Resource Dependence, Uncertainty, and The Allocation of Corporate Political Activity Across Multiple Jurisdictions
Resource Dependence, Uncertainty, and The Allocation of Corporate Political Activity Across Multiple Jurisdictions
Resource Dependence, Uncertainty, and The Allocation of Corporate Political Activity Across Multiple Jurisdictions
RICHARD A. DEVINE
DePaul University
BRUCE T. LAMONT
R. MICHAEL HOLMES, JR.
Florida State University
This paper contributes to the resource dependence theory and corporate political ac-
tivity literatures by distinguishing dependence from uncertainty and explaining how
two different types of uncertainty have opposite effects on dependence management. We
explain how some environmental factors increase the state uncertainty associated with
firms’ dependence on government jurisdictions, whereas other factors increase response
uncertainty. We hypothesize that, due to the historical influence of the media and social
movement organizations on politics, negative national media tenor and oppositional
social movement organization resources increase state uncertainty (i.e., government’s
likely behavior toward firms becomes less predictable), strengthening the relationship
between firms’ dependence on jurisdictions and their use of political contributions in
those jurisdictions. Further, we hypothesize that top management team turnover and
politician turnover increase response uncertainty (i.e., the effectiveness of firms’ efforts
to manage their dependence becomes less clear), weakening the relationship between
dependence and political contributions. We find support for our theory in an exami-
nation of state-level political contributions of firms in environmentally intensive industries
from 2009 to 2016. Interviews with senior executives, political consultants, and senior
government employees directly involved in corporate political activity help illustrate the
hypothesized relationships.
Firms’ dependence on government is an important when and where to allocate CPA is a complex task for
motivation for corporate political activity (CPA), managers because many firms are dependent on a
such as contributions to political campaigns (Hillman, large number of jurisdictions.
Keim, & Schuler, 2004; Pfeffer & Salancik, 1978). Pol- Although we know that managers use CPA to in-
iticians hold the power to change public policy in fluence politicians in jurisdictions where they are
ways that can severely affect firm performance, so dependent, less research has articulated factors that
managers use CPA in an effort to align politicians’ in- can direct (or inhibit) managerial engagement in
terests with their own, thereby reducing the uncer- dependence management, leaving our understand-
tainty of problematic public policy changes (Pfeffer & ing of CPA incomplete (Hadani, Bonardi, & Dahan,
Salancik, 1978). As one executive told us, “[CPA] can 2017; Hillman et al., 2004; Hillman, Withers, &
mean all the difference for companies, as increased Collins, 2009; Lux, Crook, & Woehr, 2011). Conven-
or decreased regulation can have an extraordinary tionally, the resource dependence theory (RDT) ex-
impact on the bottom line.” However, determining planation of CPA has been straightforward: firms
spend more on CPA as the magnitude of their de-
We are grateful to our editor Kevin Steensma and three pendence on government increases (Hillman et al.,
anonymous reviewers for the constructive guidance they 2009; Lux et al., 2011), in general and presumably
provided. We also thank Gonzalo Molina-Sieiro for his across jurisdictions. Although the relationship be-
helpful advice. tween dependence and CPA is positive, it is quite
38
Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder’s express
written permission. Users may print, download, or email articles for individual use only.
2021 Sutton, Devine, Lamont, and Holmes, Jr. 39
weak, suggesting that certain unexplored factors consider different types and sources of uncertainty,
might strengthen or weaken the effects of depen- along with the magnitude of dependence, paints an
dence (Lux et al., 2011). This is readily observed incomplete picture of CPA and perhaps dependence
in CPA data. For example, ExxonMobil is head- management more generally.
