Over half of the US states have jettisoned an exclusive focus on profit maximization for shareholders and created new corporate structures, called “benefit corporations”, which give equal standing to the achievement of social and environmental objectives. This paper aims to examine the factors leading to adoption of legislation for the business formation of benefit corporations by the US states.
Event History Analysis (EHA), a time-series technique using panel data of non-repeatable events, is used to identify and understand economic, political and diffusion factors that affect the adoption of benefit corporation enabling legislation in the US states.
The results strongly indicate that politics matters – states in which the Democratic Party or liberal ideology controls governmental functions are more likely to pass these laws. There is also evidence that states that are more innovative in their approach to policy-making are more likely to adopt these laws. Otherwise, unemployment, tax burden, political culture, enacted constituency statutes and geographic diffusion have no discernible relationship with the adoption of benefit corporation laws.
The paper provides warning signs to firms considering expending costly resources on the establishment of or conversion to benefit corporation status and the related investment in developing skills for the preparation, review and assurance of required annual benefit corporation reporting.
The findings suggest future adoption of benefit corporation enabling laws may slow considerably.
CitationDownload as .RIS
Emerald Publishing Limited
Copyright © 2018, Emerald Publishing Limited
We don’t hire people to make brownies, we make brownies to hire people. – Greyston Bakery, New York state benefit corporation
Corporate social responsibility (CSR) is an emerging issue in business. The number of US firms reporting on their CSR activities more than doubled from 32 per cent in 2005 (KPMG, 2008) to 86 per cent in 2013 (KPMG, 2013), and over half of the US states have recently jettisoned an exclusive focus on profit maximization for shareholders and created new corporate structures that give equal standing to the achievement of social and environmental objectives. This new corporate form, called a “benefit corporation”, is designed to promote CSR by recognizing the creation of a public good.
The first benefit corporation legislation was adopted in 2010, and benefit corporation legislation has been passed quickly by 30 states and the District of Columbia. Delaware, which is the location of 50 per cent of all publicly traded companies, adopted this corporate form into law in 2013. Benefit corporations are such a new entity in the business landscape that few academic studies have investigated them. Given their recent emergence, business salience and the paucity of research about them, the purpose of this article is to identify and understand the factors that lead to states adopting benefit corporations as a legal corporate form. In particular, it assesses how competing economic and political factors and diffusion patterns affect the adoption of laws that enable benefit corporations.
In this endeavor, this research proceeds as follows. Section 2 introduces the limited literature about benefit corporations and how this new corporate structure fits with conventional corporate structures. Section 3 presents a number of hypotheses regarding factors affecting the adoption of benefit corporation enabling legislation and details their theoretical motivations. Section 4 describes the state-level data used and the event history analysis used to estimate the relationships between these factors and a state’s adoption of benefit corporation legislation. Section 5 details the results, which assess economic, political and diffusion factors that may play a role in the adoption of this legislation. Finally, Section 6 offers a discussion of the results showing systematic elements related to the adoption of this legislation.
2. Background: benefit corporations, corporate social responsibility and the accounting profession
Benefit corporations require the creation of a public good – that is, a material positive impact on society and the environment – but also may seek a profit for shareholders. Operating conceptually between public sector (Governments) and private sector (traditional C- and S-corporations) organizations, benefit corporations reside in the “gray sector” along with, but distinct from, non-profit organizations as well as business forms such as low-profit (L3C) or non-profit limited liability companies, social/special/flexible purpose corporations and government-related enterprises such as turnpike authorities, the Federal National Mortgage Association (Fannie Mae) and the California Public Employees’ Retirement System.
Until recently, the US corporate law had not recognized the legitimacy of any corporate purpose other than maximizing profits, and shareholder wealth maximization was the basis for case law regarding conflicts about whose interests should be given primacy for over 100 years. By this perspective, directors have a fiduciary duty to shareholders to maximize profits, sometimes described as shareholder primacy or Revlon duties, and social and environmental goals are subordinate. One narrow and fairly recent exception (first adopted in Pennsylvania in 1983) to shareholder primacy is state constituency statutes, which have been adopted in 33 states and allow directors of acquired corporations to consider more than shareholder value during change-of-control decisions. For example, they may advance non-shareholder interests by choosing to be sold to the second highest bidder because that bidder agrees to keep the firm in the community, thus retaining employee jobs (Bisconti, 2009). Outside of change-of-control situations covered by constituency statutes, directors may consider non-shareholder interests if a logical connection exists to shareholder value. The “business judgment rule” is a legal rebuttal to shareholder primacy available to all corporations. For instance, courts have ruled that philanthropic or charitable giving, which takes money from shareholders and gives it to other stakeholders, is allowable when it is done with the intention of improving the firm’s reputation and sales. However, courts have struck down the business judgment defense where there is no clear demonstration of a legitimate threat to corporate effectiveness (Hill and McDonnell, 2012).
Benefit corporations have emerged in the absence of US laws permitting socially focused corporations and legal clarity regarding for-profit and mission-driven business actions. This new legal business form, entered into voluntarily, adds three provisions to structures that govern traditional corporations. First, benefit corporation status requires the company to create a public good and seek a profit for its shareholders. The public good may be any corporate purpose to create a material positive impact on society and the environment. Examples could be providing services to low-income individuals and underserved communities, promoting the arts, preserving the environment or improving human health. Second, directors of benefit corporations must consider the needs and desires of all stakeholders. Other than shareholders, stakeholders may include past, present and future employees, vendors, customers, creditors, community members and future generations. Third and lastly, benefit corporations must deliver an annual benefit report to shareholders and the public that includes an assessment of its social performance against a third-party standard. In some cases, the report must also be filed with the Department of State. This third aspect of benefit corporations, the required annual benefit report, is a critical factor that makes benefit corporations particularly interesting because of the cost of preparation and potential demand for assurance.
Though distinct from benefit corporations, CSR and its reporting history can shed light on the demand for a benefit corporation form and on benefit corporation reporting requirements. CSR reporting is a voluntary disclosure by traditional corporations and typically includes environmental, social justice and other sustainability reporting. Because of demand from investors for expanded disclosures through such mechanisms as the Carbon Disclosure Project and ESG funds, CSR reporting grew quickly for firms worldwide, reaching 71 per cent in 2013, with 93 per cent of the world’s largest 250 companies reporting their CSR activities (KPMG, 2013). The reporting of the US firms more than doubled from 32 per cent in 2005 (KPMG, 2008) to 86 per cent in 2013 (KPMG, 2013). The public’s demand for traditional corporations’ support of environmental and social justice could translate into demand for and support of benefit corporations with their public good requirement.
