Building a Safe State

Abstract

Although existing studies uncovered how controversial policies are adopted across American States, little is known as to how regulatory policies are diffused that are technically complex and yet have low political salience. Drawing on existing theories of policy adoption and diffusion, this article identifies the determinants of building code mandates in U.S. state governments. Using original data from state legislatures and administrative agencies from 1980 to 2006, this article finds that both emulation from neighboring states and innovation within the state government increase the likelihood of statewide building regulations. Code adoption follows a hybrid diffusion pattern, whereby external and internal mechanisms occur simultaneously rather than independently from each other.

Keywords

building codes policy diffusion public employment event history analysis

Existing studies have uncovered both external and internal mechanisms through which policies are diffused and adopted across state and local governments. Beginning with Walker (1969), scholars have amassed evidence on why and how states adopt new policies, and F. Berry and Berry (1990) pioneered the area by demonstrating that external and internal attributes are not mutually exclusive, but both can affect the likelihood policy adoption. Since their study on state lottery programs, various policy areas have been examined including taxation (F. Berry & Berry, 1992), abortion (Mooney & Lee, 1995), same-sex marriage (Hume, 2011), and public health (Pacheco, 2012; Shipan & Volden, 2008). However, despite the wealth of evidence on these “highly salient and technically simple” policies (Mooney & Lee, 1995, p. 618), it remains unclear whether the same logic applies to a large body of policies that are technically complex and yet low in political salience.¹ In 2011, about 40,000 bills and resolutions took effect in 505 states. Among them, controversial issues were included such as civil unions and health care, but a large portion of these bills are nonpolitical policies such as distracted driving, animals and wildlife, and consumer protection (National Conference of State Legislatures, 2012). These regulatory policies rarely involve intense popular debates, but their implementation has broad impacts on the everyday life of state residents.

Although it is known that low salience and high complexity policies are diffused more gradually than rapidly (Nicholson-Crotty, 2009), the conditions under which these policies diffuse with such a pace remain elusive. In light of this oversight, I develop testable hypotheses around external and internal mechanisms and examine each of them using a unique data set on building code adoption across U.S. state governments from 1980 to 2006.² Instead of treating external and internal mechanisms as separate and independent processes, this study argues that both pathways are complementary to and interactive with each other. Due to the broad implication of building codes in such areas as housing market, public health and construction industry, agents of code adoption may emulate neighboring states’ policies as well as consult with multiple departments within a state government. Therefore, leveraging one mechanism over the other may not capture a full picture of building code diffusion.

As climate change and energy efficiency become the central initiatives of the federal government, building regulation is considered one of the most efficient ways to achieve energy efficiency (Aroonruengsawat, Auffhammer, & Sanstad, 2012). Increasing attention is being paid to statewide variations in the degree of building regulations, as reported in New York Times, for example (Krauss, 2009). As such, there is an ongoing debate on the extent to which the federal government should preempt the state government laws in such areas as building codes and environmental protection (Klass, 2010). For this matter, the answer remains inconclusive: President Obama, for example, acknowledged that state governments have provided more aggressive protection than the federal government and thus cautioned against the federal preemption (Executive Office of the President, 2009), while the Bush administration upheld a stronger role of the federal government over state government laws. However, to date, there is no systematic analysis on why some states are more active in adopting, updating, and enforcing stricter building codes while others lack the incentive or capacity to do so.

This study also fills a lacuna in building regulations studies by identifying the origins of code adoption rather than variations in enforcement after the adoption. While there exists a wealth of research on how codes are enforced by administrative agencies (e.g., Van der Heijden, 2010, 2013), we know relatively less about what prompts state government’s decision to legislate uniform building standards. For example, Van der Heijden (2010) showed that private sectors are highly involved in the regulation practices in countries such as Australia and Canada and yet provided little information on the motivations of code adoption in the first place. One of the reasons could be that these countries have a more centralized government than the United States, generating little variation at the adoption stage. Diffusion patterns under the federalist system, such as one in the United States, offers an ideal setting as state governments may choose the timing and extent of regulations without the federal government mandate. As such, I consider external and internal mechanisms in the administrative, executive, legislative, and electoral domains that lead to a state’s decision to pass a statewide building code.

Patterns of Building Code Adoption

In the latter half of the 20th century, state governments began undertaking efforts to unify, adopt, and enforce statewide standards of building construction and maintenance. Initially, the movement was regional. Three model codes were developed and adopted by state and local governments: In the East Coast and Northwest, the National Building Code was published by the Building Officials Code Administrators International (BOCA). In the South, the Southern Building Code Congress International created Standard Building Code (SBC). Finally, Uniform Building Code (UBC) published by the International Conference of Building Officials covered the West Coast.

Other than having regional headquarters, these three codes did not differ much in terms of their contents and were very similar in nature. Consequently, a growing awareness emerged among building professionals on developing a nationwide state building code to facilitate smooth adjustment across geographic boundaries. International Code Council was created in 1994 to provide International Building Code (IBC) that does not have regional limitations (McFann, 2012).³ The convergence from three regional codes (BOCA, SBC, and UBC) to the uniform national standard (IBC) occurred voluntarily, as the three regional code councils agreed to suspend the development of their own individual codes and collectively promote IBC. This process of regional to national consolidation suggests that spatial diffusion through neighboring states may have occurred in the building code case.

Although regional movements on uniform codes were present, all states in the same region did not adopt building codes simultaneously. In terms of the speed of diffusion, the spatial and temporal patterns corroborate Nicholson-Crotty’s (2009) argument that low salience and high complexity policies tend not to diffuse rapidly but rather be adopted in an incremental fashion. The spatial pattern shows that a number of states in the West and the Southeast adopted building codes between 1970 and 1980, and it spread to the Northeast and the South in the 1990s and 2000s (Figure 1).

Figure 1.

Spatial diffusion of building codes by decade, 1970-2000.

In addition to the regional and later national movement, innovation within state boundaries also occurred. In each state, different government agencies led the code adoption process with various stringency and applicability to municipal governments, implying that internal conditions may have also played a role in addition to the influence of nearby states (International Code Council, n.d.). The following section elaborates on the theoretical discussions on both external and internal mechanisms.