quartered in Texas and has more production and Therefore, the purpose of this study is to explain
storage facilities in Texas than in any other U.S. state how different types of uncertainty from multiple
(United States Environmental Protection Agency, sources affect firms’ allocation of CPA across nu-
2019). Nonetheless, in 2016 ExxonMobil made po- merous government jurisdictions on which the firms
litical contributions in 20 states, and Texas ranked are dependent. We build our model by first restoring
sixth in terms of the dollars contributed by the firm the original RDT distinction between dependence
(National Institute on Money in State Politics, 2019). and uncertainty, consistent with calls to advance
Explaining this sort of heterogeneity of CPA across RDT by revisiting its past (Wry, Cobb, & Aldrich,
government jurisdictions is an opportunity for a 2013). We then integrate Milliken’s (1987) uncer-
meaningful theoretical extension of RDT’s argument tainty typology into arguments from RDT. In doing
that dependence is an important driver of CPA so, we distinguish two different types of uncertainty:
(Hillman et al., 2009). (1) uncertainty about the actions of politicians (i.e.,
Indeed, there are two important and problematic “state uncertainty”) and (2) uncertainty about the
assumptions in the research using RDT to explain proper way to use CPA to influence politicians (i.e.,
CPA: the research often assumes that (1) uncertainty “response uncertainty”). Drawing on RDT argu-
increases as dependence increases, and (2) this un- ments about how sources of uncertainty are known
certainty leads managers to take action to mitigate and monitored by firms, we identify sources of each
the dependence. Dependencies can create uncer- of these two types of uncertainty that are important
tainty for managers and their firms, but dependence in our research context, explain how they become
and uncertainty do not necessarily covary (Pfeffer & salient to firms and their managers, and explain how
Salancik, 1978). This nuanced observation from the they shape efforts to manage resource dependence in
original articulation of RDT has generally been ig- different ways. Advancing RDT, we explain how
nored in subsequent RDT theorizing. Dependence firms’ dependencies interact with the state and re-
determines the importance of the relationship or ju- sponse uncertainty from different sources to influ-
risdiction to the firm, but uncertainty in the rela- ence CPA. In this way, we extend knowledge of the
tionship or jurisdiction can amplify (or diminish) the actions managers use to mitigate their firms’ depen-
degree to which the firm actively manages the de- dence on important resource providers.
pendency. Dependence and uncertainty are foun- We develop our theory and conduct our tests in the
dational elements of RDT, so failure to account for context of state-level CPA. Specifically, we built a
both and explain how they interact could undermine database of large firms in environmentally intensive
the accuracy of predictions drawn from the theory, as industries (i.e., firms that regularly emit toxic
illustrated in the Exxon example above. chemicals) and modeled their political contributions
In addition, there are different types of uncertainty across all 50 U.S. states over an eight-year period.
(Milliken, 1987), each arising from multiple sources In this setting, we argue that firms allocate more CPA
in the environment. For instance, there might be lit- to jurisdictions (i.e., states) where they emit more
tle uncertainty about some aspects of dependence on toxic chemicals, because they are more dependent
government (e.g., what politicians plan to do), while on public policy in such jurisdictions. However,
there is significant uncertainty about other aspects drawing on RDT, we also argue that firms and their
of that dependence (e.g., how to change politicians’ managers pay more attention to stakeholders that
minds). Likewise, some types of uncertainty might historically have been, and currently are, instru-
prompt action, while other types inhibit action. mental to firm strategy. In particular, we argue that
Firms that are uncertain about impending govern- negative news media coverage and oppositional
ment policies may use CPA to manage their political social movement organizations (SMOs) (e.g., envi-
dependencies, for example, but firms that are un- ronmental groups) are important sources of state
certain about how to influence politicians may hes- uncertainty that also affect the firms’ behavior. We
itate to use CPA to manage those dependencies. chose these two sources of state uncertainty because
And firms that are dependent on multiple jurisdic- both groups historically have fueled political oppo-
tions will likely face different types and sources sition to firms in the selected industries and, in turn,
of uncertainty in different jurisdictions. Failure to are monitored closely by managers in their current
40 Academy of Management Journal February
environment. Top management team (TMT) and attention at particular points in time, typically those
politician turnover, on the other hand, are the sour- that are thought to have the greatest impact on firm
ces of response uncertainty. Because TMTs and performance (Hambrick & Mason, 1984; March &
politicians are the creators and targets, respectively, Simon, 1958).
of the CPA that firms use to manage their depen- Building on these ideas, the sensemaking perspec-
dence on government, they are the most appropriate tive explains the process of enactment, which is how
indicators of response uncertainty in this context. managers and their firms come to understand their
We submit that turnover in either group creates un- environment (Pfeffer & Salancik, 1978). Enactment is a
certainty about how to use CPA to manage depen- two-stage process: (1) managers’ preconceptions direct
dence on government. In short, our theory suggests their attention to certain factors in the environment
that firms’ dependence on various jurisdictions, and (2) managers implement structures, systems, and
along with the magnitude of these sources of state processes to monitor and perhaps influence those
and response uncertainty, combine to affect firms’ factors in the future (Weick, 1988). The sensemaking
CPA across jurisdictions. literature suggests that firms’ histories and current
Consistent with our expectations, the results reveal circumstances jointly shape this enactment by influ-
that the sources of state uncertainty strengthen, and the encing managers’ evaluations of their firms’ environ-
sources of response uncertainty weaken, the positive ments and the options available for managing in those
relationship between a firm’s dependence on a juris- environments (Weick, 1978, 1988). Managers look
diction and its CPA in that jurisdiction. In addition to back at history to determine which factors in the en-
our quantitative analysis, we also interviewed 16 in- vironment require their attention, and they develop
dustry insiders, all of whom had experience managing monitoring structures and systems to attend to these
and observing relationships between top managers and factors, assess uncertainty,1 and evaluate the options
their firms and government officials (see Appendix A for managing their dependencies. In doing so, the
for descriptions of the interviewees). We have included firms’ environmental monitoring systems, based on
their comments to shed additional light on our theory past sources of uncertainty for the firms’ managers,
and hypothesized relationships. In sum, our results direct managerial and firm attention in the present,
illuminate not only how different types of uncertainty thereby affecting the environment that managers and
shape dependence management, but also which sour- their firms enact (Weick, 1969, 1988; Weick, Sutcliffe,
ces of uncertainty managers monitor. & Obstfeld, 2005). Therefore, managers and their firms
attend to particular stakeholders because those stake-
holders have been important sources of uncertainty in
THEORETICAL BACKGROUND the past and are considered important to the firms’
current dependencies.
