Introduction
Governors used their emergency powers to deploy diverse strategies to combat rising case numbers during the COVID-19 pandemic. In April 2020, Maryland Governor Larry Hogan banned local governments from implementing mask standards that conflicted with his (Hogan, Executive Order 20-04-15-01). Colorado Governor Jared Polis allowed local governments to implement their own mask standards – only if they were stricter than his (Polis, Executive Order D 2020 138). Iowa Governor Kim Reynolds took a broader approach, prohibiting local governments from implementing mask mandates without setting statewide mask standards (National League of Cities 2022). Each of these gubernatorial strategies is an example of preemption, whereby a law passed by a higher authority takes precedence over a law passed by a lower one (Barber and Dynes Reference Barber and Dynes2021). However, each governor’s strategy differed in how they preempted their local governments. Many governors barred local regulation altogether, while others granted local governments some flexibility, allowing mandates that were more or less restrictive than their own.
A flourishing research agenda examines state preemption of local governments. This research agenda is increasingly important, given that state preemption of local governments continues to rise (Flavin and Shufeldt Reference Flavin and Shufeldt2020; Fowler and Witt Reference Fowler and Witt2019; Kogan Reference Kogan2017; Swanson and Barrilleaux Reference Swanson and Barrilleaux2020). Scholars have begun classifying the different preemption styles, including floor, ceiling, and vacuum preemption (Goodman and Hatch Reference Goodman, Hatch and McDonald2021; SoRelle and Walker Reference SoRelle and Walker2016; Wagner et al. Reference Wagner, Davidson, Haddow, Jones, McFarland and Rainwater2019). Ceiling preemptions restrict local governments from issuing laws different or stricter than their state’s, while floor preemptions restrict local governments from issuing laxer laws. Vacuum preemptions prohibit local governments from doing something without the state setting standards of its own (National League of Cities 2022). These three distinct preemption styles are not hierarchical; however, vacuum preemptions are viewed as more restrictive than ceiling and floor preemptions (National League of Cities 2022).
However, there are several shortcomings in the existing literature. Although many scholars descriptively examine the different preemption styles, empirical examinations are rare and typically focus on ceiling and floor preemptions (SoRelle and Walker Reference SoRelle and Walker2016). Existing empirical analyses often exclude vacuum preemptions, which I demonstrate in this paper are frequently employed by governors. Existing literature primarily focuses on federal preemption style, often overlooking how and why states preempt their local governments. Moreover, most studies focus on legislative and judicial preemption. Although gubernatorial preemptions occur during times of emergency, several aspects of gubernatorial preemption make it interesting to examine. For example, unlike legislative preemption, gubernatorial preemption is unilateral, effective immediately, and can be done year-round while lawmakers are out of session. Therefore, gubernatorial preemption can reflect a governor’s political aspirations, personal experiences, and ideology without the moderating effect of other elected officials.
I remedy these shortcomings by unveiling the institutional and political conditions influencing gubernatorial preemption style during the COVID-19 pandemic. I introduce original data to complement the dataset of COVID-19-related executive orders introduced by Weissert et al. (Reference Weissert, Uttermark, Mackie and Artiles2021) by classifying their style of preemption: ceiling, floor, or vacuum (Goodman and Hatch Reference Goodman, Hatch and McDonald2021; Wagner et al. Reference Wagner, Davidson, Haddow, Jones, McFarland and Rainwater2019). Findings indicate that local autonomy, partisanship, and ideological asymmetry are associated with different preemption strategies. Governors in states with high local autonomy rely on ceiling and floor preemptions. Republican governors are more likely to issue ceiling preemptions and less likely to issue vacuum preemptions. Governors in states with ideologically dissimilar local governments are more likely to issue vacuum preemptions and less likely to issue ceiling and floor preemptions.
When non-preempting previsions are dropped from the analysis, local autonomy does not significantly affect issuing one type of preemption over another. On the other hand, Republican governors are more likely to issue both ceiling and floor preemptions over vacuum preemptions. Governors in states with high ideological asymmetry are less likely to issue ceiling and floor preemptions over vacuum preemptions. These findings have significant insights for scholars trying to understand gubernatorial behavior, the relationships between state and local governments, and how theories of federalism can teach us more about how governments respond to crises.