Theoretical Framework: Mechanisms of Policy Adoption

Spatial Diffusion

Past studies have taught us that policy diffusion processes occur through both external and internal mechanisms. The external mechanism focuses on whether neighboring states have adopted the policy or not. Beginning with Walker’s (1969) insight on the diffusion of innovation, scholars have found substantial evidence on the diffusive nature of policy adoption across geographic boundaries. Specifically, Shipan and Volden (2008) listed four ways of spatial diffusion: learning, competition, imitation, and coercion. Subsequent studies found evidence of this spillover effect in many policy areas as state lottery (Berry & Berry, 1990), antismoking ban (Pacheco, 2012; Shipan & Volden, 2008), same-sex marriage (Hume, 2011), and abortion regulation (Mooney & Lee, 1995). These mechanisms commonly assume that states are influenced by what neighbors do, especially by those that share borders. Because contiguous states share similar political, social, and economic characteristics, it is more likely that states look to what neighbors do and follow them. Furthermore, gradual learning is all the more probable for low salience policies, as adoption of these policies is less motivated by political needs and therefore does not have to occur rapidly (Nicholson-Crotty, 2009). Based on the existing literature and the patterns of diffusion in code mandates, we can hypothesize that building codes may follow the gradual learning pattern identified by Nicholson-Crotty (2009):

Hypothesis 1: Spatial Diffusion Hypothesis. When a state’s neighbors adopt building codes, the probability of building code adoption will increase.

State Government Resources

Along with the external influence, I also consider state’s internal attributes. First, the growth of state government, in human and financial capacity, may facilitate regulatory policy adoption. To the extent that policy innovation is equivalent to the innovation in the private sector and state bureaucrats to entrepreneurs, we can extrapolate the Schumpeterian hypothesis on firm size and innovation to the growth of state government and policy innovation. According to Schumpeter (1942), large firms stimulate innovation, as these firms enjoy the economy of scale, invest more in research and development, and therefore are capable of promoting entrepreneurial activities. Walker (1969) makes a similar argument about the economy of scale, arguing that large, wealthy and industrial states tend to innovate faster than small, poor, and rural states. Later studies also found empirical support for Walker’s hypothesis. Clark (1985), for example, argues that while symbolic policies do not require significant resource commitments, complex policies may need slack resources because their implementation involves new administrative infrastructures. Tolbert, Mossberger, and McNeal (2008) also found that state wealth and education are significant predictors of e-government innovation across states. These studies commonly find that when states have a higher median income, more urban and educated population, there are ample opportunities to build administrative infrastructures throughout the state.

The accumulation of human (personnel) and physical (financial) capital in state government can contribute to policy adoption in two ways. First, parallel to the Schumpeterian hypothesis, there is a greater chance of policy innovation when the administrative agency grows in size. In other words, government agencies can serve as a laboratory of innovation as a greater number of bureaucrats are involved in public safety and construction, as they can expect rewards from expanding their expertise such as prestige, increased budget, and greater power within the bureaucracy. Existing studies on building codes also emphasize the role of government officials in promulgating building regulations. For example, May (1997) argues that state officials’ willingness and capacity to develop and execute regulatory programs are crucial to determining the level of intervention in building construction. Oster and Quigley (1977) also found evidence that the education level of building officials is the most important factor that facilitates building regulations at the municipal level. Second, increased capacity of state government also means the capacity to counter resistance from the building sector. Often, private actors in the building industry are concerned with the increasing cost of construction at the onset of more regulations (May, 2005). Delays in permit approval, inconsistencies in regulation and inspection, and corruption all add to the heightened cost of housing construction and development. As a solution, a series of innovations in the regulatory and administrative processes are suggested (May, 2005): A facilitative permit process, integration of regulatory functions, and conflict reduction are some examples. State government should have sufficient capacity and resources to minimize resistance and prepare for postadoption enforcement strategies.

To the extent that these state bureaucrats are policy experts, they use their expertise to establish and promulgate new regulations. Callander (2008) argues that policy expertise is not just a possession of information, but also knowledge of the “process by which policy instruments are transformed into outcomes” (p. 123). In other words, policy experts usually have knowledge not only about the policy itself but also about the adoption process and policy consequences. Furthermore, a legislative body is more willing to delegate to the agency when a policy is complex (Elgie & McMenamin, 2005), and state bureaucrats can have substantial autonomy to determine both the contents of the policy and the timing of policy adoption and implementation.

Furthermore, the process of regulatory policy adoption enables administrators to exert influence in the legislation process. In addition to the standard legislative adoption process in which most political issues are discussed and passed, there exists a regulatory pathway through which state government agencies can exert influence on policy adoption and implementation (U.S. Department of Energy, n.d.). Compared with a common process where a voting body should approve a proposal, regulatory action bypasses the legislative process because the legislature delegates its power to the executive agencies of state government. Instead, state government may appoint a regulatory body that reviews the code provisions. Then, the code goes through a public review process from where comments are incorporated, and the approval authority conducts a final review, and it is filed to become law. Therefore, building regulations are less likely to emerge as a politically salience issue not only because of its substantive technicality but also because of the procedure that does not involve a legislative process. Based on this reasoning, I derive the second hypothesis:

Hypothesis 2a: Public Employment Hypothesis. As the size of state government employees related to building sector increases, the probability of building code adoption also becomes higher.

The second reason why government size may matter lies in its financial capacity. In addition to innovation by policy entrepreneurs, state government size is also relevant with satisfying citizen demands. A great portion of regulatory policies derives from citizens who demand better and more extensive public service provision. Increasing economic and social activities produce externalities that require government intervention, and state government responds to these growing demands by creating regulatory institutions. As such, regulatory policies such as building codes or speed limits require a sufficient amount of resources for enforcement, and only those governments that have reasonable means to meet these demands are willing to propose and adopt policies. Therefore, the next hypothesis reflects this fiscal dimension:

Hypothesis 2b: Public Finance Hypothesis. As the fiscal magnitude of state government increases, the probability of building code adoption also becomes higher.

With regard to the state government capacity hypotheses, an inevitable question is the direction of causality (May, 1997). One might question whether state government capacity has enabled policy adoption, or the adopted policy stimulated state capacity. Although both claims are logically possible, there is a reason to doubt the latter proposition. The growth of state and local government has been a nationwide phenomenon since the World War II, and it is difficult to posit that this national trend is a sole function of regulatory policies, let alone building code. For instance, Garand (1988) found that state government growth is a function of “bureau voting” power and the increase of federal grant. As public employees tend to vote more than ordinary citizens, their political preferences (i.e., greater government spending) are more likely to be realized than the public preference for an optimal government size. Furthermore, he found support for the claim that the amount of transfers from the federal government increased over time and contributed to the growth of the public sector. In this sense, regulatory policy adoption is a result of the increased government spending and personnel rather than the cause thereof.