Resource Dependence
Although the importance of individual stakeholders
According to RDT, firms depend on external actors can vary across firms, due to their unique histories or
to provide the resources they need (Pfeffer & Salancik, their managers’ misreading of their dependencies
1978). This resource dependence grows as the need for (Pfeffer & Salancik, 1978), several factors generate some
the resources increases and as alternative sources of commonality in the sources of uncertainty that receive
the resources decrease. Resource dependence is often attention in particular industries. First, managers in the
accompanied by uncertainty, because external actors same industries face similar environmental conditions
can be unpredictable and because firms often are (Pfeffer & Salancik, 1978; Porter, 2008). Because regu-
unable to control those actors’ behavior perfectly lation is an important source of uncertainty for all
(Hillman et al., 2009). In turn, RDT suggests that man- firms in the oil and gas industry, for example, firms
agers take actions, such as acquiring key suppliers, to in this industry tend to monitor government closely
manage this uncertainty. (Cho, Laine, Roberts, & Rodrigue, 2018; Hess, 2014).
A seldom-explored aspect of RDT is the manner in Additionally, intra-industry alliances, intra-industry
which firms monitor their dependencies. Many firms
are dependent on a large number of stakeholders, but 1
Although Weick (1969, 1978, 1988) preferred the term
managers lack the time and cognitive resources to scan “equivocality” over “uncertainty” to emphasize the diffi-
and respond to every aspect of the environment con- culty of interpreting the environment, we use the term
tinuously. In response to cognitively overwhelming “uncertainty” to reflect environmental change that poses
tasks, managers simplify their task in some way problems for managers and their firms, which is more in
(Schwenk, 1984). They isolate a few variables for closer line with RDT logic.
2021 Sutton, Devine, Lamont, and Holmes, Jr. 41
movement of managers, commonality in trade publi- et al., 2011). Thus, it appears that our understanding of
cations and other media (Aldrich & Pfeffer, 1976), and how dependence and associated uncertainties shape
the conversations managers across firms have with CPA is incomplete, suggesting the need for additional
each other (Starbuck, 1975) all lead to some degree of theory to probe these relationships further. We contend
knowledge convergence in an industry. Thus, man- that (a) dependence and uncertainty may not always
agers in the same industries often hold similar view- coincide and (b) there are different types of uncertainty
points about various aspects of the environment confronting managers, which in turn might affect CPA
(Sutcliffe & Huber, 1998) and address the uncertainties in unexpected and potentially contrasting ways.
surrounding their firms’ dependencies in similar ways, RDT, as it is typically used, conflates uncertainty
even when using CPA (Hillman et al., 2009). and dependence. Greater dependence is typically
viewed as involving greater uncertainty, which mo-
tivates firm action (e.g., Hillman et al., 2009). That is,
Dependence on Government
even though uncertainty is central to RDT, it is typ-
Dependence on government is especially impor- ically assumed to be a linear function of the magni-
tant for many firms. Governments establish “rules of tude of the dependence confronting firms. However,
the game” that enable or constrain certain firm ac- the original conceptualization of RDT suggests that
tivities and can have significant effects on firm stakeholders are not problematic merely because
strategy and performance (North, 1990).2 Further, firms depend on them. Rather, they are problematic
public policy is influenced by a complex set of fac- if they are not dependable (Pfeffer & Salancik, 1978:
tors and can change quickly, unexpectedly, and 3). In other words, dependence is problematic when
significantly, all of which can create uncertainty. it involves uncertainty. In our context, the magni-
Firms attempt to manage this dependence and un- tude of dependence on a political jurisdiction re-
certainty by using CPA to create the policy environ- flects the importance of the jurisdiction to the firm,
ment they desire (Hillman et al., 2009; Pfeffer & and the extent of uncertainty associated with the
Salancik, 1978). Political contributions give firms jurisdiction determines how problematic the juris-
access to politicians (Kalla & Broockman, 2016; diction is. We argue that, by distinguishing sources
Langbein, 1986) and improve the election odds of of uncertainty and magnitude of dependence across
politicians they prefer (Magee, 2002; Stratmann, jurisdictions and examining how they interact, we
1991). The ability to influence public policy helps restore fidelity to RDT theorizing and get a better
reduce the uncertainty associated with dependence understanding of how firms manage dependencies.