Preemption and the COVID-19 pandemic
State–local preemption in the US
American federalism continues a trend toward centralization, with the federal government centralizing over state governments and state governments centralizing over local governments (Bowman and Kearney Reference Bowman and Kearney2011; Kincaid Reference Kincaid2018; Stephens Reference Stephens1974). State centralization has become increasingly preemptive. States preempt local governments over numerous policy domains, such as environmental regulations (Welland Reference Welland2000), gun control (Pomeranz, Silver, and Lieff Reference Pomeranz, Silver and Lieff2021), the minimum wage (Riverstone-Newell Reference Riverstone-Newell2017), discrimination in public facilities (Dorosin Reference Dorosin2020), rent control (Treskon and Docter Reference Treskon and Docter2020), and immigration customs and enforcement (Olivas Reference Olivas2007). State preemption has also become more punitive, with states threatening noncompliant local governments with fines or state aid cutoffs (Briffault Reference Briffault2018; Weissert et al. Reference Weissert, Uttermark, Mackie and Artiles2021).
Gubernatorial behavior: what do we know?
Governors are responsible for implementing state laws and overseeing the operation of the state executive branch (National Governors Association 2023). Governors have various tools to fulfill these responsibilities, including executive orders, executive budgets, and legislative proposals and vetoes (National Governors Association 2023). A limited but growing literature suggests that governors utilize these tools, particularly executive orders, for many purposes. Namely, governors issue executive orders to shape public policies such as immigration (Arellano Reference Arellano2012), environmental regulations (Gakh, Vernick, and Rutkow Reference Gakh, Vernick and Rutkow2013), LGBT rights (Sellers Mitchell Reference Sellers Mitchell2017), and public health (Curley and Federman Reference Curley and Federman2020). Furthermore, governors issue executive orders in response to legislative delegation or to circumvent legislative authority (Barber, Bolton, and Thrower Reference Barber, Bolton and Thrower2019; Cockerham and Crew Reference Cockerham and Crew2017; Ferguson Margret and Bowling Reference Ferguson Margret and Bowling2008). In sum, as chiefs of the state, governors are significant entities who can change the course of a state’s political trajectory.
Why gubernatorial preemption?
Preemptions can and have been issued by all three branches of government. An abundant literature examines legislative preemption (Boeckleman and Day Reference Boeckelman and Day2021; Einstein and Glick Reference Einstein and Glick2017; Flavin and Shufeldt Reference Flavin and Shufeldt2020; Goodman Reference Goodman2019; Pomeranz, Silver, and Lieff Reference Pomeranz, Silver and Lieff2021; Riverstone-Newell Reference Riverstone-Newell2017; SoRelle and Walker Reference SoRelle and Walker2016), while some studies shed light on the role of the judiciary in preemption (Briffault Reference Briffault2018; Stahl Reference Stahl2020; Swanson and Barrilleaux Reference Swanson and Barrilleaux2020).
However, we know little about gubernatorial preemption compared to legislative and judicial preemption. This is because gubernatorial preemption is only accessed through the emergency powers awarded to governors during crises, making this form of preemption less frequent than legislative and judicial preemption (McDonald, Goodman, and Hatch Reference McDonald, Goodman and Hatch2020). However, gubernatorial preemption warrants attention in the literature because governors behave differently than legislators and judges concerning preemption. For example, state legislators interested in preempting local action face significant time constraints due to the cartelized agenda-setting system (Cox and McCubbins Reference Cox and McCubbins2005). On the other hand, governors can issue preemptions immediately and unilaterally. Therefore, this preemption can reflect a governor’s political ambition, personal experiences, partisanship, and ideology without the moderating effect of other elected officials and interests (Ferguson Margret and Bowling Reference Ferguson Margret and Bowling2008).
The COVID-19 pandemic allowed governors to become the primary preemptors of local action (Bowman and McKenzie Reference Bowman and McKenzie2020; Weissert et al. Reference Weissert, Uttermark, Mackie and Artiles2021). This is because, during times of emergency, states grant governors significant control over state and local affairs – including dictating what local governments can and cannot do (Waugh Reference Waugh2007). For example, Florida statutes grant governors the ability to “assume direct operational control over all or any part of the emergency management functions.” Florida governors may direct emergencies “until the Governor finds that the threat or danger has been dealt with” (Florida Statutes 2021). Pennsylvania statutes allow governors to “utilize all available resources of the Commonwealth Government and each political subdivision of this Commonwealth as reasonably necessary to cope with the disaster emergency” (Pennsylvania Statutes 2022). Governors across the US did precisely this – preempting local governments over business closures, mask mandates, social distancing, and more – without legislative constraint. Given the governors’ unique role, the COVID-19 pandemic provides an excellent opportunity to study gubernatorial preemption style.