Now that CSR reports are widespread, stakeholders are now exerting pressure to have this information provided with financial reports. For example, the Technical Working Group (TWG) of the Carbon Disclosure Standards Board supports a single reporting framework with environmental information placed in “mainstream reports”, such as annual reports for the US firms (Climate Disclosure Standards Board, 2009). This integrated reporting is quickly gaining popularity. In 2012, 51 per cent of companies that reported on their CSR included this information in their financial reports, while only 9 per cent did in 2008 (KPMG, 2013). If the reporting of both financial and non-financial data are connected, such as with enhanced reporting that connects strategy, risk and performance, key performance indicators (including financial measures) and sustainability opportunities and impacts, then this reporting is typically already externally assured.
Even when CSR reports are not integrated with the financial statements, there is more pressure by stakeholders to have this information and to have it externally assured. In 2013, of the 82 per cent of G250 (the top 250 of the Fortune 500) that reported on their CSR activities using the Global Reporting Initiative (GRI, the de facto standard), reporting framework of sustainability, 59 per cent obtained external assurance. For those not externally assured, internal audit departments can provide an objective appraisal (Ackers, 2008). The next logical step is a demand for assurance of the benefit corporations’ required annual benefit reports, adding to the cost to firms for choosing this corporate form.
3. Theory and hypotheses development
Researchers have identified two primary categories of factors that affect the adoption of policies in the US states: internal state determinants and external diffusion effects (Berry and Berry, 1990; Mooney and Lee, 1995). Internal determinants broadly consist of the economic, political and social characteristics of a state that make a policy more or less appealing to a state’s policymakers.
With respect to economic characteristics, policymakers in the states with a higher unemployment rate are more likely to be attracted to economic development issues to improve the employment prospects of their citizens. Benefit corporations are a very low-cost, low-risk mechanism for competing with other states for certain types of businesses and entrepreneurs (Rawhouser et al., 2015). Further, states with higher unemployment must fund higher levels of social welfare services (e.g. unemployment benefits). The benefit corporation structure can provide a mechanism for businesses to provide social services and public goods at little to no cost to government. Consequently, we propose the following:
The probability a state will adopt benefit corporation legislation will increase as the state’s unemployment rate increases.
Also, with respect to economic characteristics, tax burden may impact passage of benefit corporation legislation. States with higher tax rates need mechanisms to compete with states with lower tax rates (Miller and Richard, 2010). Offering benefit corporation status is a very low-cost way to partially offset this disadvantage and to compete with other states for socially concerned businesses and entrepreneurs. Consequently, we propose the following hypothesis:
The probability a state will adopt benefit corporation legislation will increase as the state’s tax burden increases.
Shifting to political characteristics, partisans often disagree about policies related to CSR. For instance, Democrats are more likely to support policies that provide a social safety net (Pew Research Center, 2012) and counter climate change than Republicans (Coley and Hess, 2012; Guber, 2012; Pew Research Center, 2013). The literature finds that political preferences influence corporate decisions such that firms with higher (lower) investments in CSR are more likely to be located in Democratic (Republican) states (Rubin, 2008). Furthering this research, Di Giuli and Kostovetsky (2014) add that the political campaign contributions of CEOs, founders and independent directors to Democratic candidates are associated with a higher expenditure on CSR, even to the detriment of firm value. Higher Democratic support for CSR could translate into increased demand for and support of benefit corporations with their public good requirement in Democratic states.
Further, political party control of the state government may affect adoption of benefit corporation enabling legislation. Institutionally, both the legislative (i.e. state House and Senate) and executive (i.e. governor) branches of state government must consent to a bill for it to be enacted into law. Proposed bills are more likely to pass when the same political party, Democratic or Republican, controls the state legislature and governorship because both branches of government are more likely to be in ideological agreement about policies and therefore cooperate. This is a condition called “unified” government. On the other hand, when one party does not control both the legislative and executive branches, proposed bills are less likely to pass because of decreased ideological agreement on policy solutions and decreased cooperation. This is a condition called “divided” government. Consequently, the following hypothesis:
The probability a state will adopt benefit corporation legislation will increase with unified Democratic control of the state government and decrease with unified Republican control of the state government.
State political culture also may affect adoption of benefit corporation legislation. Political culture broadly refers to attitudes, orientations and values, outside of partisanship, regarding the role of citizens and government in the political system. In the state government, political culture establishes the range and boundaries of: permissible government actions, participation by citizens and groups in policymaking and government practice including innovativeness and tolerance of corruption (Leckrone, 2015; Lieske, 2010). Because the USA is highly ethnically, culturally and economically diverse (Koven and Mausolff, 2002), it is not surprising that states have different political cultures and, therefore, that state political culture could affect the adoption of state policies (Elazar, 1966). While there have been a number of related estimates of state political culture (Elazar, 1966; Sharkansky, 1969), Rice and Sumberg (1997) focus on civic culture in terms of levels of civic engagement; political equality; solidarity, trust and tolerance; and social structures of cooperation. These authors conclude that states with greater levels of each are more civic minded. In particular, states with high levels of cooperative social structures and solidarity may be more open to socially sensitive legislation. Consequently, we propose the following:
The probability a state will adopt benefit corporation legislation will increase as the state’s civic mindedness increases.
Policy innovativeness is related to state political culture. Some states are more open to enacting unique or new legislation than other states (Boehmke and Skinner, 2012; Walker, 1969). Studies indicate that factors such as state size (larger), wealth (richer) and urban density (denser) are related to (more) policy innovativeness (Boushey, 2010). For instance, states such as California are rated highly in numerous measurements of innovativeness, while states such as Wyoming lag in adoptions of policies across numerous policy areas. Because benefit corporations only recently emerged as a corporate structure and because the business community and governments are still assessing how to address CSR effectively, more innovative states may be more receptive to benefit corporation legislation. Consequently, the following hypothesis:
The probability a state will adopt benefit corporation legislation will increase when the state is more innovative in policy adoption.