on government. As a result, firms that are more de- Equally important, in our view, uncertainty is a
pendent on government tend to spend more on CPA multifaceted construct, in that managers can be un-
(Hillman et al., 2004; Lux et al., 2011). certain about several aspects of the environment
Meta-analytic evidence suggests that we lack a simultaneously (Duncan, 1972; March, 1978). In par-
complete understanding of the relationship between ticular, Milliken (1987) drew the following distinction
dependence and CPA. In particular, CPA is often in- between state uncertainty and response uncertainty:
effective at encouraging politicians to enact policies “state uncertainty” exists when managers and their
that the firms favor (Hadani et al., 2017), which raises firms do not know what the environment will be
questions about why managers continue to engage in it. like in the future, and “response uncertainty” exists
Many of the factors thought to drive CPA, such as firms’ when they do not know what actions are appropriate
levels of exposure to or dependence on regulations, for their environments.3 In the CPA context, state un-
appear to have positive albeit weak effects on CPA (Lux certainty exists when managers and their firms cannot
2 3
RDT studies have revealed two potential bases for firm Milliken (1987) defined three types of uncertainty:
dependence on government: government-based revenue state, response, and effect. We do not hypothesize about
(Hansen & Mitchell, 2001; Schuler, Rehbein, & Cramer, “effect uncertainty,” which is unpredictability about how a
2002) and regulatory control (Grier, Munger, & Roberts, certain event will affect the firm. Effect uncertainty is, in a
1994; Hart, 2001; Mitchell, Hansen, & Jepsen, 1997). In our sense, negatively related to state uncertainty: managers
context of firm dependence on U.S. state governments, must know what events will occur before they can perceive
revenue-based dependence is negligible, because relatively uncertainty about the effects of those events (Milliken,
few large firms (especially in our sampled industries) derive 1987: 137). In other words, effect uncertainty is less rele-
significant portions of their revenue from state government vant when there is state uncertainty. Thus, we focus on
contracts. However, this may not hold in other contexts. “state uncertainty” and “response uncertainty.”
42 Academy of Management Journal February
predict the actions of government and, thus, do not jurisdiction, but we also expect to find that other
know what public policies they will face in the future. variables shape the strength of this relationship.
Response uncertainty exists when managers and their
firms do not know how to influence politicians effec-
State Uncertainty
tively, perhaps because they lack experience with
CPA, do not know how to use it effectively, do not “State uncertainty” exists when managers and
know which politicians are amenable to influence, their firms find that the external environment is un-
and so on. Whereas state uncertainty resembles the predictable. For example, managers may not under-
type of uncertainty considered influential in RDT— stand how major stakeholders will behave (Milliken,
managers who are uncertain about the actions of 1987: 136). In the context of politics, state uncertainty
resource providers are more likely to take action to reflects the possibility of unpredictable changes in
reduce that uncertainty (Pfeffer & Salancik, 1978)— public policy. The timing and specifics of such
RDT does not consider response uncertainty. If man- changes can be difficult for firms to predict ex ante,
agers experience response uncertainty and are unsure but the changes can happen rather quickly and can be
about what actions to take, they might be less likely substantial.
to take action in response to their resource depen- Two stakeholder groups are especially likely to
dencies. Failure to account for response uncertainty shape the state uncertainty for firms in environ-
might account for some of the mixed or weak results mentally intensive industries: (1) the news media
in the meta-analyses on CPA. (hereafter, “the media”) and (2) environmental SMOs.