Defining the three preemption styles
Governors preempt local action in several ways. In all cases, the executive orders passed by a governor take precedence over orders passed by their local governments (Barber and Dynes Reference Barber and Dynes2021). Goodman and Hatch (Reference Goodman, Hatch and McDonald2021) and Wagner et al. (Reference Wagner, Davidson, Haddow, Jones, McFarland and Rainwater2019) identified three distinct modalities of preemption. Ceiling preemptions prevent local governments from establishing different or stricter ordinances than those established by state law; local governments can only issue equally restrictive or less restrictive measures than their state. For example, West Virginia Governor Jim Justice prohibited local governments from issuing mask mandates that were stricter than his own. During the pandemic, ceiling preemptions were desirable to local officials who felt that mask and vaccine mandates were a government overreach that infringed on personal freedom (Vuolo, Kelly, and Roscigno Reference Vuolo, Kelly and Roscigno2020). Floor preemptions are essentially the opposite of ceiling preemptions. Floor preemptions prevent local governments from establishing laxer ordinances than those established by state law; local governments can only issue equally restrictive or more restrictive measures than the state (Goodman and Hatch Reference Goodman, Hatch and McDonald2021). For example, Nevada Governor Steve Sisolak declared that local governments would follow the state’s minimum mask requirements but could adopt additional protective measures. During the pandemic, floor preemptions appealed to many local officials because they established a statewide standard and still left them with the power to enact more robust regulations (Vuolo, Kelly, and Roscigno Reference Vuolo, Kelly and Roscigno2020).
Vacuum preemptions occur when a state government chooses not to regulate a particular policy but still forbids local governments from doing so, creating a regulatory vacuum. For example, Iowa Governor Kim Reynolds did not issue a mask mandate in her state but declared that local governments had no authority to implement mask mandates (National League of Cities 2022). Vacuum preemption is often undesirable to local officials because it leaves a policy area of concern entirely unregulated. Table 1 provides additional examples of each type of preemption from the dataset.Footnote 1
Table 1. Examples of the types of preemption affecting local governments
What influences a governor’s preemption style?
A governor’s preemption style is influenced by the institutional and political factors they face. The first factor relates to the institutional condition of a governor’s state, while the final three factors relate to the political landscape of a governor’s state.
Local autonomy
Localities do not have the same immunity from state preemption that states have from federal preemption (Blair et al. Reference Blair, Cooper, Wolfe and Worker2020; Kamal et al. Reference Kamal, Cox, Fehr, Ramirez, Hortsman and Levitt2018; Rosenbaum and Westmoreland Reference Rosenbaum and Westmoreland2012).
The local autonomy movement aimed to institutionalize municipal protections, such as Home Rule, to provide localities with legal frameworks to challenge state preemption (Bunch Reference Bunch2014). Such frameworks protect localities from state overreach; if a state wishes to reclaim power from a municipality, it must first prove the existence of conflict between state and local ordinances (Einstein and Glick Reference Einstein and Glick2017). As a result, localities across many states enjoy increased importance, discretion, and capacity (Wolman et al. Reference Wolman, Manmon, Bell, Brunori, Bell, Brunori and Youngman2010).
However, states are increasingly testing the boundaries of local autonomy by preempting issues that were once within the purview of local government authority, such as zoning and public safety (Stahl Reference Stahl2020). When local governments have considerable autonomy, preemption becomes the only way for states to guarantee local compliance. Legislators and judges have preempted high-autonomy local governments over a wide variety of issues, including land use (Stahl Reference Stahl2020), hydrofracking (Knight and Gullman Reference Knight and Gullman2015), firearms (Pomeranz, Silver, and Lieff Reference Pomeranz, Silver and Lieff2021), and the minimum wage (Goodman Reference Goodman2019). A growing body of evidence suggests that governors also override high-autonomy local governments. For example, Weissert et al. (Reference Weissert, Uttermark, Mackie and Artiles2021) found that higher local autonomy was associated with increased gubernatorial preemption during the pandemic. Moreover, many state courts ruled that municipalities could not exercise autonomy if it conflicted with the broad emergency powers reserved for governors (McDonald, Goodman, and Hatch Reference McDonald, Goodman and Hatch2020).
I anticipate that governors will constrain local governments with high autonomy through ceiling preemptions. The logic behind this is that local governments with high autonomy tend to innovate and adopt various policies. For example, high-autonomy local governments were at the forefront of mitigation during the pandemic, often instituting bans on large gatherings and closing bars and restaurants (Brasch and Lutz Reference Brasch and Lutz2020).
To avoid having to contend with a myriad of local regulations during a public health crisis, governors will impose uniformity by banning local governments from taking action different from the statewide policy (Wagner et al. Reference Wagner, Davidson, Haddow, Jones, McFarland and Rainwater2019). In other words, governors will issue ceiling preemptions to prevent policy diffusion among their local governments (Mallinson Reference Mallinson2020).
The use of ceiling preemption to strategically circumvent local governments with high autonomy was apparent throughout the pandemic. For example, local governments throughout Tennessee – which have considerable autonomy – experienced unprecedented ceiling preemptions throughout the pandemic. Governor Bill Lee established a regulatory ceiling over stay-at-home orders by preempting “any other order issued by a local official or local government entity that contravenes or would limit the application of the provisions of this order.” Governors in Arizona, Georgia, Florida, Mississippi, South Carolina, Texas, and West Virginia issued orders with similar regulatory ceilings (Davidson and Haddow Reference Davidson and Haddow2020).