State Political Conditions

In this section, I consider three political conditions that may prompt a state’s decision to establish regulatory mandates in the building sector. I begin with the assumption that the primary initiator of building regulations is bureaucrats rather than politicians, and then discuss when bureaucratic agents may attempt to introduce the bill to politicians. In particular, I focus on administrative and electoral fluctuations as important political conditions to maximize a chance for regulatory policy adoption. Furthermore, state’s latent ideological characteristics may also facilitate or hinder the likelihood of adopting building regulations.

To legislate regulatory policies, bureaucrats have to deal with state politicians. In this case, state officials will prefer an environment with greater political uncertainty because politicians are more likely to delegate to nonpolitical agencies when the level of political uncertainty is high (Elgie & McMenamin, 2005). Under a situation with high political uncertainty, incumbent politicians tend to delegate more independence to agencies that deal with complex issues rather than insulating bureaucracy. As a result, bureaucrats are less constrained in their ability to propose a policy, persuade the legislature, and put them to vote when the state government goes through an administrative change.

A period when political uncertainty is high when there is an administrative change, especially with regard to a new governor. Administrative change encourages an adoption of new management innovations, because a newly elected politician often wants to promote policy innovation to show her leadership in the beginning of her administration (F. Berry, 1994). Policymakers will also prefer a new governor who is less knowledgeable about the policymaking process and yet more enthusiastic and open-minded about new policies. Berry found that policy innovation is more likely to occur in the first year of a new governor’s office, as policy innovation (strategic planning in her study) can serve as a symbol of new political leadership. Newly elected governors are also more willing to take risks in their first year in office than later in their tenure. Therefore, I construct the following hypothesis:

Hypothesis 3a: Administrative Change Hypothesis. The probability of building code adoption will be higher in the first year of new gubernatorial administration.

Compared with the administrative change, state elections have mixed implications for building code adoption. F. Berry and Berry’s (1992) political opportunity argument states that politicians use policy innovation for their electoral benefits, by adopting popular policies and delaying unpopular and controversial ones. Politicians tend to promote policies that will maximize their chance of winning an election, and economic development policies are likely to be adopted in election years. This argument originates from a political budget cycle theory (Nordhaus, 1975; Rogoff, 1990), which argues that politicians may contract or expand economic programs according to election cycles. Politicians are likely to expand their tax revenue and expenditure in election years and a year prior to the election year, in the hope that greater spending will translate into more votes. Just as politicians manipulate the fiscal budget, they can also use policy innovation for their electoral benefits. For instance, F. Berry and Berry (1990) found that state lottery program is more likely to be adopted during election years because of its perceived popularity among voters. Politicians tend to promote policies that will maximize their chance of winning an election, and therefore revenue generating policies are likely to be adopted in election years.

Conversely, controversial or unpopular policies are legislated once elections are over. Politicians are more careful with socially controversial policies during elections. For instance, Mooney and Lee (1995) found that more regulations on abortion were adopted when fewer elections occur in a given year. For these controversial policies, elections constrain innovation because politicians do not want to take extra risks by initiating a controversial policy. Another study found that tax policy is less likely to be adopted in election years, as taxation is generally not very appealing to voters (F. Berry & Berry, 1992). Voters generally do not support any type of tax increase, and therefore politicians find political opportunities to adopt tax increase policies in off-election years.

In terms of building regulations, candidates may not view them as their primary policy device for winning an election. In light of those low salience policies, Nicholson-Crotty (2009) emphasizes the lack of interest for politicians in adopting regulatory policies that are politically invisible. As it is difficult for politicians to calculate the predicted benefits and costs of politically neutral policies such as building codes, they may simply delay or shelve these low salience policies during election years. Therefore, an inverse relationship may exist with regard to electoral cycle and code adoption, yielding the following hypothesis:

Hypothesis 3: Electoral Cycle Hypothesis. The probability of building code adoption will be lower during state gubernatorial and legislative election years.

Finally, past studies have emphasized the role of state political ideology on policy adoption (e.g., Grossback, Nicholson-Crotty, & Peterson, 2004). Economically redistributive and socially progressive policies are more likely to be adopted by government with the Democratic majority, whereas conservative government tends to support policies that promote private sector growth and minimize government intervention. As a regulatory barrier, building codes generally adds construction costs and are believed to depress the housing market (Burby, Malizia, & May, 1999), and the construction sector tends to oppose additional layer of government regulation. In this sense, Democratic government will be more willing to adopt regulatory policies, as these policies can strengthen government authority to oversee the private sector.

Hypothesis 3c: Political Climate Hypothesis. The probability of building code adoption will be higher under the Democratic government.

Hybrid Diffusion

Finally, I hypothesize that external and internal mechanisms of diffusion are mutually reinforcing. Agents of code adoption do not only look at neighboring states under similar conditions but also consider internal attributes to gauge the possibility of code adoption. Officials may look to the successful case of code adoption and implementation in their region. However, emulation will only be possible when a state has sufficient manpower or budget to carry out code regulation. Likewise, when a state possesses internal political and economic capacity, learning from adjacent states is still important to ensure successful implementation.

Hypothesis 4: Hybrid Diffusion Hypothesis. Interstate diffusion and intrastate innovation are more likely to occur simultaneously rather than occurring independently from each other.

Event History Analysis (EHA) of Building Code Adoption

To test these hypotheses, I conduct discrete time EHA of pooled cross-sectional data of building codes from 1980 to 2006. Pioneered by Berry and Berry (1990), EHA is the most common approach to analyzing policy diffusion. I fit the EHA model with generalized linear model with a complementary log log link function (cloglog). Cloglog produces similar results with logit function, and cloglog is used in this analysis since code adoption is a rare event (Carter & Signorino, 2010). This takes the following form:

\Pr (Y_{i, t} = 1) = 1 - \exp (- \exp (x_{i, t} β + γ_{1} t_{i} + γ_{2} t_{i}^{2} + γ_{3} t_{i}^{3})),

where the probability (or hazard) of building code adoption in state i at year t( $Y_{i, t}$ ) is estimated with covariates as well as time-dependent variables $γ_{1} t_{i}, γ_{2} t_{i}^{2}, and γ_{3} t_{i}^{3} .$ The following section explains the dependent and independent variables used in the model.