Thus, prior literature describes the nature of firms’ Returning to our earlier discussion of enactment and
resource dependencies, why they are important, sensemaking, both the media and environmental
how firms attend to them, and the strategies used to SMOs have a history of influencing public policy.
manage them. While much has been accomplished Both groups have sought to be disruptive and influ-
with RDT to date, we posit that the theorizing re- ential, often publicizing information about firms’
mains incomplete in at least two ways. We argue negative effects on the natural environment and
that RDT, as it is typically used, conflates uncer- prompting policy change in some instances. Because
tainty and dependence and has generally empha- of their historical influence, the media and environ-
sized state uncertainty while neglecting response mental SMOs are monitored closely by firms.
uncertainty. We also contend that these two limi- Indeed, monitoring the media and environmental
tations in current applications of RDT limit the SMOs was important to the firms of the individuals
explanatory power of the theory. Therefore, one we interviewed (see Results section). Their firms’
critically important question remains: Why do firms monitoring systems explicitly tracked these sources
respond more to some dependencies than others (of of uncertainty. As one manager explained, “We
the same objective magnitude)? We now turn to the spend a lot of time tracking media and local advo-
hypotheses, which focus on sources of two different cacy groups. We track it on a daily basis, and I get a
types of uncertainties (state and response) and how report each week.” Another manager similarly
the sources interact with dependence magnitude to explained, “Public relations firms are routinely en-
affect firms’ CPA. gaged to monitor activists and the media, because, if
you don’t watch them, they can create regulatory
change. You have to get ahead of it.” Thus, moni-
HYPOTHESES
toring these groups was important, at least partly
Because of CPA’s potential efficacy, firms use it because such monitoring helped the firms anticipate
to manage their dependence on government (Grier and respond to policy changes.
et al., 1994; Hart, 2001; Mitchell et al., 1997; Schuler Importantly, however, neither the media nor en-
et al., 2002). Similar to other forms of dependence vironmental SMOs are likely to affect CPA directly,
management, the use of CPA is positively related to because neither poses consequences for the firm per
the magnitude of a firm’s dependence on govern- se. Firms that are not dependent on a jurisdiction are
ment (Hillman et al., 2004; Lux et al., 2011). Meta- unlikely to be very concerned about the prospect of
analytical evidence suggests positive effects of re- the media or SMOs affecting public policy in that
source dependence on CPA, though these effects are jurisdiction. By contrast, dependence on a jurisdic-
small (Lux et al., 2011). Thus, we expect to find a tion does carry potential consequences for firms,
positive relationship between firms’ magnitudes of because dependence means the firm’s operations are
dependence on a jurisdiction and their CPA in that vulnerable to government action in that jurisdiction.
2021 Sutton, Devine, Lamont, and Holmes, Jr. 43
As the magnitude of the dependence increases, so publications such as The Wall Street Journal (Maher
does the magnitude of potential consequences if the & McWhirter, 2017). Since the pollution scandal
government changes policy in a way that is hostile to began receiving national attention, DuPont and its
the firm’s interests. In this way, dependence prompts recently spun-off operations at Chemours have re-
CPA because dependence creates concerns about ceived more government scrutiny about C8 and a
firm performance. State uncertainty amplifies these related chemical in other locations, including North
concerns. If managers of Firm F are already con- Carolina (CBS News, 2017; Weber & Fong, 2017) and
cerned about the firm’s exposure to public policy in the Netherlands (DutchNews, 2017). As this example
Jurisdiction J because the firm is heavily dependent illustrates, national media coverage poses a greater
on that jurisdiction, then the state uncertainty asso- threat of contagion than does local coverage. The
ciated with Jurisdiction J only amplifies those con- new government scrutiny also has been focused
cerns. Thus, as we explain below, different sources of directly on DuPont itself, and new regulations re-
state uncertainty are likely to strengthen the rela- garding these specific chemicals would probably be
tionship between firms’ dependence on particular costlier for DuPont than for others because the
government jurisdictions and their CPA in those chemicals are patented.
jurisdictions. Thus, negative national media coverage makes
Media. The media investigates firms suspected of political dependencies seem more problematic than
inappropriate behavior and distributes that infor- they would otherwise seem, leading firms to manage
mation in ways that can affect stakeholders’ per- large dependencies more aggressively than they
ceptions (Elsbach, 1994; Pollock & Rindova, 2003). otherwise would. Negative national media coverage
Politicians may form negative views of firms that gets managers’ attention and creates more uncer-
receive negative media coverage, because the cov- tainty about the future of the policy environment
erage can expose the firms’ misdeeds and encourage in those jurisdictions where their firms are most
other stakeholders to lobby politicians for policy dependent. Whereas firms can be expected to use
changes (Maguire & Hardy, 2009). more CPA in jurisdictions where their dependence
The tenor of media coverage also drives state un- is high, to tilt public policy to their advantage,
certainty because negative coverage poses the threat CPA should be even higher for firms that have re-
of a political contagion. Managers are concerned that ceived negative national media coverage to address
the firm’s misdeeds in one jurisdiction could be no- the heightened concern of political backlash (e.g.,
ticed by politicians and regulators in other jurisdic- policy changes or increased scrutiny) resulting from
tions, potentially multiplying the cost of poor the negative press. Large dependencies combined
environmental performance. Indeed, one of our in- with the high state uncertainty associated with
terviewees told us that one of his firm’s greatest risks negative national media coverage, in other words,
is the media publicizing environmental harm his could pose a performance and survival threat to
firm caused, potentially leading to bigger problems firms, so firms can be expected to scale up their al-
for the firm. Although firms receive coverage in a ready high CPA in jurisdictions where they have
variety of media outlets, national media coverage large dependencies.