-
• Local Autonomy Hypothesis: Governors in states with higher local autonomy are likelier to issue ceiling preemptions than governors in states with lower local autonomy.
Partisanship
For decades, the Republican Party has supported decentralization, while the Democratic Party has supported centralization. Following this logic, one might expect that Democratic governors are more likely to preempt local governments than Republican governors (Dinan and Heckelman Reference Dinan and Heckelman2020). However, it is increasingly the case that Democratic and Republican governors alike preempt their local governments (Barber and Dynes Reference Barber and Dynes2021; Einstein and Glick Reference Einstein and Glick2017; McDonald, Goodman, and Hatch Reference McDonald, Goodman and Hatch2020; Treskon and Docter Reference Treskon and Docter2020).
However, Democratic and Republican governors use different strategies to preempt local governments. It is increasingly the case that Republican-controlled states tend to issue ceiling preemptions, allowing local governments only to implement equally or less restrictive measures than their state. In contrast, Democrat-controlled states tend to issue floor preemptions, allowing local governments only to implement equally or more restrictive measures than their state. Republicans tend to issue ceiling preemptions in the areas of business and commerce to prevent increased regulations and restrain the scope of government, while Democrats tend to issue floor preemptions on civil rights and healthcare, ensuring minimum protections for individuals (Conlan and Posner Reference Conlan and Posner2016; Meyer-Gutbrod Reference Meyer-Gutbrod2018; SoRelle and Walker Reference SoRelle and Walker2016).
The increase in ceiling preemption has been tied to battles between Republican state governments and Democrat local governments (Mallinson Reference Mallinson2020). For example, Flavin and Shufeldt (Reference Flavin and Shufeldt2020) found that states with Republican trifectas are significantly more likely to prohibit local governments from raising the minimum wage above the state ceiling. Blair et al. (Reference Blair, Cooper, Wolfe and Worker2020) found that Republican-dominated states tend to issue ceiling preemptions over workers’ rights laws, including the minimum wage, project labor agreements, and paid leave.
Throughout the pandemic, Republican and Democratic governors alike continued these preemption trends, with Republican governors limiting local government restrictions on business closures, mass gatherings, and quarantines and Democrats setting minimum standards on these issues. Republican governors issued ceiling preemptions over public health by banning or limiting local mask and vaccine restrictions, while Democrats issued floors requiring masks and vaccines. For example, Mallinson (Reference Mallinson2020) discussed the increased reliance on ceiling preemption by Republican governors – including in Arizona, Georgia, Florida, and Mississippi – to override local governments and set a statewide standard. McDonald, Goodman, and Hatch (Reference McDonald, Goodman and Hatch2020) leveraged a database of executive orders focused on staying at home and sheltering in place. They found that Republican governors tend to prevent local government from exceeding statewide standards, while Democrat governors tend to allow local governments to set their policies at higher levels. Chen, Boadu, and Xiao (Reference Chen, Boadu and Xiao2022) found that Democrat governors are significantly more likely to adopt regulatory floors through mask mandates than Republican governors.
-
• Republican Governor Hypothesis: Republican governors are more likely to issue ceiling preemptions than Democratic governors.
-
• Democratic Governor Hypothesis: Democrat governors are likelier to issue floor preemptions than Republican governors.
Ideological asymmetry
Ideology heavily influences state preemption activity. Barber and Dynes (Reference Barber and Dynes2021) found that local officials whose cities are ideologically distant from their state government are likelier to report being preempted. Swanson and Barrilleaux (Reference Swanson and Barrilleaux2020) found similar results concerning judicial preemption: state courts are more likely to preempt local ordinances when local and state governments are ideologically dissimilar. A limited but emerging literature suggests that governors also preempt ideologically distant local governments. For example, Weissert et al. (Reference Weissert, Uttermark, Mackie and Artiles2021) found that governors in ideologically asymmetric states were significantly more likely to preempt local action during the COVID-19 pandemic.
Governors have policy preferences that can be negated or weakened by the various actions of local governments in their state (Tausanovitch and Warshaw Reference Tausanovitch and Warshaw2014). By utilizing ceiling preemptions, governors can prevent ideologically dissimilar local governments from moving policies away from their preferred positions by barring them from passing anything that deviates from the statewide policy. For example, Texas Governor Greg Abbott issued a series of ceiling preemptions prohibiting local governments from implementing any restrictions on mask mandates, stay-at-home orders, or restrictions on religious gatherings that differed from the statewide restrictions (Platoff Reference Platoff2020). Notably, these preemptions disproportionately targeted liberal cities – including Austin, Dallas, and San Antonio – that had attempted to impose stricter restrictions than the state. Attorney General Ken Paxton warned these cities to roll back their “unenforceable” restrictions beyond the state’s or face lawsuits (Davidson and Haddow Reference Davidson and Haddow2020).