Data and Variables

Dependent Variable

To examine when states formally adopted building codes, I created a variable based on a novel data set that traces the history of building codes adoption in each state government (bc_adopt). The data were retrieved from state government websites and legislative records. It is coded 1 if a state mandates all municipalities adopt the statewide building code in a given year, and 0 otherwise. Often, states publish a code and “encourage” municipalities to adopt the code, and in this study, this informal recommendation is not considered as full-fledged adoption and thus coded 0. Furthermore, if a state did not adopt any code by 2006, the value for each year is 0. I included all states that had adopted regional codes prior to 1997 as well as those that adopted IBC afterward.

Table 1 presents years that each state adopted a statewide building code. Following previous studies (e.g., F. Berry & Berry, 1990), Alaska and Hawaii are excluded from analysis. Alabama and Arkansas adopted the first statewide building codes in the 1950s, and Connecticut in 1969. These states are truncated from the data set, which leaves a total of 45 states in the analysis.

Table 1.

Code Adoption Year by State.

State	Year	State	Year
AR	1955	NE	1987
AL	1957	NV	1978
AZ	—	NH	2002
CA	1978	NJ	1975
CO	1970	NM	1983
CT	1969	NY	1981
DE	1953	NC	1973
FL	1998	ND	1979
GA	1981	OH	1981
ID	1975	OK	2009
IL	—	OR	1974
IN	1998	PA	2003
IA	1972	RI	1976
KS	—	SC	1986
KY	1980	SD	1978
LA	2005	TN	1974
ME	2010	TX	2001
MD	1993	UT	1988
MA	1974	VA	1973
MI	1972	VT	1972
MN	1977	WA	1985
MS	—	WV	1988
MO	—	WI	1980
MT	1972	WY	1977

Source. Compiled by Author.

Note. “—” denotes that no code is adopted as of 2006.

Independent Variables

Spatial diffusion

To test the spatial diffusion hypothesis (i.e., Hypothesis 1), I calculate the proportion of those states that had previously adopted statewide building codes among the total number of contiguous states (nbors_adopt). Consistent with other studies of policy diffusion (e.g., Boehmke, 2009), I use F. Berry and Berry (1990)’s index of contiguous neighbors for each state. Given that statewide building code began as a regional movement, considering the influence of adjacent states reflects the historical patterns of regionalized code adoption discussed above.

State public employment and finance

To measure public employment, I calculate the proportion of state government employees in the sectors related to buildings and safety relative to two general employment statistics: one in overall state employment (building_total_ratio), and the other in state government⁴ (building_employee), both lagged by 1 year (Census Bureau). The former measures the relative size of the building-related public employment in the general state labor market, whereas the latter captures the proportion within state government.^5,6

Given that the total number of state government employees has been steadily increasing (Figure 2), one might argue that the relative size of the public workforce has also increased linearly over time, only to increase the chance of policy adoption as time went on. However, the temporal trend of public employment vis-à-vis the entire state employment does not expand over time. Figure 3 demonstrates the proportion of code-related public employment in select states (Mississippi, Los Angeles, Maine, Kansas, Arizona, Illinois). There is no increasing amount of public employment in these states; rather, these states have experienced a slight decline in their share of building-related employment in state government. For example, Maine’s ratio fell gradually from 1.2% in 1981 to below 1% in 2006. Arizona’s public employment share had decreased until the late 1980s, increased again until the mid-1990s, and then steadily declined afterward. Illinois, however, had a consistent share (0.5%) throughout the period, with no further fluctuation. Although the divergent patterns of state public employment is out of the scope of this article, it is clear that building-related public employment, however small the size may be, follows a dynamic path.

Figure 2.

Number of state government employees by year: 1947-2010.

Figure 3.

Percentage of public building employees, select states: 1980-2006.

Second, to capture the magnitude of state government’s financial contribution to code-related industries, I use an indicator that measures the size of this expenditure relative to the entire state economy (building_size). Building size is the percentage of building-related expenditures (construction, transportation, natural resources, parks and recreation, and sanitation and sewerage) out of total state economic output. These categories are equivalent with what I included in employment statistics so that both measures point to state government’s involvement with construction and maintenance of buildings. As Garand (1988) suggests, there is no systematic measure available for total economic output, and I follow his personal disposable income for the measure of state economic output.

Building expenditure size = \frac{total building expenditure}{total disposable income} \times 100 .

As shown in the public sector employment patterns, no uniform pattern is found across states in the percentage of state government spending (Figure 4). If any, the proportion of government spending vis-à-vis the overall economy has been fluctuating over the years, with greater than 4% in states like Mississippi and Louisiana and about 2% in Illinois. This nonpattern once again shows that expenditure trend is not time-dependent.

Figure 4.

Percentage of public building expenditure, select states: 1980-2006.

Although these observational measures are imperfect to gauge the exact capacity of state government to initiate and implement statewide building code, there is a trade-off. Despite policy experts’ potential contribution to policy adoption, the difficulty of conducting a comparable nationwide survey of state bureaucrats has plagued analyses of policy experts in the policy adoption process (Karch, 2007). As a result, existing studies usually capture the quality of bureaucracy at a particular moment rather than tracking a temporal trend (e.g., Oster & Quigley, 1977). In this respect, measuring human and fiscal capacity through employment and spending patterns can provide valuable insight on the relative strength of state government to influence building code adoption.

Administrative and electoral changes

Finally, I test the third set of hypotheses (i.e., Hypotheses 3a, 3b, and 3c) using data sets on state partisan balance and governors.⁷ I created dummy variables for administration change (new_governor) and electoral seasons (gubernatorial and state legislative elections).⁸ In other words, it is coded 1 when a new governor is in office, or there is a gubernatorial/state legislative election. The overall political ideology of a state is measured with state NOMINATE score (gov_nom), which measures state legislator’s ideal points to infer state government ideology (Berry et al., 2010). NOMINATE scores are an updated and improved measure from the W. Berry Ringquist, Fording, and Hanson’s (1998) state ideology indicator based on interest-group ratings. Governor’s partisanship is measured in a variable govparty, where 0 is Republican, 1 Democrat, and 0.5 if neither. To measure which party controls the state legislature, two variables are used: one for House (house), and the other for Senate (senate). Both are coded 0 if controlled by Republicans, 1 if Democrats, and 0.5 if the chamber is split.