Another internal determinant is the existence of constituency statutes within a state. Constituency statutes allow a corporation to consider issues other than profit maximization or shareholder primacy in change-of-control situations. The existence of a constituency statute may indicate a state is more open to social considerations in corporate actions. Consequently, we propose the following:
The probability a state will adopt benefit corporation legislation will increase when the state has already adopted a constituency statute.
Some policy analysts also consider internal determinants such as legislative professionalism, election proximity, news coverage, conflicting interest groups and education levels. The attributes of benefits corporation legislation suggest that these other measures are not relevant to understanding the adoption of this legislation. The proposed and adopted laws have not been complex, which reduces the need for longer legislative sessions and better staff resources generally found in more professional legislatures (Squire, 1992). Further, the laws have not been highly salient in the states, suggesting that election proximity, conflicting interest groups and education levels are less important than for some other policies. For instance, few companies obtain B-certification in a state before the state’s passage of the law (an average of fewer than eight firms per state in states that passed the legislation). There is often a lack of internet search interest and media coverage regarding proposed benefit corporation legislation, and there has been a limited committee testimony regarding proposed benefit corporation legislation (fewer than ten per state).
On the other hand, external factors or diffusion acknowledge that the probability a state adopts a law is influenced by the policy choices of other states (Berry and Berry, 1999, p. 310), typically thought to occur through mechanisms of learning, imitation and/or competition (Maggetti and Gilardi, 2015). Geographic diffusion assumes, and a number of studies have shown, that state governments are influenced by states with similar economic and social problems and channels of influence, such as states with shared borders and states within their region (Berry and Berry, 1999, p. 317). Researchers have found that contiguous (neighboring; Heyndels and Vuchelen, 1998) and regional neighbors (Hageman and Robb, 2011) influence policy adoption. Diffusion may occur because a state is competing with another, such as when a state adjusts tax structures to attract businesses, or because a state is learning about a more effective policy approach, such as enacting building code (Go, 2015) or renewable energy (Stoutenborough and Beverlin, 2008) legislation. Consequently, we propose:
The probability a state will adopt benefit corporation legislation will increase when geographically proximate states have adopted similar legislation.
4. Method and data
4.1 Event history analysis
Like many policy adoption studies, this research uses event history analysis (EHA), a time-series technique used to test relationships with rare events (Berry and Berry, 1992; Hageman and Robb, 2011; Hayes and Dennis, 2014). EHA uses panel data of non-repeatable events to identify, in this case, internal state determinants of policy adoption and to detect external policy diffusion. Individual US state-years (e.g. California 2010, CA 2011, FL 2010, FL 2011 and so on) serve as the unit of analysis. By comparing states, this study controls for shared national-level political and economic institutions and identifies state-level factors that account for the observable divergence in the adoption of this legislation.
4.2 Dependent variable
The dependent variable (STATUS) is the adoption or not of legislation enabling the formation of benefit corporations in a given state, in a given year (Benefit Corporation Information Center, 2015). The sample period begins in 2010, the year the first state passed benefit corporation legislation, and ends in November 2015, the last period available to collect data before this research was completed. During this time frame, 30 states adopted benefit corporation legislation. Following EHA protocol, states are coded 0 for each year they have not passed legislation, then coded 1 in the year they adopt benefit corporation legislation, after which they are dropped from the data set in succeeding years. For instance, STATUS for California is coded 0 in 2010, 0 in 2011 and 1 in 2012, the year it was adopted, then missing for 2013 to 2015. As such, Maryland, the state adopting the first year, contributes one state-year observation, while later adopting states contribute between two and six observations, such as California, which contributes three observations. The 20 non-adopting states contribute six observations, each coded 0 for years 2010 to 2015. There are a total of 234 observations. See Appendix 1 for state status and, if applicable, adoption year.
4.3 Independent variables
The literature identifies several economic, political and geographic factors that affect policy adoption. In these analyses, the economic explanatory variables include unemployment rate (UNEMPLOYMENT) and tax burden (TAXBURDEN). The economic data are lagged one year to avoid endogeneity (Miller and Richard, 2010). Unemployment rate is the number of eligible workers who are unemployed as a per cent of the labor force, as reported by the Bureau of Labor Statistics, United States Department of Labor (2015). Tax burden is proxied by the total effective business tax rate, the ratio of state business taxes to private-sector gross state product, as reported by Ernst & Young and Council on State Taxation (2014). Data for 2014 for this measure have not been reported, so the value used for 2015 is imputed by calculating the mean of the previous five years.
The political explanatory variables include political party control (STATECONTROL), political culture (POLCUL), policy innovativeness (INNOVATE) and the existence of constituency statutes (CONSTITUENCY). STATECONTROL is obtained from the National Conference of State Legislatures (2014). Its effect is captured through indicator variables for Republican control of the executive and legislative branches of government and divided government (both coded 1 for the specified condition and 0 otherwise) with Democratic control serving as the comparison group. Importantly, this measure captures both the institutional and ideological aspects of policy adoption. Political culture (POLCUL) is captured using Rice and Sumberg’s (1997) measure of civic mindedness, a continuous measure ranging from −1.5 to 1.5, in which larger values indicate more civic-mindedness. Policy innovativeness (INNOVATE) is captured by a methodology developed by Walker (1969) that has been updated and expanded by Boehmke and Skinner (2012). INNOVATE is a continuous measure indicating the proportion of possible adoptions undertaken and ranges from 0.03 to 0.10, in which larger values indicate a greater propensity for policy innovation. Finally, the existence of a constituency statute (CONSTITUENCY) in a state is obtained from Loewenstein (2013) and coded 1 if the state had previously adopted and continues to hold constituency statutes and 0 otherwise.
The external diffusion explanatory variable is tested using a methodology for assessing geographic diffusion that captures two important sources of diffusion. Researchers often model geographic diffusion as a process in which policies spread from next-door-neighbor states (Mooney, 2001) or states within the same identifiable region (e.g. Census or Bureau of Economic Analysis region; Allen et al., 2004; Chamberlain and Haider-Markel, 2005; Miller and Richard, 2010). Following Hageman and Robb (2011), who found that a combined measure demonstrated geographic diffusion of state R&D tax credit laws, this study uses a measure that captures both neighbor and regional diffusion. That is, geographic diffusion (REGPROXPROP) is calculated as the proportion of contiguous states and states within the same region as identified by the US Department of Commerce’s Bureau of Economic Analysis (without double counting) that have previously adopted benefit corporation legislation.