-
• Ideological Asymmetry Hypothesis: Governors in states with greater ideological asymmetry are likelier to issue ceiling preemptions than governors in states with less ideological asymmetry.
In sum, I expect a governor’s preemption style to be influenced by both the institutional and political circumstances within their state. I now turn to the empirical analysis.
Empirical analysis
Data collection
I extend Weissert et al.’s (Reference Weissert, Uttermark, Mackie and Artiles2021) database containing 897 gubernatorial executive order provisions issued in response to the pandemic from March 2020 to August 2020.Footnote 2 Weissert et al. (Reference Weissert, Uttermark, Mackie and Artiles2021) collected every provision from the Council of State Governments (CSG) COVID-19 Resources for State Leaders and filtered to only include provisions referencing cities, municipalities, counties, towns, townships, villages, local governments, local orders, local authorities, other governmental units, and preemption.
The extended database adds an additional 321 provisions issued from September 2020 to December 2021, making for a total of 1,219 provisions from March 2020 to December 2021. As observed in Weissert et al. (Reference Weissert, Uttermark, Mackie and Artiles2021), states continued to vary significantly in their use of executive orders in response to the pandemic. New Jersey issued the most provisions affecting local government (150), while Nebraska issued the least (1). The average state issued twenty-five provisions that affected local government.
Dependent variable
My dependent variable measures the type of preemption issued by a governor during the pandemic. Applying the preemption framework developed by Wagner et al. (Reference Wagner, Davidson, Haddow, Jones, McFarland and Rainwater2019), provisions were coded into one of three categories: ceiling preemption, floor preemption, and vacuum preemption.Footnote 3 Including all executive orders that mention local government is essential to represent the universe of executive orders. Therefore, a fourth baseline category is reserved for all executive orders that do not preempt local governments. This means that these provisions either empower local governments or are neutral. Empowering provisions allow local governments to act on a policy issue. For example, Wisconsin Governor Tony Evers enabled local governments to address local coronavirus outbreaks through local executive orders (Evers, Executive Order 2020-30). Neutral provisions typically contain coordination, technical language, or clarifying language. For example, Minnesota Governor Tim Waltz encouraged local governments to work collaboratively with establishments to allow for outdoor dining services (Waltz, Executive Order 20-0).
Two coders independently coded a random sample of ten percent of the provisions, assigning them to one of the four categories of the dependent variable. Table 2 shows the inter-coder reliability scores between the two coders.
Table 2. Inter-coder reliability scores
Table 3 reports the percentage of each type of preemption throughout the pandemic. Ceiling preemptions are common, representing 34.29 percent of the data. These were followed closely by non-preempting provisions, which are 33.31 percent of the data. Floor and vacuum preemptions were 16.08 percent and 16.32 percent of the data, respectively. As one can see, although vacuum preemptions are often neglected in the literature on preemption style, they are quite commonplace.
Table 3. Distribution of executive order provisions
There exists great variation in the types of preemption used across states. Figure 1 reports the percentage of each type of preemption issued by the governor in each state throughout the pandemic. Some states rely heavily on a single type of preemption, while others utilize a combination of preemption types. Ceiling preemptions were heavily used in Iowa, Maryland, and North Dakota. Floor preemptions were most common in Alabama, Montana, and Rhode Island, and vacuum preemptions were frequent in Louisiana and Oregon. Non-preempting provisions were common in Ohio and New Hampshire.
Figure 1. Use of preemption style by state.
Independent variables
The first independent variable of interest is local autonomy. I utilize Wolman et al.’s (Reference Wolman, Manmon, Bell, Brunori, Bell, Brunori and Youngman2010) index of local autonomy. The Wolman et al. (Reference Wolman, Manmon, Bell, Brunori, Bell, Brunori and Youngman2010) index utilizes three dimensions of local autonomy: local importance, local discretion, and local capacity. The authors identify twenty-one variables within these three dimensions and use factor analysis to create an overall local autonomy score. Higher scores indicate greater levels of local autonomy, while lower scores indicate lower levels. A score of zero indicates moderate levels of local autonomy. The state with the highest autonomy score is Kansas (0.861), while the state with the lowest is Hawaii (-1.213). States with moderate levels of local autonomy include Indiana (-0.04) and North Carolina (-0.036). Measures of local autonomy, such as Home Rule and Dillon’s Rule, are binary and do not capture variations in local authority levels; therefore, I opt for the measure by Wolman et al. (Reference Wolman, Manmon, Bell, Brunori, Bell, Brunori and Youngman2010).