Control Variables

Private sector demand and natural disasters

To control for the effect of the private sector economy, I include per capita housing units (percapitahousing) and construction sector size (construction) as measures of construction industry in a state. Per capita housing units is measured with the total number of housing units built in a given year divided by annual population, and the data were collected from Census Building Permits Survey.⁹ Construction sector size is calculated as the proportion of employees in the construction sector out of total state employment, lagged by 1 year (Bureau of Economic Analysis). State’s housing market can be a crucial component of adopting building regulations because as the construction market expands, there are higher chances of producing lower quality buildings and therefore greater citizen demands for strict regulation. Construction industry is a countervailing force to such demands, as constructors and developers tend to oppose a layer of regulations for fear of increased construction cost. Furthermore, scholars have also found that strong code depresses housing market in urban central city areas and motivates developers’ exist to less regulated suburbs (Burby et al., 1999).

Second, I calculate the annual frequency of FEMA disaster declarations in each state (disaster), lagged by 1 year. Natural disaster is one of the critical reasons for ensuring building safety and standards (Czajkowski & Simmons, 2013), and it is possible that states exposed to various kinds of natural hazards, such as hurricane, tornado, and wildfire, may experience a push for stricter building regulations.

Temporal variation

Finally, I take into account time dependence in the data set. In analyzing binary cross-sectional data, duration dependence can occur because the probability of an event occurring will not only depend on covariates but also with elapsed time. Overlooking the duration dependence leads to omitted variable bias, which yields biased coefficient estimates. Therefore, I address this issue by including three time dummies, t, t², and t³, as suggested by Carter and Signorino (2010). Carter and Signorino have shown that using t, t², and t³ captures any form of hazard rate (linear, curvilinear, etc.) and is a simpler and more efficient alternative to time dummies or splines (p. 34). Table 2 provides descriptive statistics for the entire set of variables.

Table 2.

Descriptive Statistics.

Variable	Observation	M	SD	Minimum	Maximum	Source
Bc_adopt	338	0.0325	0.1777	0	1	Author’s Compilation
Nbors_adopt	338	0.654	0.209	0.2	1	Author’s Compilation
Building_total	338	0.36	0.0026	0.75	2.06	Census Bureau
Building_employee	338	86.23	3.79	60.50	96.24	Census Bureau
Building_size	338	3.047	0.906	1.466	6.005	Bureau of Economic Analysis
New_gov	338	0.172	0.378	0	1	Klarner Politics
Gub_election	338	0.272	0.446	0	1	Klarner Politics
Leg_election	338	0.453	0.498	0	1	Klarner Politics
Gov_ideo	338	48.61	11.11	23.64	70.36	Berry et al. (2010)
Gov_party	338	−0.024	0.989	−1	1	Klarner Politics
Senate	338	0.554	0.492	0	1	Klarner Politics
House	338	0.659	0.473	0	1	Klarner Politics
Percapita_housing	338	0.005	0.0037	0	0.0256	Census Building Permits Survey
Construction	338	0.057	0.01	0.031	0.087	Bureau of Economic Analysis
Disaster	338	17.3	34	0	306	FEMA
Time	338	21.6	7.28	11	36	—
Time_sq	338	519.42	332.11	121	1,296	—
Time_cube	338	13,606.4	12,307.1	1,331	46,656	—

Note. FEMA = Federal Emergency Management Agency.

Findings

Table 3 displays the findings from EHA for four models.¹⁰ The first two models, respectively, test the external and internal mechanisms of policy diffusion. When the external factor alone is considered (Model 1), the model does not fit well with no variables showing statistical significance. Models 2 and 3 estimate internal factors. Model 2 includes the share of building employees in total state employment, whereas Model 3 includes the share of building employees within the state government employment. Among internal factors, building-related expenditure, new governorship, and the Republican Senate are positively associated with code adoption. One percentage increase in building-related expenditure, for example, increases the odds ratio by 3.46 (exp(1.242)), which means that the odds of successful adoption is about 3.5 times greater than nonadoption (0.78 vs. 0.22).

Table 3.

Event History Analysis of External and Internal Mechanisms in Building Code Adoption, 1980-2006.

Variable	Model 1 external diffusion: Neighbors only coefficient (SE)	Model 2 internal diffusion I: State bureaucracy and politics coefficient (SE)	Model 3 internal diffusion II: State bureaucracy and politics coefficient (SE)	Model 4 hybrid diffusion: No controls coefficient (SE)	Model 5 hybrid diffusion: With controls coefficient (SE)
External mechanism
Nbors_adopt	1.040 (1.594)			4.707** (2.142)	6.73** (2.62)
Internal mechanisms
State bureaucracy
Building_total_ratio		−0.314 (1.694)
Building_employee			0.143 (0.147)	0.379** (0.172)	0.459** (0.234)
Building_size		1.242**(0.505)	1.156**(0.487)	1.104** (0.487)	1.889*** (0.669)
Administrative change
New_gov		1.655* (0.914)	1.502* (0.902)	1.632 (1.027)	2.231* (1.165)
Elections
Gub_election		1.203 (0.879)	1.188 (0.901)	0.955 (0.906)	1.506 (1.005)
Leg_election		−0.557 (0.940)	−0.328 (0.938)	−1.030 (1.044)	−0.835 (1.085)
Political climate
Gov_ideo		−0.009 (0.044)	−0.019 (0.044)	−1.030 (1.044)	−0.068 (0.056)
Gov_party		0.175 (0.440)	−0.237 (0.445)	0.630 (0.464)	0.398 (0.521)
Senate		−2.890**(1.155)	−2.836*** (1.160)	−2.776** (1.081)	−4.068*** (1.416)
House		1.548 (1.051)	1.580 (1.079)	2.362** (1.190)	2.988** (1.380)
Control variables
Percapita_housing	−114.2 (119.5)	−385.921 (245.771)	−315.1 (257.9)		−428.715* (254.025)
Construction	−11.57 (35.55)	42.417 (50.767)	20.52 (54.93)		9.941 (54.585)
Disaster	−0.037 (0.025)	−0.056* (0.032)	−0.055 (0.032)*		−0.071** (0.034)
Time	0.085 (1.451)	0.887 (1.797)	1.276 (1.858)	−0.472 (1.762)	0.746 (2.098)
Time_sq	−0.011 (0.066)	−0.052 (0.084)	−0.068 (0.086)	0.021 (0.080)	−0.044 (0.098)
Time_cube	0.000 (0.000)	0.001 (0.001)	0.001 (0.001)	0.000 (0.001)	0.001 (0.001)
Intercept	−2.603 (9.906)	−12.608 (13.333)	−26.88 (20.15)	−39.45* (21.78)	−55.307** (28.022)
n	337	330	330	330	330
AIC	106.03	98.41	97.415	95.127	91.311

Note. AIC = Akaike Information Criterion.