poses the greatest threat of such a contagion, due to Positive national media coverage, by contrast,
its scope, power, and legitimacy compared with makes large political dependencies less concerning
more regional or local coverage. Additionally, be- and reduces the attention those dependencies re-
cause the TMT is centrally involved in CPA deci- ceive. Therefore, in the context of positive national
sions and the TMT has limited time, they are more media coverage, firms will use less CPA in high de-
likely to read and be influenced by national media pendence jurisdictions than they otherwise would
than by local media. have. At low levels of dependence in a jurisdiction,
For example, DuPont’s plant in Parkersburg, West the firm has little at stake, making CPA generally less
Virginia, contaminated regional water supplies for necessary, regardless of the national media coverage
decades with a toxic chemical, C8 (Blake, 2015). As the firm receives. Therefore, the interaction between
early as 2001, the pollution received frequent cov- magnitude of dependence and state uncertainty re-
erage from local newspapers in West Virginia and flected in national media coverage is mostly driven
Ohio. However, it received almost no coverage from by the heightened uncertainty of the negative cov-
national media until 2015, when a long-form Huf- erage in high dependence situations. The positive
fington Post article was published (Blake, 2015), relationship between firms’ dependence on jurisdic-
followed by critical coverage from other national tions and CPA in those jurisdictions should become
44 Academy of Management Journal February
more positive (i.e., stronger) as the negativity of firms’ The history of SMOs fighting for policy change has
national media coverage increases. Thus: led managers in many firms to attend to SMOs. Due
to their community embeddedness, their efforts to
Hypothesis 1a. The tenor of national media coverage
educate and shape their communities, and the net-
of the firm moderates the relationship between mag-
nitude of dependence and corporate political activity, work effects that follow (Rowley, 1997), the SMOs
such that negative tenor strengthens the positive re- based in a jurisdiction are likely to have more direct
lationship between dependence and corporate polit- and indirect influence over the politicians in that
ical activity. jurisdiction than are SMOs from other jurisdictions
(Pacheco & Dean, 2015; Schneiberg & Bartley, 2001;
Additionally, it is plausible that managers receive York, Hargrave, & Pacheco, 2016). Indeed, our in-
information about local media coverage from the terviewees consistently said that they monitor SMOs
employees and professional services firms that because of their goal and ability to influence politics.
monitor the firm’s coverage in the media. Negative One executive said, “Monitoring activists is becom-
media coverage by publications in a jurisdiction ing increasingly important, because elected officials
might lead managers to fear that the coverage could will respond to their pressure.” As environmental
trigger public policy changes in that jurisdiction. In SMOs in a jurisdiction gain more resources, man-
other words, local media coverage specific to a ju- agers of environmentally intensive firms that depend
risdiction could increase state uncertainty specific to
on that jurisdiction, in turn, experience more state
that jurisdiction. Therefore, we propose:
uncertainty about the future of the public policy
Hypothesis 1b. The tenor of local media coverage in a environment. Speaking about this need to under-
jurisdiction moderates the relationship between a firm’s stand how capable the oppositional SMOs are, one of
magnitude of dependence on that jurisdiction and its our interviewees told us, “The environmental left
corporate political activity in the jurisdiction, such that takes advantage and tries to influence legislation.
negative tenor strengthens the positive relationship be- There are always people trying to stir things up, but
tween dependence and corporate political activity. do they have money?” As the SMOs gather resources,
they become more influential, create more state un-
Social movement organizations. SMOs are dedi- certainty, and are worthy of more attention.