The second independent variable is the political party of a state’s governor, retrieved from Ballotpedia. Republican governors are coded 1, while Democratic governors are coded 0. In 2020, there were 27 Republican governors and 23 Democratic governors; in 2021, there were 28 Republican governors and 22 Democratic governors.Footnote 4
The third independent variable is the ideological distance between a state and its localities. I utilize Tausanovitch and Warshaw’s (Reference Tausanovitch and Warshaw2014) state and county ideology measure, which is based on aggregating 275,000 policy-related survey responses from US adults. I averaged the county ideology scores from Tausanovitch and Warshaw (Reference Tausanovitch and Warshaw2014). Then, I subtract the average county ideology score from its state government’s ideology score to form an asymmetry score. A state with complete asymmetry is coded as a 1; a 0 represents a state with complete symmetry. The state with the most ideological asymmetry is Alaska (0.419); the state with the least is Rhode Island (0.03).
Control variables
I control whether a governor’s state government is divided since governors may rely on preemption via executive orders when bargaining with their legislature is costly (Cockerham and Crew Reference Cockerham and Crew2017; Sellers Mitchell Reference Sellers Mitchell2017). Divided governments include when both parties control at least one chamber of the legislature or the governor’s office (Barber and Dynes Reference Barber and Dynes2021). Divided governments are coded 1, while unified governments are coded 0. I also control for two characteristics of a governor that may influence their preemption style. I control for a governor’s institutional powers; stronger governors may preempt more aggressively than weaker governors. I utilize Ferguson’s (Reference Ferguson, Gray, Hanson and Kousser2013) index of gubernatorial power, which accounts for separately elected executive branch officials, tenure potential, appointment power, budget power, veto power, and party control. I also control for whether a governor is a lame-duck; governors may be less concerned about the consequences of preempting their local governments when they are ineligible for reelection. This measure is binary, where 1 is a lame-duck governor, and 0 is a governor eligible for reelection. I also capture the overall partisan composition of a state’s electorate, which will likely influence a governor’s actions. This is measured using a state’s vote share for President Trump in the 2020 election.
The pandemic unequivocally produced both economic and public health downfall within states; the magnitude of these crises is likely to influence a governor’s preemption style. Therefore, I control for a state’s average monthly COVID-19 cases per 100,000, retrieved from the Center for Disease Control (CDC), and a state’s monthly unemployment rate, retrieved from the Bureau of Labor Statistics (BLS). Finally, states with higher populations might be adversely impacted by the spread of COVID-19; therefore, I control for the log of a state’s population in 2020. Table 4 details the summary statistics for every independent and control variable used for analysis.
Table 4. Summary statistics
Estimating the models
I estimate two multinomial logistic regression models to test the political and institutional influences on preemption style. I use this method because the dependent variable – preemption style – is categorical, with two or more unordered levels. In the first model, I examine the likelihood of adopting one type of preemption over a non-preempting provision. In the second model, I examine the likelihood of adopting one type of preemption over another type of preemption. To address the potential problems with clustering, both models include phase-fixed effects and state-clustered standard errors.Footnote 5, Footnote 6
Since multinomial logits are not directly interpretable from estimated coefficients beyond their direction and significance levels, I also report the predicted probabilities of each independent variable. The predicted probabilities examine how the likelihood of each preemption style changes compared to one another.
Results
Preempting vs. non-preempting provisions
Table 5 contains the multinomial logistic regression for adopting one type of preemption over a non-preempting provision. I first turn to the results for the Local Autonomy Hypothesis. The multinomial logit indicates that a one-unit increase in local autonomy increases the odds of a governor issuing a ceiling preemption over a non-preempting provision, significant at the p<0.001 level. Similarly, a one-unit increase in local autonomy increases the odds of a governor issuing a floor preemption, significant at the p<0.05 level. At the highest observed level of local autonomy, governors are most likely to issue ceiling preemptions (53%), followed by vacuum preemptions (16%), non-preempting provisions (16%), and floor preemptions (15%) (see Figure 2).
Table 5. Multinomial logistic regressions, preemption vs. non-preemption
Note: Clustered standard errors in parentheses. Phase-fixed effects are included but are omitted to facilitate presentation.
* p < 0.05.
** p < 0.01.
*** p <0.001.
Figure 2. Predicted probabilities for local autonomy.
Note: Predicted probability of observing different preemption styles across different levels of local autonomy. Outer lines indicate 95% confidence intervals.
Therefore, I find support for the Local Autonomy Hypothesis: governors in states with higher local autonomy are significantly more likely to issue ceiling preemptions and floor preemptions. These results bring back into question whether local governments are truly protected by institutions such as Home Rule.
