject the null. In circumstances in which the null hypothesis is the expectation under investigation, a common approach is to utilise a more generous standard and reject the null hypothesis when the probability of committing a Type I error is at least 10% and perhaps as high as 25% (Blaylock Reference Blalock1979; Julnes and Mohr Reference Julnes and Mohr1989; Hall and Grofman Reference Hall and Grofman1990). Given that such probability levels are exceeded in the analysis of HAVA resistance, the results confirm the null hypotheses in question.

Expectations regarding the societal interests and government officials who influence policy making vary across types of stakeholders in the context of mixed salience and complexity issues such as NCLB and REAL ID. For REAL ID, an issue that is low in salience and high in complexity, the expectation is that characteristics of specialised interest groups, but not of citizens or elected officials, affect government decision making. This configuration of expectations is confirmed by the fact that the parameter estimate for Interest Groups is statistically significant, whereas the estimates for Liberalism and Republican do not allow for the rejection of their respective null hypotheses.

The expectation for NCLB, an issue of high salience and low complexity, is that characteristics of citizens and elected officials, but not of specialised interest groups, separate resisting states from states that have not enacted statutes or resolutions in response to the Act. This pattern of expectations is affirmed by the statistically significant parameter estimate for Liberalism and the estimate for Interest Groups that does not cross the threshold of rejecting the null hypothesis. Among all of the variables that test hypotheses derived from the salience-complexity typology, only Republican in the analysis of NCLB resistance does not match its theoretical expectation. Although elected officials are hypothesised to condition state resistance to NCLB, the estimate for Republican is not statistically significant.

Not only are the overall results of the analysis consistent with expectations derived from the salience-complexity typology, these results are substantively meaningful as well.Footnote ¹⁹ For example, a relatively conservative state in which Liberalism was one standard deviation below the mean had an 86.23% likelihood of enacting statutes or constitutional amendments opting out of ACA or challenging mandatory ACA provisions. The likelihood decreased to 4.61% for a relatively liberal state in which Liberalism was one standard deviation above the mean.Footnote ²⁰

Similarly, as the number of institutions of state government controlled by the Republican Party increased, states were substantially more likely to resist the mandates of ACA. A state in which the Republican Party did not control either chamber of the state legislature or the governorship had a 14.44% likelihood of ACA resistance. As the value of Republican increased to 1 or 2, the likelihood of ACA resistance increased to 30.28 and 56.04%, respectively. In states in which both chambers of the state legislature and the governorship were controlled by the Republican Party, the likelihood of enacting statutes or constitutional amendments in opposition to ACA reached 75.34%.

The prevalence of interest groups active in state policy making had a substantively significant impact on state resistance to ACA. A state in which Interest Groups was one standard deviation below the mean had a 15.41% likelihood of enacting statutes or constitutional amendments in opposition to ACA. The likelihood increased to 64.57% for a state in which Interest Groups was one standard deviation above the mean.Footnote ²¹

In sum, the results offer evidence in support of expectations, drawn from the salience-complexity framework (Gormley Reference Gormley1986), regarding the importance of characteristics of citizens, elected officials and specialised interest groups in conditioning state resistance to federal mandates. Overall, the analysis suggests that state resistance can be characterised as a strategic response to federal action that varies in systematic ways across broad classes of policies.

State resistance and the politics of public policy

Over the past several decades, a significant shift has occurred in the character of American federalism. A long-standing emphasis on cooperation and fiscal tools has to some extent given way to coercion and regulation (Kincaid Reference Kincaid1990). State resistance to the dictates of coercive federalism occurred with rising frequency during the presidency of George W. Bush (Krane Reference Krane2007), a trend that has continued under the Obama administration (Metzger Reference Metzger2011). Despite its emerging importance, as well as insights generated by existing research (Krane Reference Krane2007; Posner Reference Posner2007), state resistance is not yet well understood as a general, strategic phenomenon. The aim of this article is to lay the groundwork, both theoretically and empirically, for analysing the overarching factors that separate resisting states from states that do not oppose federal policy.

The analysis utilises a well-established public policy framework (Gormley Reference Gormley1986) to develop expectations regarding the conditions under which states are likely to resist federal action. The framework identifies the societal interests and government officials who are hypothesised to influence policy making on issues of varying salience and complexity (Gormley Reference Gormley1986). The veracity of these hypotheses is confirmed in the context of state resistance to four federal laws (ACA, NCLB, HAVA and REAL ID).

Given the contemporary prevalence of the passage of nullifying legislation as a means of opposing federal mandates, the analysis of the four laws investigated in the article represents an initial evaluation of state resistance as a strategic response to coercive federalism. State officials have pursued nullification in a wide variety of policy areas. For example, dozens of legislatures pushed back against federal gun policy proposals in the aftermath of the 2012 Newtown, Connecticut mass shooting by introducing bills declaring such proposals null and void within state borders (Simon Reference Simon2013). In Alabama, the Senate passed a bill stating that, “All federal acts, laws, orders, rules or regulations regarding firearms are a violation of the Second Amendment” (Associated Press 2013). Such examples suggest the importance of extending the research presented here across policy areas.

The article demonstrates the promise of the salience-complexity typology as an orienting framework for the analysis of state resistance. Salience and complexity condition the circumstances under which characteristics of citizens, elected officials and specialised interest groups influence state resistance. In this regard, the framework complements analyses of specific instances of state resistance by drawing attention to factors that, as a general matter, separate resisting states from states that do not oppose federal policy.

More broadly, the application of the salience-complexity typology to state resistance illustrates the insights that can be derived from approaching issues in federalism from the perspective of the politics of the policy-making process. The typology theorises about the intersection of political mobilisation and characteristics of public policies. In making such connections, the typology, and by extension the article, integrates the study of state resistance into the wider research programme of utilising theoretical insights from political science to advance the understanding of the making and implementation of public policy.