cated to changing social institutions (Zald & Ash, We expect firms to increase their CPA in jurisdic-
1966; Zald & McCarthy, 1979), and they have histor- tions where they are more dependent, but the size of
ically had significant consequences for govern- that increase should depend on the resources of op-
ments and firms (Hiatt, Sine, & Tolbert, 2009; King & positional SMOs in that jurisdiction. As oppositional
Haveman, 2008). As a result, SMOs can shape firms’ SMOs gather more resources in a jurisdiction, state
attention and behavior in important ways (Henriques uncertainty increases and managers become more
& Sadorsky, 1999; Kassinis & Vafeas, 2006). As SMOs
concerned about their dependence on that jurisdic-
gather more resources, they are more capable of dis-
tion, prompting firms to manage their large depen-
tributing information to those capable of affecting in-
dencies with even greater CPA. High levels of
stitutional change, including politicians (Cress &
dependence and high levels of oppositional SMO
Snow, 1996; McCarthy & Zald, 1977). Even in the
resources are likely to produce the highest levels of
absence of immediate policy victories, well-funded
CPA. Conversely, low SMO resources reduce con-
SMOs may influence public opinion and culture
cerns about a firm’s high dependence on a jurisdic-
(Meyer & Whittier, 1994), thereby indirectly influ-
tion, leading the firm to reduce its use of CPA to
encing politicians to make policy changes.
manage that dependency. At low magnitudes of de-
Stakeholders have heterogeneous preferences,
pendence, CPA would be expected to be low, due to
making it difficult or impossible for firms to satisfy
the firms’ limited exposure to public policy in the
all stakeholders simultaneously (Dorobantu, Henisz,
jurisdiction, regardless of the level of SMO resources
& Nartey, 2017; Freeman, 1984). Such preference
heterogeneity is even observed within stakeholder in the jurisdiction. Therefore, the level of SMO re-
groups (Wolfe & Putler, 2002). Preference heteroge- sources in a jurisdiction should make the positive
neity can be observed not only in the goals pursued relationship between dependence and CPA more
by SMOs, but also in how they seek to accomplish positive. Thus:
those goals. Managers may find some SMOs eager to Hypothesis 2. SMO resources in a jurisdiction mod-
work cooperatively with the firm and others working erate the relationship between a firm’s magnitude of
to impose regulatory burdens on the firm. dependence on that jurisdiction and its corporate
2021 Sutton, Devine, Lamont, and Holmes, Jr. 45
political activity in the jurisdiction, such that an conversations with one another, and shape one an-
increase in SMO resources strengthens the positive other’s views. In this way, they engage in collective
relationship between dependence and corporate sensemaking and develop an approximately shared
political activity. interpretation and understanding of the firm’s op-
tions for managing its dependencies (Smircich &
Stubbart, 1985). Sensemaking and enactment are
Response Uncertainty
fundamentally concerned with managing uncer-
“Response uncertainty” is a lack of understanding tainty (Weick et al., 2005), so any disruption to the
about what response options exist, or about the TMT’s sensemaking process should increase re-
consequences of response options, for addressing sponse uncertainty experienced by the team.
conditions in the environment (Milliken, 1987: 137). TMT turnover disrupts the sensemaking process
Response uncertainty could stem from a lack of in- underlying response formulation in several ways.
formation about the response options or from the First, TMT turnover may dislodge extant routines
inability to quickly process information about them. (Nystrom & Starbuck, 1984), including methods used
In the context of CPA, response uncertainty exists to gather and share information, and can reorient
when either the source (i.e., TMT members) or the managers’ assessments of the decision options they
objects (i.e., politicians) of firms’ CPA strategies confront (Lant, Milliken, & Batra, 1992). In this way,
changes. These changes limit understanding of the it can create uncertainty about the adequacy of
available responses and their consequences by dis- existing response repertoires. Second, the presence
rupting information processing and managers’ confi- of new TMT members is likely to generate more in-
dence that they can influence politicians effectively. depth discussion of response options due to the new
Thus, TMT and politician turnover are two salient members’ different experiences and preferences
sources of response uncertainty in this context. (Melone, 1994) and due to the disruption of inter-
As with state uncertainty, response uncertainty is personal dynamics within the group (Wiersema &
unlikely to affect CPA directly. Response uncer- Bantel, 1992), both of which can make consensus
tainty is meaningless in the absence of a resource more elusive (Carpenter, 2002). By disrupting the
dependency that requires a response. In other words, sensemaking and enactment process and disconnect-
the turnover of TMT members or politicians is un- ing it from extant and accepted response repertoires,
likely to shape firms’ CPA in a particular jurisdiction TMT turnover creates response uncertainty for firms
if they have little to no presence in that jurisdiction. attempting to manage their dependencies. More spe-
However, response uncertainty can shape the cifically, TMT turnover can increase the perceived
effect of dependence on CPA. Responding to de- complexity of CPA options and raise doubts about the
pendence on a political jurisdiction requires under- efficacy of these options. Therefore, TMT turnover
standing the available response options and their inhibits a firm’s ability to respond to political depen-
likely effectiveness, and response uncertainty disrupts dencies with CPA expeditiously.