EHA typically assumes a constant “hazard” rate (adoptions/risk set) (Hageman and Robb, 2011). Analyses indicate these data do not have a constant hazard rate (see Appendix 2). Following common practice (Chamberlain and Haider-Markel, 2005; Mooney and Lee, 1995; Sylvester and Haider-Markel, 2015), then, the models include a nonlinear trend variable (TREND), which is constructed by taking the square root of the number of years between a given state-year and the year with the highest adoption rate or hazard rate, in this case 2014.
The relations are estimated using probit regression because the dependent variable, STATUS, is dichotomous. The following probit model is used to test the relationships between the proposed explanatory variables and status:
5.1 Descriptive statistics, correlation coefficients and tests for multicollinearity
Descriptive statistics for the variables are presented in Table I. Correlation coefficients are presented in Table II. There are a total of 234 observations. Appendix 1 lists the 30 states that have passed the legislation through November 2015.
5.2 Regression results
Diagnostics tests suggest little to no multicollinearity but consequential heteroskedasticity, so the model is estimated using robust standard errors. Table III displays the results for the primary probit model. The overall model fit is statistically significant (Wald x2 = 22.81, p = 0.007; log likelihood = −77.13) and the McFadden’s pseudo R2 is 0.14. The model correctly classifies adoption of benefit corporation enabling legislation in 88 per cent of cases (with a probability cutoff of 0.5) with a proportional reduction of error of 3.3 per cent.
The multivariate results support two of the seven hypothesized explanations for the adoption of benefit corporation enabling legislation. The model suggests Democratically controlled governments are much more likely to adopt benefit corporation legislation than Republican-controlled governments (26 per cent probability versus 8 per cent probability; p = 0.005, one-tailed). For complete reporting, it is important to note that states with divided government had a 13 per cent probability of adopting this legislation (versus Democratic states p = 0.04, one tailed; versus Republican states p = 0.13, one-tailed). Further, more innovative states are more likely to adopt benefit corporation legislation (p = 0.02, one-tailed) such that the most innovative states have a 42 per cent probability of adopting while the least innovative states have a 5 per cent probability. Other than the two political characteristics of state party control and policy innovativeness, the results do not support expectations regarding economic characteristics (unemployment and tax burden), political culture, constituency statutes or geographic diffusion.
5.3 Additional analysis/robustness checks
Although the above variables represent the most valid representations of the theoretical relationships to be tested, reasonable scholars could question whether the reported results are robust, or are the artifacts of the data used. As such, the model is re-estimated a number of times using alternative measures for political party control, political culture, policy innovativeness and geographic diffusion. The results (shown in Table IV) are substantially similar.
Though political party control captures both institutional factors related to passing legislation and ideological orientation toward benefit corporations, in studies such as this researchers have used measures of state citizen ideology and state government ideology. Citizen ideology is a measure of the mean position of state voters on the liberal–conservative continuum as reported by Berry et al. (1998) and updated by Fording (2015). The measure ranges from 13.5 to 91.9, with a mean of 47.9 (sd = 14.7) and with higher scores indicating greater citizen liberalism. State government ideology is a measure of state legislative ideology along the liberal–conservative continuum based on Poole’s (1998) methodology for measuring congressional ideology as applied by Berry et al. (2010) to state legislatures and updated by Fording (2015). The measure ranges from 2.6 to 92.5 with a mean of 43.0 (sd = 28.2) and with higher scores indicating greater legislator liberalism. Citizen and state government ideology are strongly correlated with each other (r = 0.67, p < 0.001) and state party control (Spearman rho = −0.62, p < 0.001 and Spearman rho = −0.90, p < 0.001, respectively). In each case, and consistent with the main findings, the alternative model specifications indicate more liberal/Democratic states are more likely to adopt benefit corporation legislation than more conservative/Republican states (p = 0.04 and p = 0.02, one tailed). That is, the probability of adopting benefit corporation legislation drops from 36 per cent and 26 per cent in the most liberal/Democratic states to 4 per cent and 6 per cent in the most conservative/Republican states. Further, the effects of the other explanatory variables remain substantially the same. Policy innovativeness continues to exert a consequential effect (p = 0.03 and p = 0.01, one tailed) in these alternative models. Otherwise, none of the other explanatory variables exerts a consequential effect.
Two alternative measures of political culture appear with some frequency in the literature. Elazar (1966) puts the states into three categories of civil society: individualistic, moralistic and traditionalistic. Broadly speaking, politics in individualistic and moralistic states focuses on enhancing private versus community interests, respectively, while politics in traditionalistic states focuses on maintaining the status quo. The Elazar measure identifies 15 individualistic states (e.g. Illinois and Nevada), 17 moralistic states (e.g. Oregon and Wisconsin) and 16 traditionalistic states (e.g. Florida and Oklahoma). This specification includes indicator variables for moralistic and traditionalistic states as well as an indicator for Alaska and Hawaii, making individualistic states the comparison group. Sharkansky (1969; updated by Koven and Mausolff, 2002 and Witko and Newmark, 2005), on the other hand, places states on a continuous scale that ranges from 1 to 9 with a mean of 4.9 (sd = 2.6) and with lower scores tending to be associated with Elazar’s moralistic states and higher scores with his traditionalistic states. The results suggest Elazar’s individualistic states may be more likely to adopt benefit corporation legislation than moralistic states (18 per cent versus 9 per cent; p = 0.09, two tailed), but none of the other Elazar or Sharkansky effects is statistically significant. Further, the effects of the other explanatory variables remain substantially the same in these alternative models. Democratic Party political control relative to Republican Party control (p = 0.01 and p = 0.009, one tailed) and policy innovativeness (p = 0.02 and p = 0.02, one tailed) continue to exert a consequential effect. None of the other explanatory variables exerts a consequential effect.
Boushey (2010) offers an alternative measure of policy innovativeness to Boehmke and Skinner (2012). This measure, which views policy diffusion as an epidemiological process, places the states on a continuous scale that ranges from 0.24 to 0.62, with a mean of 0.39 (sd = 0.07) and with higher numbers indicating greater innovativeness. This measure is highly correlated with the Boehmke and Skinner measure (r = 0.81, p < 0.001). Despite the high correlation between the two measures of policy innovativeness, this measure only approaches conventional levels of statistical significance (p = 0.09, one tailed). Democratic Party control continues to affect adoption relative to Republican control (p = 0.003, one tailed). None of the other explanatory variables exerts a consequential effect besides the presence of a constituency statute, which experiences a jump to approach conventional levels of statistical significance (p = 0.07, one tailed).