Turning to the results for the Republican Governor Hypothesis , the multinomial logit model indicates that having a Republican governor increases the odds of a governor issuing a ceiling preemption over a non-preempting provision, significant at the p<0.01 level. On the other hand, Republican governors are significantly less likely to issue a vacuum preemption over a non-preempting provision, significant at the p<0.01 level. Republican governors are most likely to issue ceiling preemptions (45%), followed by non-preempting provisions (31%), floor preemptions (14%), and vacuum preemptions (9%) (see Figure 3). Therefore, I find support for the Republican Governor Hypothesis : Republican governors are significantly more likely to issue ceiling preemptions. The multinomial logit model does not indicate that Democrat governors will issue a floor preemption over a non-preempting provision. Therefore, I do not find support for the Democrat Governor Hypothesis.
Figure 3. Predicted probabilities for gubernatorial partisanship.
Note: Predicted probability of observing different preemption styles under a Democrat vs. Republican governor. Outer lines indicate 95% confidence intervals.
The results for the Republican Governor Hypothesis are consistent with SoRelle and Walker’s (Reference SoRelle and Walker2016) study of the US Congress, which found that Republicans tend to cap regulations by utilizing ceilings that curtail localities’ ability to regulate. These preemption trends had observable implications during the pandemic. Take, for example, a Republican governor who issues a ceiling preemption, banning counties from issuing stricter mask mandates than their own. In counties with high levels of COVID-19 cases, this preemption can prevent local governments from meeting the public health needs in their area.
Turning to the results for the Ideological Asymmetry Hypothesis , the multinomial logit model indicates that a one-unit increase in ideological asymmetry decreases the odds of a governor issuing a ceiling preemption over a non-preempting provision, significant at the p<0.05 level. A one-unit increase in ideological asymmetry also decreases the odds of a governor issuing a floor preemption over a non-preempting provision, significant at the p<0.05 level. On the other hand, a one-unit increase in ideological asymmetry increases the odds of a governor issuing a vacuum preemption over a non-preempting provision, significant at the p<0.01 level. At the highest observed level of ideological asymmetry, governors are most likely to issue vacuum preemptions (52%), followed by non-preempting provisions (30%), ceiling preemptions (13%), and floor preemptions (6%) (see Figure 4). Therefore, I do not find support for the Ideological Asymmetry Hypothesis: greater ideological distance between a state and its localities is associated with decreased ceiling preemptions. However, the findings for Ideological Asymmetry still prove interesting: governors prohibit ideologically dissimilar local governments from regulating policy but also do not set a statewide standard.
Figure 4. Predicted probabilities for ideological asymmetry.
Note: Predicted probability of observing different preemption styles across different levels of ideological asymmetry. Outer lines indicate 95% confidence intervals.
Results for the control variables also prove interesting. Stronger governors are likelier to issue a ceiling preemption over a non-preempting provision. Lame-duck governors are likelier to issue a vacuum preemption over a non-preempting provision. Governors in states with a higher vote share for President Trump are less likely to issue a ceiling preemption over a non-preempting provision. A state’s COVID-19 cases, unemployment rate, and population size do not significantly influence gubernatorial preemption style.
Dropping non-preempting provisions
The first model in this paper explores what influences a governor’s decision to issue a preempting provision over a non-preempting provision. The first model addresses most states’ circumstances, given that most governors issue preempting and non-preempting provisions. However, some governors primarily interact with local governments through preemption, rarely empowering them (Weissert et al. Reference Weissert, Uttermark, Mackie and Artiles2021). Therefore, the second model in this paper examines what influences a governor’s decision to issue one type of preemption over another.
Table 6 contains the multinomial logistic regression for adopting one type of preemption over another. This model drops non-preempting provisions from the analysis and explores the likelihood of a governor issuing a ceiling or floor preemption over a vacuum preemption. Local autonomy does not significantly influence a governor’s decision to adopt a ceiling or floor preemption over a vacuum preemption. Republican governors are significantly more likely to issue a ceiling preemption over a vacuum preemption, significant at the p<0.001 level. Interestingly, Republican governors are likelier to issue a floor preemption over a vacuum preemption, significant at the p<0.05 level. Republican governors are most likely to issue ceiling preemptions (66%), followed by floor preemptions (20%) and vacuum preemptions (14%).
Table 6. Multinomial logistic regression, preemption vs. preemption
Note: Clustered standard errors in parentheses. Phase-fixed effects are included but are omitted to facilitate presentation.
* p < 0.05.
** p < 0.01.
*** p < 0.001.
A one-unit increase in ideological asymmetry decreases the odds of a governor issuing ceiling and floor preemptions over a vacuum preemption, significant at the p<0.001 level. At the highest observed level of ideological asymmetry, governors are most likely to issue vacuum preemptions (74%), followed by ceiling preemptions (17%) and floor preemptions (9%).