Denotes significance at .10 level.

Denotes significance at .05 level.

***

Denotes significance at .01 level.

Models 4 and 5 include both external and internal variables. When both external and internal mechanisms are considered, the model fit is improved, with smaller Akaike Information Criterion (AIC) values. In other words, including both external and internal variables still improves the model compared with the information lost by adding more parameters. This supports the hybrid diffusion hypothesis (i.e., Hypothesis 4): Compared with the models that test internal and external mechanisms separately, the hybrid models increase the model fit as well as produce more significant coefficients. In the full model, the estimated coefficients for public employment are positive and statistically significant, confirming the public employment hypothesis (i.e., Hypothesis 2a). In particular, the increase in building-related employment within state government, rather than vis-à-vis general state employment, is positively associated with building code adoption. In Model 4, 1% change in building-related public employment increases the log odds ratio of code adoption by 0.459, and the actual odds ratio of code adoption (vs. nonadoption) by 1.58 (exp(0.459)). In other words, the odds of adoption are about 50% greater than nonadoption (1.24 / 2.24 = 0.61 vs. 0.39). Furthermore, Models 4 and 5 also support the public finance hypothesis (i.e., Hypothesis 2b). One percent increase in building-related expenditure leads to the increase of log odds ratio by 1.889, which means that the odds of successful adoption (0.87) is more than six times as great as nonadoption (0.13).

Newly elected governors have a positive impact on the likelihood of code adoption, supporting the administrative change hypothesis (i.e., Hypothesis 3a). When a new governor is in office, the log odds ratio increases by 2.231 and makes code adoption much more likely than other years (0.90 vs. 0.10). If a new governor is in power, the probability of adoption is almost nine times as high as nonadoption. This finding, combined with public employment and finance hypotheses, supports a view that both bureaucratic agents and a new governor may find administrative change to be a good environment for strengthening regulatory provisions. Election timing hypothesis (i.e., Hypothesis 3b) is not supported throughout the analysis. Given that neither gubernatorial nor legislative elections do not affect the adoption timing, building code adoption does not seem to follow a particular election cycle but is facilitated by new leadership.

Among state’s ideological positions, only state legislature is statistically significant, lending partial support for the political climate hypothesis (i.e., Hypothesis 3c). However, the direction of coefficients is the opposite for the Senate and the House, not supporting the idea that building code adoption is more likely under the Democratic control. Instead, the Republican control of the state senate is positively associated with code adoption, whereas the Democratic partisanship of the state house legislature is statistically significant. The minimal effect of electoral timing and the inconsistent direction of the state legislature suggest that building code policy occupies a flexible position along the liberal and conservative political spectrum. Indeed, one can view code regulations as a layer of government-led regulations or a reform measure to overhaul the building industry. To the extent that building code is perceived as a proxy for government intervention or for part of the administrative reform, both Democratic and Republican legislators may support or oppose its passage.

Discussion

What determines a state’s likelihood of adopting a building regulation policy? Findings in this study indicate that external and internal factors are not mutually exclusive, but both mechanisms work simultaneously to facilitate code adoption by state governments. The establishment of regional building code councils and their consolidation into the national institution have suggested that state governments emulate and interact with one another to create more efficient regulatory regimes. This spatial interaction, however, is not the sole driver of policy innovation but is conditional upon each state’s governing capacities, among which state government agencies and new gubernatorial leadership are significant and consistent factors. In this sense, both emulation and innovation occurs in the process of regulatory policy adoption.

The willingness and capacity of state bureaucrats to learn and implement statewide regulations is crucial because the benefits of code adoption are invisible compared with the immediate cost. The ultimate aim of imposing a set of rules on how to build, maintain, and demolish a structure is to ensure public health and safety and prevent damages from external harms, and yet these benefits are hard to quantify. Compared with these long-term benefits, the sunk cost of code adoption is more immediate: State government may establish an interdepartmental commission to create or approve existing codes, build a department and hire personnel to oversee regulatory activities such as permit issuance and code enforcement. This is true even when there is an established uniform code ready for adoption, because the projected cost comes mainly from enforcement. Therefore, code adoption becomes possible when a state government has sufficient resources to ensure effective administration and enforcement in the future. These findings are consistent with the long-standing theory of policy innovation, which emphasizes both motivation and resources as key determinants of policy adoption and diffusion (e.g., F. Berry & Berry, 1990; Mohr, 1969). Neither motivation nor resources alone can stimulate policy innovation, but rather it is the combination of these two conditions that lead to policy adoption.

In state government, motivation and resources originate from a wide range of agencies rather than a single, identifiable department. In adjacent states, disparate agencies are involved in controlling the building regulation process. For example, in New York, the Office of Planning and Development administers statewide building codes, whereas in New Jersey, the Department of Community Affairs is responsible for code adoption and enforcement. Connecticut, on the other hand, has the Department of Administrative Services in charge of building regulations. The diversity of administrative titles and functions mirrors the variety of issues related with building codes, such as public safety, real estate market, energy conservation, and urban planning.

Furthermore, interagency cooperation within state government also propels the code adoption process. The adoption process in Wyoming showcases how multiple agencies are at work in promulgating building codes (Wyoming State Fire Marshall’s Office, 2010). In Wyoming, the State Fire Marshall’s Office was in charge of adopting statewide building codes. Since its creation, the Office expanded to become the Department of Fire Prevention in 1963, which includes subdivisions such as State Electrical Board and State Fire Commission. By 1964, the Department maintained five appointees, namely, State Fire Marshal, Chief Deputy Fire Marshal, Electrical Inspector, Secretary and Clerk. As the Department became bigger in size and more hierarchical in organization, it also conducted various activities, including inspections on state-run facilities and fire investigation. At the same time, the Department of Fire Prevention cooperated with the Wyoming Health Department for inspecting nursing homes and hospitals, and the Department of Public Assistance and Social Services for child care facilities. It was through these activities that agency staff realized the need for statewide building regulation and pushed for the adoption of the UBC. The variety of responsible agencies, interagency merger, and cooperation make it difficult to argue that one particular agency’s budget and personnel will influence building code adoption. Given the diversity and multiplicity of governmental agencies that are involved in code adoption processes, it is reasonable to assume that the sufficient growth of government capacity enables interagency collaboration as well as the transfer of authorities across departments.