Finally, scholars typically estimate the effects of geographic diffusion using either an immediate-neighbor or fixed-region method. The main model, which suggests no geographic diffusion, uses a combined measure proposed and found effective by Hageman and Robb (2011). To confirm the null results reported from the main model, the model is re-estimated using a measure of neighbor diffusion that indicates the proportion of contiguous states that have previously adopted a benefit corporation law following the tradition of Mooney (2001) and Chamberlain and Haider-Markel (2005). The measure ranges from 0 to 1 with a mean of 0.18 (sd = 0.24). To further confirm the results, the model is also re-estimated using measures of fixed-regional diffusion, which indicate the proportion of states in an identified region that have previously adopted a benefit corporation law. The first regional categorization, following Hageman and Robb (2011) and Miller and Richard (2010), is the eight US Bureau of Economic Analysis regions, which group states based on economic and social homogeneity. The measure ranges from 0 to 0.8, with a mean of 0.12 (sd = 0.19). The second regional categorization, following Chamberlain and Haider-Markel (2005) and Allen et al. (2004), is the four Census Bureau, United States Department of Commerce (2015) statistical regions, which group states based on location, economic systems, ethnicity of settlers, climate, topography and systems of the local government. The measure ranges from 0 to 0.9, with a mean of 0.14 (sd = 0.19). The BEA and Census regional measures are strongly correlated with each other (r = 0.83, p < 0.001). Consistent with the main model, the results indicate benefit corporation legislation does not diffuse geographically, either through immediate neighbors or regions (minimum p > 0.30, one tailed, for BEA regions). In these alternative models, Democratic Party control continues to exert a consequential effect relative to Republican control (maximum p = 0.005, one tailed, for immediate neighbor and BEA regions), as does policy innovativeness (maximum p = 0.02, one tailed, for BEA regions). None of the other explanatory variables exerts a consequential effect.
In summary, the robustness checks confirm the findings of the main model. Political party control and policy innovativeness consistently affect the passage of benefit corporation enabling legislation. The remaining factors consistently do not. When the model is run with just the two significant predictors the model fit remains statistically significant (Wald x2 = 11.01, p = 0.012; log likelihood = −84.36), and the model continues to correctly classify adoption of benefit corporation enabling legislation in 88 per cent of cases (with a probability cutoff of 0.5) and a proportional reduction of error of 3.3 per cent, but the McFadden’s pseudo R2 drops from 0.14 in the full model to 0.06 in the limited model.
This research is designed to identify and understand the factors associated with the adoption of benefit corporation laws by states. The results strongly indicate that politics matters. States in which the Democratic Party or liberal ideology controls governmental functions are more likely to pass these laws than states controlled by the Republican Party or conservative ideology. Further, there is evidence to support that states that are more innovative in their approach to policymaking are more likely to adopt these laws. Otherwise, unemployment, tax burden, political culture, enacted constituency statutes and geographic diffusion have no discernible relationship with the adoption of benefit corporation laws.
Benefit corporation laws have spread quickly in the USA, being adopted in 30 states in the 6 years since the first law passed in 2010. These results suggest the spread of benefit corporation laws may slow, possibly dramatically. Of the 20 states that have not passed a benefit corporation law to this point, none is Democratically controlled, while 13 are controlled by Republicans. As previously indicated, the probability that a Republican state passes benefit corporation legislation is only 8 per cent (versus 26 per cent for a Democratic state and 13 per cent for a state with divided control). Because changes in partisanship and ideology are slow, as are changes in state policy innovativeness, the most likely progress will be made in the seven divided states without enabling legislation, which, even with complete success, would leave almost one-quarter of the states without benefit corporation legislation.
Overall, these results suggest that there are systematic elements to the adoption of benefit corporation enabling laws by states. While the future is notoriously hard to predict, if current effects hold true over the next several years, future adoption of these laws may slow considerably, and firms considering establishment or conversion to benefit corporation status may wish to delay this financially costly decision.
Descriptive statistics for model variables
|Variable||N||Mean or proportion†||SD||Minimum||Maximum|
|Tax – state and local||234||0.053||0.020||0.033||0.179|
Where Status = adoption (coded 1) or not (coded 0) of legislation enabling the formation of benefit corporations in a given state in a given year; Unemployment = number of eligible workers who are unemployed as a per cent of the labor force (lagged one year); Tax = total effective business tax rate, the ratio of state business taxes to private-sector gross state product (lagged one year and 2014 is repeated for 2015); State control = indicator variables for Republican control of both the executive and legislative branches of government, divided government or Democratic control of both the executive and legislative branches of government and divided government; Political culture = measure of civic mindedness; Innovativeness = propensity for policy innovation; Constituency statute = existence of constituency statutes legislation (coded 1) or not (coded 0); Proximity = proportion of contiguous states and states within the same Bureau of Economic Analysis (without double counting) that have previously adopted benefit corporation legislation; Trend = square root of the number of years between a given state-year and the year with the highest adoption rate or hazard rate, in this case 2014.