Conclusions and discussion
This paper offers a first look at governors’ strategies to preempt local governments. Extending Weissert et al.’s (Reference Weissert, Uttermark, Mackie and Artiles2021) dataset and utilizing Wagner et al.’s (Reference Wagner, Davidson, Haddow, Jones, McFarland and Rainwater2019) novel typology, this paper finds that a combination of institutional and political factors influence the different preemption styles used by governors during the COVID-19 pandemic.
Governors in states with high local autonomy are more likely to issue vacuum preemptions. Republican governors are more likely to issue ceiling preemptions, placing a cap on the regulations that local governments can put in place. Governors in states with ideologically distant local governments are more likely to issue vacuum preemptions. When non-preempting previsions are dropped from the analysis, local autonomy does not significantly affect issuing one type of preemption over another. On the other hand, Republican governors are more likely to issue both ceiling and floor preemptions over vacuum preemptions. Governors in states with high ideological asymmetry are less likely to issue ceiling and floor preemptions over vacuum preemptions.
The results of this paper support the literature that finds increased state preemption activity over high-autonomy local governments. For example, Weissert et al. (Reference Weissert, Uttermark, Mackie and Artiles2021) found that governors facing high-autonomy local governments tend to issue more preemptions. These results further this finding, showing governors’ specific strategies to circumvent high-autonomy local governments. Moreover, the results comport with the literature that illustrates the impact of partisanship on preemption activity. Namely, this paper extends SoRelle and Walker’s (Reference SoRelle and Walker2016) study on legislative preemption style, which found that Republican lawmakers are more likely to utilize ceiling preemptions in the case of governors. The results also mirror the literature showing that ideological asymmetry increases state preemption activity. Specifically, this paper matches Barber and Dyne’s (Reference Barber and Dynes2021) findings on legislative preemption and Swanson and Barrilleaux’s (Reference Swanson and Barrilleaux2020) findings on judicial preemption, confirming that ideology also influences gubernatorial preemption.
The contribution of this paper is twofold. First, it offers insight into the understudied area of gubernatorial behavior. Governors are influential figures who can unilaterally preempt local action. Therefore, understanding whether and how they choose to do so is vital. For example, if we know that governors respond to high-autonomy local governments with more restrictive preemptions, we must examine how to empower these local governments despite these preemptions.
Second, this paper demonstrates that it is necessary to differentiate between preemption in general versus the different styles of preemption. Many studies examine what institutional and political factors influence state preemption activity; however, the preemption strategies used have been long overlooked. Preemption strategies range in severity and restrictiveness; ceiling and floor preemptions limit local governments, while vacuum preemptions tie local governments’ hands altogether. It is, therefore, essential to understand what drives states’ decisions to utilize each.
These findings provide several avenues for future scholarship on intergovernmental relations. Future work should explore the determinants of state preemption style in contexts beyond the COVID-19 pandemic. COVID-19 was an unprecedented national emergency requiring quick and prolonged action among governors. It will be interesting to see whether, for example, legislators or judges pursue similar preemption styles outside of emergencies as governors did during emergencies. Using this study as a foundation, other scholars can evaluate the generalizability of these findings beyond the COVID-19 pandemic.
Furthermore, scholars should evaluate how local government structure affects the amount and type of preemption used by governors. Weissert et al. (Reference Weissert, Uttermark, Mackie and Artiles2021) found that local autonomy increased preemption rates. This paper adds to their results, finding that the most restrictive type of preemption was leveraged against high-autonomy local governments. However, what specifically about high-autonomy local governments drives governors to preempt them so severely remains unknown. Future work should unravel which fiscal, political, and administrative variables within high-autonomy local governments lead to increased preemption.
Finally, future work should examine how local officials respond to the different types of state preemption. We know that local officials react to state preemption in various ways, such as filing a lawsuit, refusing to comply, or initiating local legislation (Swindell, Svara, and Stenberg Reference Swindell, Svara and Stenberg2018). However, we do not know which types of preemption influence the most local resistance. By gauging local officials’ reactions to state preemption, we can learn more about the consequences of state preemption activity, which continues to rise rapidly.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/spq.2024.12.
Data availability statement
Replication materials are available on SPPQ Dataverse at https://doi.org/10.15139/S3/QQTASK (Artiles Reference Artiles2024).
Funding statement
The authors received no financial support for the research, authorship, and/or publication of this article.
Competing interest
The author declares none.
Author biography
Dr. Alexandra Artiles is an Assistant Professor of Political Science at Boise State University and former Fulbright Scholar at the University of Ottawa. Her research interests include public policy, state politics, and gender.