In addition to the coexistence of external and internal influences, findings from this study also recall our attention to the dual significance of administrative and political motivations in the regulatory policy adoption. For politically salient policies, state legislators and executives are interested in advancing their political promises and may use policies as a tool to achieve their goals. In this setting, administrators have little room to maneuver as these policies require little specialized knowledge or prearranged discussions on technical details. Regulatory policies are less sensitive to election timing or ideological orientations, and the bureaucratic resources of state government and the gubernatorial turnover constitute important preconditions. To the extent that building codes are viewed as a reform measure rather than a regulatory barrier, governors, especially those newly elected, will be willing to adopt a statewide regulation. Furthermore, as governors serve a broader range of constituents than localized districts, they are incentivized to initiate a uniform, statewide regulation to counter local fragmentation in building regulations. Much of land use and zoning regulations are delegated to local governments, and building code is an arena where state government can establish a uniform set of standards and encourage the compliance of local governments.

Provided that building codes cannot be made overnight and requires a considerable period of time until adoption, a new governor is more likely to take over her predecessor’s initiative rather than creating it anew by herself. Florida’s case exemplifies this thought (Florida Department of Business and Professional Regulation, 1998). After Hurricane Andrew hit the state in 1992, a series of investigations on hurricane damage revealed that many properties were exposed to natural hazards, but their structure was not resilient enough to withstand another disaster. Furthermore, it was also found that locally enforced building codes were often poorly administered and frequently violated. Prior to state code adoption, five different codes had been used by local jurisdictions, and there was a growing need for a consolidated building code.

As a result, Florida Building Codes Study Commission was established in 1996 to evaluate the existing system and recommend ways to improve the code compliance and enforcement. Consisting of 23 members with the chairman appointed by the Governor, the Commission went through intensive research and developed a uniform statewide code, that is, the Florida Building Code. The Code made a number of recommendations, most of which were approved by the legislature as part of the House Bill 4181 in 1998. It was also the year when Jeb Bush ran for governor as a Republican candidate and defeated his Democratic competitor Buddy MacKay. It was a transformative period for Florida politics, as the state would have both a Republican governor and a Republican majority state legislature for the first time since the Reconstruction (Navarro, 1998). The Florida case demonstrates that rather than being a creator of new regulations, a newly elected governor can serve as a facilitator for adopting them. She may be willing to expedite the process and see it pass during her first year of tenure, and the proposals that are waiting to be passed in the legislature may gain momentum during the new gubernatorial takeover.

Conclusion and Policy Implication

Establishing statewide building codes strengthens the state regulatory power in at least two ways. First, state government can intervene in and collaborate with local government agencies to administer and enforce regulation. Because of the different needs and characteristics across localities, the state allows for variation in the scope of enforcement at the local level, as long as it meets the minimum standards. In New York State, for example, the Division of Building Standards and Codes nested within the Department of State oversees the regulation of building practices through local administrations. The Division set up 11 regional offices, which provides technical assistance, administers variances, delivers educational courses, and oversees the enforcement practices of local governments (New York Division of Building Standards and Codes, n.d., para. 9). Therefore, state government can effectively manage local practices by letting local government choose what is best for their town or city among uniform building, fire and energy standards in the state code. Second, state-centered building regulations increase administrative flexibility. Despite the escalated interests in nationwide issues such as natural disaster and energy conservation, the federal government is limited in its authority to force strict regulations upon land and buildings in state and local jurisdictions. Many regulations come in the form of “recommendations” instead of mandates, limiting the scope of the administrative oversight by the federal regulatory agencies. The inter and intrastate learning and innovation not only reduces administrative costs but also empowers state government to focus on each state’s needs rather than complying with uniform standards.

Studying the origins of code adoption in the 20th century has implications for a broad set of building-related policies. Increasingly, state governments seek to promulgate a set of codes for environmentally friendly and cost-effective building practices. Much education is necessary to induce cooperation from the private sector, for these green practices place a considerable burden on the industry by increasing the cost of construction. Findings from this study suggest that to promote code adoption, state administrators should have a long-term strategy to augment administrative capacity as well as facilitate interstate collaboration, rather than politicizing the issue and relying on political opportunities such as elections. Furthermore, interstate network is useful after a state builds bureaucratic capacity, partly because these policies are less sensitive to political situations and emulating ideas from neighboring states is less consequential than securing sufficient resources to expend throughout the state when adopted.

On a final note, I recognize limitations for the present research. Although local building codes are still prevalent, their impact on statewide code adoption was not fully addressed in the analysis. Part of the reason is that no comprehensive data set exists on which and when local governments adopted the first building code, prohibiting further examination on their relationship with state governments. Fragmented local codes may pressure a state government to adopt a statewide mandate. Or, big cities and their regulatory practices may affect the state’s decision to implement regulation at the state level. The direction is not very clear, however. Would a state like to learn from these local governments and be willing to enforce a statewide code, or would the state government feel it unnecessary to enforce additional standards once these cities have their own building codes? Investigating this local and state relationship will further illuminate the dynamics of regulatory policy adoption. Furthermore, future studies can enrich building regulation studies by investigating determinants of the variation in the stringency of code regulation. Each state has different scope and intensity of code regulation. For example, Florida has strengthened their building codes as the state experienced a series of floods and hurricanes while Louisiana adopted a relatively loose code only recently in 2005. It is possible that compared with the process of nonadoption to adoption, negotiating stringency is more sensitive to election timing and state ideology. Uncovering the mechanisms of this variation will further illuminate our understanding of building regulations in American states.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biography

Min Hee Go is assistant professor of political science at Brooklyn College, City University of New York. Her curent research interests concern representation and participation at the state and local level. Her research has been published in Urban Affairs Review and State and Local Government Review.

References

Aroonruengsawat

Auffhammer

Sanstad

(2012). The impacts of state level building codes on residential electricity consumption. Energy Journal, 33, 31-52.

Berry

(1994). Innovation in public management: The adoption of strategic planning. Public Administration Review, 54, 322-330.

Berry

(1990). State lottery adoptions as policy innovations: An event history analysis. The American Political Science Review, 84, 395-415.

Berry

(1992). Tax innovation by American states: Capitalizing on political opportunity. American Journal of Political Science, 36, 715-742.

Berry

Ringquist

Fording

Hanson

(1998). Measuring citizen and government ideology in the American states, 1960-93. American Journal of Political Science, 42, 327-348.

Berry

Fording

Ringquist

Hanson

Klarner

(2010). Measuring citizen and government ideology in the US States: A re-appraisal. State Politics & Policy Quarterly. 10, 117-135.