Proportions are presented for indicator variables and means and standard deviations are presented for continuous variables
Correlation coefficients (Spearman and Pearson†) for dependent variable, independent variables and control (n = 234)
|4. State control||−0.150*||−0.123*||0.072||1|
|5. Political culture||0.025||−0.399*||0.172*||−0.312*||1|
|7. Constituency statute||0.054||0.040||−0.015||0.074||0.142*||0.114||1|
All variables defined in Table I;
Correlation coefficients are Spearman for categorical and Pearson for continuous;
indicates significance at the 5 per cent level
Determinants of state adoption of benefit corporation enabling legislation
|Variable||Coef.||Robust Std. Err.||z-score||Hypothesis||Supported|
|N (state years)||234|
|Wald χ2 (degrees of freedom)||22.81** (9)|
All variables defined in Table I;
Democratic is the omitted category;
indicates significance at the 5 per cent level, using one-tailed tests
Alternative specifications of determinants of state adoption of benefit corporation enabling legislation
|Variable||Original Coefficient (Robust Std. Err.)||H3a||H3b||H4a||H4b||H5a||H7a||H7b||H7c|
|H1||Unemployment||−8.13 (8.20)||−10.52 (8.29)||−10.56 (8.68)||−9.48 (7.42)||−7.96 (7.95)||−6.13 (8.03)||−8.13 (8.20)||−8.17 (8.20)||−8.53 (8.19)|
|H2||Tax||−6.81 (6.44)||−11.31** (6.86)||−8.37 (7.11)||−7.96 (7.26)||−6.27 (6.33)||−9.80* (6.27)||−7.12 (6.38)||−7.46 (6.78)||−7.34 (6.47)|
|Divided||−0.51** (0.30)||–||–||−0.40* (0.29)||−0.50** (0.29)||−0.59** (0.30)||−0.52** (0.31)||−0.54** (0.31)||−0.52** (0.30)|
|Republican||−0.83** (0.32)||–||–||−0.71** (0.31)||−0.81** (0.31)||−0.93** (0.32)||−0.84** (0.33)||−0.83** (0.32)||−0.82** (0.32)|
|H3b||State ideology||–||–||0.01** (0.01)||–||–||–||–||–||–|
|H4||Political culture||−0.12 (0.21)||−0.14 (0.21)||−0.17 (0.21)||–||–||−0.15 (0.22)||−0.12 (0.21)||−0.12 (0.21)||−0.13 (0.21)|
|H4b||Sharkansky culture||–||–||–||–||0.04 (0.05)||–||–||–||–|
|H5||Innovation||22.35** (10.99)||23.15** (11.79)||26.72** (11.56)||24.98** (11.64)||23.56** (11.40)||–||22.18** (10.63)||22.05** (10.81)||21.89** (10.82)|
|H5a||Boushey innovation||–||–||–||–||–||2.37* (1.81)||–||–||–|
|H6||Constituency statute||0.31 (0.25)||0.10 (0.25)||0.24 (0.25)||0.11 (0.27)||0.30 (0.25)||0.35* (0.24)||0.31 (0.25)||0.33* (0.25)||0.34 (0.25)|
|H7||Proximity||−0.06 (0.60)||−0.31 (0.64)||−0.18 (0.59)||−0.27 (0.58)||−0.09 (0.58)||−0.08 (0.56)||–||–||–|
|H7b||Census region||–||–||–||–||–||–||–||0.32 (0.64)||–|
|H7c||BEA region||–||–||–||–||–||–||–||–||0.29 (0.56)|
|Trend||−0.54** (0.23)||−0.55** (0.23)||−0.59** (0.24)||−0.60** (0.24)||−0.55** (0.23)||−0.53** (0.23)||−0.55** (0.20)||−0.45** (0.23)||−0.46** (0.23)|
|Intercept||−0.35 (0.91)||−1.28* (0.83)||−1.21* (0.83)||0.11 (0.93)||−0.61 (0.89)||−0.07 (0.98)||−0.29 (0.98)||−0.51 (0.92)||−0.48 (0.91)|
|Model statistics (N = 234 state years)|
|Wald χ2 (df)||22.81*** (9)||26.28*** (8)||21.17** (8)||24.16** (11)||23.45** (9)||21.81** (9)||22.61** (9)||22.85** (9)||24.05** (9)|
Alternative models testing alternative variables (original variables defined in Table I); Column H3a replaces State control with Citizen ideology = mean position of state voters on the liberal-conservative continuum; Column H3b replaces State control with State ideology = a measure of state legislative ideology along the liberal-conservative continuum; H4a replaces Political culture with Elazar culture = three categories of civil society: individualistic (comparison group), moralistic and traditionalistic; H4b replaces Political culture with Sharkansky culture = Sharkansky’s scale of political culture; H5a replaces Innovation with Boushey innovation = Boushey’s measure of policy innovativeness; H7a replaces Proximity with Contiguous = proportion of contiguous states that have previously adopted a benefit corporation law; H7b replaces Proximity with Census region = proportion of states in the four US Department of Commerce Census Bureau statistical regions that have previously adopted a benefit corporation law; and H7c replaces Proximity with BEA region = proportion of states in the eight US Department of Commerce Bureau of Economic Analysis regions that have previously adopted a benefit corporation law;
***Indicates significance at the 10, 5 and 1 per cent levels, respectively, using one-tailed tests;
The omitted category
Annual hazard rates for a state adopting benefit corporation legislation
|Year||Adoptions a||Risk set b||Hazard rate c (%)|
Adoptions = total number of states passing legislation;
Risk Set = total number of states not previously having passed legislation; and
Hazard Rate = Adoptions/Risk Set (Mooney and Lee, 1995)
In 12 states (as of May 2016, Brewer (2015), verified as updated via personal communication September 2016).
In 4 states (as of May 2016, Brewer (2015), verified as updated via personal communication September 2016).
Dodge v. Ford (1919), Unocal Corporation v. Mesa Petroleum Company (1985), Revlon, Inc. v. MacAndrews & Forbes Holding, Inc. (1986), and eBay Domestics Holdings, Inc. v. Newmark (2010).
Delaware, where half of the US publically traded companies are incorporated, has never enacted a constituency statute.
Non-US “social hybrid legal forms include the Community Contribution Company in British Columbia in Canada, the UK’s Community Interest Company, Italy’s sociali impresa, France’s social solidarity co-operative, and Belgium’s Social Purpose Company” (Rawhouser et al., 2015, 34).
Delaware did not adopt this provision.
The TWG is made up of members of the Association of Chartered Certified Accountants, Canadian Institute of Chartered Accountants, International Federation of Accountants, Institute of Chartered Accountants in England and Wales, Japanese Institute of Certified Public Accountants, Deloitte LLP, Ernst & Young, KPMG LLP and PricewaterhouseCoopers (Climate, 2009).
A firm with B-certification, often referred to as benefit certification or certified b-corporations, is distinct from legal status as a benefit corporation. B-certification is a voluntary certification any for-profit company can apply for. Firms self-assess the same goals as benefit corporations (social and environmental performance, accountability and transparency), and certification is issued, for a cost, for two years by the nonprofit B Lab. B Lab randomly audits approximately 10 per cent of certified firms for compliance. In 2015, B Lab began requiring firms (LLC, partnerships, S-corporations, and C-corporations) in states with benefit corporation laws to convert to benefit corporation status for recertification. B-certifications have been issued to 1677 firms worldwide, with 906 in the USA (as of September 2016). B Lab is a driving force behind benefit corporation enabling legislation as it engages in activities such as drafting model legislation and lobbying for its passage.