Boehmke

(2009). Policy emulation or policy convergence? Potential ambiguities in the dyadic event history approach to state policy emulation. Journal of Politics, 71, 1125-1140.

Burby

Malizia

May

(1999). Beating the building code burden: Code enforcement strategies and Central City success in capturing new housing (University of New Orleans College of Urban and Public Affairs [CUPA] Working Papers, 1991-2000). Retrieved from http://scholarworks.uno.edu/cupa_wp/17

Callander

(2008). A theory of policy expertise. Quarterly Journal of Political Science, 3, 123-140.

10.

Carter

Signorino

(2010). Back to the future: Modeling time dependence in binary data. Political Analysis, 18, 271-292.

11.

Chatterjee

Hadi

(2012). Regression analysis by example (5th ed.). Hoboken, NJ: Wiley.

12.

Clark

(1985). Policy diffusion and program scope: Research directions. Publius: The Journal of Federalism, 15(4), 61-70.

13.

Czajkowski

Simmons

(2013). Convective storm vulnerability: Quantifying the role of effective and well-enforced building codes in minimizing Missouri hail property damage. (University of Pennsylvania Wharton School Working Paper #2013-08). Retrieved from http://opim.wharton.upenn.edu/risk/library/WP2013-08_Building-Codes-in-MO.pdf.

14.

Elgie

McMenamin

(2005). Credible commitment, political uncertainty or policy complexity? Explaining variations in the independence of non-majoritarian institutions in France. British Journal of Political Science, 35, 531-548.

15.

Executive Office of the President. (2009). Preemption. Federal Register, 74(98). Retrieved from http://www.gpo.gov/fdsys/pkg/FR-2009-05-22/pdf/E9-12250.pdf

16.

http://www.fema.gov/media-library-data/20130726-1520-20490-2241/549_ch2.pdf.

17.

Florida Department of Business and Professional Regulation. (1998). History of the Florida building commission. Retrieved from http://www.floridabuilding.org/fbc/information/building_commission.htm

18.

Garand

(1988). Explaining government growth in the U.S. states. The American Political Science Review, 82, 837-849.

19.

Gormley

(1986). Regulatory issue networks in a federal system. Polity, 18, 595-620.

20.

Grossback

Nicholson-Crotty

Peterson

(2004). Ideology and learning in policy diffusion. American Politics Research, 32, 521-545.

21.

Hume

R. J.

(2011). Comparing institutional and policy explanations for the adoption of state constitutional amendments: The case of same-sex marriage. American Politics Research, 39, 1097-1126.

22.

International Code Council. (n.d.). Code adoption process by state. Retrieved from http://www.iccsafe.org/gr/Documents/AdoptionToolkit/HowStatesAdopt_I-Codes.pdf

23.

Karch

(2007). Emerging issues and future directions in state policy diffusion research. State Politics & Policy Quarterly, 7, 54-80.

24.

Klass

(2010). State standards for nationwide products revisited: Federalism, green building codes, and appliance efficiency standards. Harvard Environmental Law Review, 34, 335-368.

25.

Krauss

(2009, July 18). Wide variation in energy codes. The New York Times. Retrieved from http://www.nytimes.com/imagepages/2009/07/18/business/energy-environment/18codesGrfx.ready.html

26.

May

(1997). State regulatory roles: Choices in the regulation of building safety. State and Local Government Review, 29, 70-80.

27.

May

(2005). Compliance motivations: Perspectives of farmers, homebuilders, and marine facilities. Law & Policy, 27, 317-347.

28.

McFann

(2012). History of the building codes. Texas Professional Real Estate Inspectors Association. Retrieved from http://www.tpreia.com/historyofcodes.html

29.

McNichol

(2012). Some basic facts on state and local government workers. Center on Budget and Policy Priorities. Retrieved from http://www.cbpp.org/cms/?fa=view&id=3410

30.

Mohr

L. B.

(1969). Determinants of innovation in organizations. The American Political Science Review, 63, 111-126.

31.

Mooney

Lee

(1995). Legislative morality in the American states: The case of pre-roe abortion regulation reform. American Journal of Political Science, 39, 599-627.

32.

National Conference of State Legislatures. (2012). NCSL news: New laws ring in the new year. Retrieved from http://www.ncsl.org/press-room/new-laws-ring-in-the-new-year.aspx

33.

Navarro

(1998, October 23). In race for governor, democrat is in trouble. The New York Times. Retrieved from http://www.nytimes.com/1998/10/23/us/the-1998-campaign-florida-in-race-for-governor-democrat-is-in-trouble.html

34.

New York Division of Building Standards and Codes. (n.d.) About the BSC. Retrieved from http://www.dos.ny.gov/DCEA/About_DCEA.html

35.

Nicholson-Crotty

(2009). The politics of diffusion: Public policy in the American states. Journal of Politics, 71, 192-205.

36.

Nordhaus

(1975). The political business cycle. Review of Economic Studies, 42, 169-190.

37.

Oster

Quigley

J. M.

(1977). Regulatory barriers to the diffusion of innovation: Some evidence from building codes. The Bell Journal of Economics, 8, 361-377.

38.

Pacheco

(2012). The social contagion model: Exploring the role of public opinion on the diffusion of antismoking legislation across the American states. Journal of Politics, 74, 187-202.

39.

Rogoff

(1990). Equilibrium political budget cycles. American Economic Review, 80, 21-36.

40.

Schumpeter

(1942). Capitalism, socialism and democracy (1st ed.). New York, NY: Harper & Row.

41.

Shipan

Volden

(2008). The mechanisms of policy diffusion. American Journal of Political Science, 52, 840-857.

42.

Tolbert

Mossberger

McNeal

(2008). Institutions, policy innovation, and E-government in the American states. Public Administration Review, 68, 549-563.

43.

U.S. Department of Energy. (n.d.). Adoption process. Retrieved from http://www.energycodes.gov/adoption/process

44.

Van der Heijden

(2010). One task, a few approaches, many impacts: Private-sector involvement in Canadian building code enforcement. Canadian Public Administration, 53, 351-374.

45.

Van der Heijden

(2013). Privatization of building code enforcement: A comparative study of regimes in Australia and Canada. International Journal of Law in the Built Environment, 2, 60-75.

46.

Walker

(1969). The diffusion of innovations among the American states. The American Political Science Review, 63, 880-899.

47.

Wyoming State Fire Marshall’s Office. (2010). History of the Wyoming State Fire Marshall’s Office. Retrieved from http://wyofire.state.wy.us/pdf/HistoryofSFMO.pdf