Following Rawhouser et al. (2015), an alternative measure of tax burden, overall tax attractiveness, the Tax Foundations’ overall index rank, yields substantially similar results.
Alternative analyses, in which the 2015 value is identical to the 2014 value, yield substantially similar results.
Nebraska’s legislature is technically non-partisan, though most members identify as Republican (Ballotpedia, 2015). Because Nebraska’s governor is Republican, Nebraska is de facto subject to unified Republican control and coded as such for the reported analyses. Alternative analyses, in which Nebraska is coded as divided, yield substantially similar results.
Alaska and Hawaii have no contiguous neighbors but, following previous diffusion studies (Chamberlain and Haider-Markel 2005), are included in the analysis such that Washington and Oregon are neighboring states for Alaska, and Washington, Oregon and California are neighboring states for Hawaii.
Alaska and Hawaii became states after Elazar completed his work and, therefore, are uncategorized.
As before, Alaska’s neighbors are Washington and Oregon, while Hawaii’s neighbors are Washington, Oregon and California (following Chamberlain and Haider-Markel 2005).
For an updated listing of benefit corporations by state, see https://driehaus.depaul.edu/about/centers-and-institutes/institute-for-business-and-professional-ethics/Pages/benefit%20corporation%20portal%20subcategories/bcir%20-%20practitioners/state-by-state-guide.aspx or http://benefitcorp.net/businesses/find-a-benefit-corp
As of February 2017.
Appendix 1. State adoptions by year of legislation
2010: Maryland, Vermont
2011: New Jersey, Virginia, Hawaii, California, New York
2012: Louisiana, South Carolina, Illinois, Massachusetts, Pennsylvania
2013: Arkansas, Arizona, Colorado, Nevada, Oregon, Delaware, Rhode Island
2014: West Virginia, Utah, Nebraska, Connecticut, Minnesota, Florida, New Hampshire
2015: Idaho, Montana, Indiana, Tennessee
Not adopted: Alabama, Alaska, Georgia, Iowa, Kansas, Kentucky, Maine, Michigan, Mississippi, Missouri, New Mexico, North Carolina, North Dakota, Ohio, Oklahoma, South Dakota, Texas, Washington, Wisconsin, Wyoming
Ackers, B. (2008), “The need for corporate social responsibility assurance”, Paper presented at the International Conference on Environment, Penang.
Allen, M.D., Pettus, C. and Haider-Markel, D.P. (2004), “Making the national local: specifying the conditions for national government influence on state policymaking”, State Politics & Policy Quarterly, Vol. 4 No. 3, pp. 318-344.
Ballotpedia (2015), available at: http://ballotpedia.org/Nebraska_State_Senate_(Unicameral)#Partisan_affiliation (accessed 25 August 2015).
Benefit Corporation Information Center (2015), “State by State Legislative status”, www.benefitcorp.net/state-by-state-legislative-status (accessed 30 November 2015).
Berry, F.S. and Berry, W.D. (1990), “State lottery adoptions as policy innovations: an event history analysis”, American Political Science Review, Vol. 84 No. 2, pp. 395-415.
Berry, F.S. and Berry, W.D. (1992), “Tax innovation in the states: capitalizing on political opportunity”, American Journal of Political Science, Vol. 36 No. 3, pp. 715-742.
Berry, F.S. and Berry, W.D. (1999), “Innovation and diffusion models in policy research”, in Sabatier, P.A. (Ed.), Theories of the Policy Process, Westview Press. Boulder, CO, pp. 169-200.
Berry, W.D., Fording, R.C., Ringquist, E.J., Hanson, R.L. and Klarner, C. (2010), “Measuring citizen and government ideology in the American states: a re-appraisal”, State Politics & Policy Quarterly, Vol. 10 No. 2, pp. 117-135.
Berry, W.D., Ringquist, E., Fording, R. and Hanson, R. (1998), “Measuring citizen and government ideology in the American states, 1960–1993”, American Journal of Political Science, Vol. 42 No. 1, pp. 327-345.
Bisconti, A. (2009), “The double bottom line: can constituency statutes protect socially responsible corporations stuck in Revlon land”, Loyola of Los Angeles Law Review, Vol. 42, pp. 765-805.
Boehmke, F.J. and Skinner, P. (2012), “State policy innovativeness revisited”, State Politics & Policy Quarterly, Vol. 12 No. 3, pp. 303-329.
Boushey, G. (2010), Policy Diffusion Dynamics in America, Cambridge University Press, Cambridge.
Brewer, C.V. (2015), “Social enterprise entity comparison chart”, available at: http://dx.doi.org/10.2139/ssrn.2304892
Bureau of Labor Statistics, United States Department of Labor (2015), “Unemployment rates for states”, available at: www.bls.gov/lau/lastrk1X.htm (with X being the last digit of each year) (accessed 21 August 2015).
Census Bureau, United States Department of Commerce (2015), “Census regions and divisions of the United States”, available at: www2.census.gov/geo/pdfs/maps-data/maps/reference/us_regdiv.pdf (accessed 23 July).
Chamberlain, R. and Haider-Markel, D.P. (2005), “Lien on me’: state policy innovation in response to paper terrorism”, Political Research Quarterly, Vol. 58 No. 3, pp. 449-460.
Climate Disclosure Standards Board (2009), Reporting Framework Exposure Draft, Climate Disclosure Standards Board, London.
Coley, J. and Hess, D. (2012), “Green energy laws and republican legislators in the United States”, Energy Policy, Vol. 48, pp. 576-583.
Di Giuli, A. and Kostovetsky, L. (2014), “Are red or blue companies more likely to go green? Politics and corporate social responsibility”, Journal of Financial Economics, Vol. 111 No. 1, pp. 158-180.
Elazar, D.J. (1966), American Federalism: A View from the States, Thomas Y. Crowell Company, New York, NY.
Ernst & Young and Council on State Taxation (2014), “Total state and local business taxes: state-by-state estimates for FYXX”, available at: www.cost.org/WorkArea/DownloadAsset.aspx?id=87982FY13,www.cost.org/WorkArea/DownloadAsset.aspx?id=84767FY12,www.cost.org/WorkArea/DownloadAsset.aspx?id=81797FY11,www.cost.org/WorkArea/DownloadAsset.aspx?id=79162FY10, www.cost.org/WorkArea/DownloadAsset.aspx?id=76116FY09 (accessed 20 August 2015).
Fording, R. (2015), “State ideology data”, available at: https://rcfording.wordpress.com/state-ideology-data/ (accessed 23 July 2015).
Go, M.H. (2015), “Building a safe state: hybrid diffusion of building code adoption in American states”, The American Review of Public Administration, Vol. 46 No. 6, doi: 0275074014563827.
Guber, D.L. (2012), “A cooling climate for change? Party polarization and the politics of global warming”, American Behavioral Scientist, Vol. 57 No. 1, pp. 93-115.
Hageman, A.M. and Robb, S.W. (2011), “A regional diffusion theory explanation for states’ proposal and adoption of anti-passive investment company laws”, Journal of Accounting and Public Policy, Vol. 30 No. 6, pp. 551-569.
Hayes, T.J. and Dennis, C. (2014), “State adoption of tax policy new data and new insights”, American Politics Research, Vol. 42 No. 6, pp. 929-955.
Heyndels, B. and Vuchelen, J. (1998), “Tax mimicking among selgian municipalities”, National Tax Journal, Vol. 51 No. 1, pp. 89-101.
Hill, C.A. and McDonnell, B.H. (2012), “Fiduciary duties: the emerging jurisprudence”, in Hill, C.A., Krusemark, J.L., McDonnell, B.H. and Robbins, S. (Eds), Research Handbook on the Economics of Corporate Law, Edward Elgar Publishing, Northampton, MA, pp. 133-151.
Koven, S.G. and Mausolff, C. (2002), “The influence of political culture on state budgets: another look at Elazar’s formulation”, The American Review of Public Administration, Vol. 32 No. 1, pp. 66-77.
KPMG (2008), International Survey of Corporate Responsibility Reporting, KPMG, Amsterdam.
KPMG (2013), Survey of Corporate Responsibility Reporting, KPMG, Amsterdam.
Leckrone, J.W. (2015), “State and local political culture”, in Shally-Jensen, M. (Ed.), American Political Culture: An Encyclopedia, ABC-CLIO, Santa Barbara, CA, Vol. 1.
Lieske, J. (2010), “The changing regional subcultures of the American states and the utility of a new cultural measure”, Political Research Quarterly, Vol. 63 No. 3, pp. 538-552.
Loewenstein, M.J. (2013), “Benefit corporations: a challenge in corporate governance”, Corporate Governance: The International Journal of Business in Society, Vol. 13 No. 4, pp. 1007-1030.
Maggetti, M. and Gilardi, F. (2015), “Problems (and solutions) in the measurement of policy diffusion mechanisms”, Journal of Public Policy, Vol. 36 No. 1, doi: S0143814X1400035X.
Miller, C.R. and Richard, B. (2010), “The policy diffusion of the state R&D investment tax credit”, State and Local Government Review, Vol. 42 No. 1, pp. 22-35.
Mooney, C.Z. (2001), “Modeling regional effects on state policy diffusion”, Political Research Quarterly, Vol. 54 No. 1, pp. 103-124.
Mooney, C.Z. and Lee, M. (1995), “Legislating morality in the American states: the case of pre-roe abortion regulation reform”, American Journal of Political Science, Vol. 39 No. 3, pp. 599-627.
National Conference of State Legislatures (2014), “20XX State and legislative partisan control”, available at: www.ncsl.org/documents/statevote/2010_Legis_and_State_pre.pdf, www.ncsl.org/documents/statevote/LegisControl_2011.pdf, www.ncsl.org/documents/statevote/2012_legis_and_state.pdf, www.ncsl.org/documents/statevote/legiscontrol_2013.pdf, www.ncsl.org/documents/statevote/legiscontrol_2014.pdf, www.ncsl.org/Portals/1/Documents/Elections/Legis_Control_2015_Feb4_11am.pdf (accessed 31 August 2015).
Pew Research Center (2012), “American values survey: trends in American values 1987-2012”, available at: www.people-press.org/files/legacy-pdf/06-04-12%20Values%20Release.pdf (accessed 23 July 2015).
Pew Research Center (2013), “GOP deeply divided over climate change”, www.people-press.org/files/legacy-pdf/11-1-13%20Global%20Warming%20Release.pdf (accessed 12 August 2015).
Poole, K.T. (1998), “Recovering an issue space from a set of issue scales”, American Journal of Political Science, Vol. 42 No. 3, pp. 954-993.
Rawhouser, H., Cummings, M.E. and Crane, A. (2015), “Benefit corporation legislation and the emergence of a social hybrid category”, California Management Review, Vol. 57 No. 3, pp. 13-35.
Rice, T.M. and Sumberg, A.F. (1997), “Civic culture and government performance in the American states”, Publius: The Journal of Federalism, Vol. 27 No. 1, pp. 99-114.
Rubin, A. (2008), “Political views and corporate decision making: the case of corporate social responsibility”, Financial Review, Vol. 43 No. 3, pp. 337-360.
Sharkansky, I. (1969), “The utility of Elazar’s political culture: a research note”, Polity, Vol. 2 No. 1, pp. 66-83.
Squire, P. (1992), “Legislative professionalization and membership diversity in state legislatures”, Legislative Studies Quarterly, Vol. 17 No. 1, pp. 69-79.
Stoutenborough, J.W. and Beverlin, M. (2008), “Encouraging pollution‐free energy: the diffusion of state net metering policies”, Social Science Quarterly, Vol. 89 No. 5, pp. 1230-1251.
Sylvester, S.M. and Haider-Markel, D.P. (2015), “Buzz kill: state adoption of DUI interlock laws, 2005-2011”, Policy Studies Journal, Vol. 44 No. 4, doi: 10.1111/psj.12128.
Walker, J.L. (1969), “The diffusion of innovation among the American states”, American Political Science Review, Vol. 63 No. 3, pp. 880-889.
Witko, C. and Newmark, A.J. (2005), “Business mobilization and public policy in the US states”, Social Science Quarterly, Vol. 86 No. 2, pp. 356-367.
Bureau of Economic Analysis, United States Department of Commerce (2015), “BEA Regions”, www.bea.gov/regional/docs/regions.cfm (accessed 23 July 2015).