that knowledge. Response uncertainty, therefore, in- In this regard, one of our interviewees told us,
terferes with the response itself, which weakens the “New players [TMT members] don’t know how to
relationship between dependence and CPA. Firms play the game. They tend to be reluctant to have the
experiencing low response uncertainty, by contrast, firm engage with politicians.” This interviewee also
are better able to adjust their CPA to manage their explained that experience allows TMTs to appreci-
dependence as needed. ate and respond to their firms’ political dependen-
TMT turnover. TMT members are important in cies more aggressively. When there is less TMT
the enactment process, because they are the indi- turnover, in other words, the TMTs coalesce into a
viduals who must formulate responses. Accordingly, shared understanding about the importance of man-
the TMT is responsible for making CPA decisions aging their dependence on particular jurisdictions
(Hadani, Dahan, & Doh, 2015; Hillman & Hitt, 1999; proactively and about how to do so effectively.
Ozer, 2010). “TMT turnover” creates response un- Thus, when TMT turnover is low, firms are better
certainty by disrupting sensemaking and enactment able to match their CPA with the magnitude of de-
related to response formulation. Although individ- pendence their firms confront in various jurisdictions.
ual TMT members analyze the environment in their High dependence makes it important for firms to
own unique ways, group processes help form an shape the legal and regulatory environment in their
overall organizational interpretation (Daft & Weick, favor; and, with an intact TMT, they are able to quickly
1984). Executives gather and share information, have and appropriately develop political responses to the
46 Academy of Management Journal February
high dependence. High TMT turnover, by contrast, One of our interviewees described this situation by
makes response options more uncertain and pre- saying that politicians have become more cautious
vents a rapid reformulation of how to use CPA most when interacting with donors, especially early in
effectively in high dependence jurisdictions. The their tenures. Explaining why his firm was cautious
high response uncertainty associated with high with new legislators, another executive said, “If you
TMT turnover results in lower CPA in jurisdictions say something to the wrong person behind closed
of high dependence than would be expected when doors, it won’t stay behind closed doors for long.” By
TMTs are stable. At low dependence on jurisdic- contrast, managers have had more time to under-
tions, little CPA response is required, making TMT stand and engage with longer-tenured politicians.
stability or turnover largely ineffectual. Thus, the Thus, high political turnover increases response
primary effect of TMT turnover on the relationship uncertainty not only because there is less informa-
between dependence and CPA occurs in the high tion about the priorities and effectiveness of new
dependence situations, where it suppresses firms’ politicians, but also because it might be unclear how
use of CPA to manage its dependencies. Thus: to influence the new politicians effectively. In this
way, politician turnover in a jurisdiction increases
Hypothesis 3. TMT turnover moderates the relation-
uncertainty about how to respond to political de-
ship between the magnitude of dependence and cor-
pendencies with CPA.
porate political activity, such that an increase in TMT
turnover weakens the positive relationship between
Therefore, as with TMT turnover and the response
dependence and corporate political activity. uncertainty it creates, we expect politician turnover
to mainly temper firms’ use of CPA in jurisdictions of
Politician turnover. “Politician turnover” refers higher dependence. At low dependence in jurisdic-
to changes in the executive and legislative branches tions, firms have little need to use CPA regardless of
in a jurisdiction, usually when politicians retire the degree of politician turnover in the jurisdiction.
or lose re-election. Politicians are objects of the Conversely, high dependence motivates greater CPA,
enactment process (i.e., the entities TMTs seek to and low politician turnover enables it. Low politician
understand and influence) and are key to firms’ ef- turnover allows the firm to vigorously manage their
forts to manage their environments. We argue that dependence by engaging with politicians in ways
politician turnover also increases response un- managers believe to be effective. Conversely, high
certainty by disrupting environmental enactment politician turnover hampers firms’ environmental
processes, weakening the link between firms’ de- enactment and their abilities to respond, making
pendencies on jurisdictions and their CPA in those them less willing to use CPA to manage their high
jurisdictions. dependence than they otherwise would have been.
Politicians’ tenure is important to donors, as sug-
Hypothesis 4. Politician turnover in a jurisdiction
gested by the large fundraising advantage enjoyed by
moderates the relationship between the magnitude of
incumbent politicians (Hillman et al., 2004; Krasno,
a firm’s dependence on that jurisdiction and its cor-
Green, & Cowden, 1994). This politician-level obser- porate political activity in that jurisdiction, such that
vation is related to donor-level uncertainties about an increase in politician turnover weakens the posi-
new politicians. New politicians may be quite vocal tive relationship between dependence and corporate
about an issue, but only time will reveal how much political activity.
effort they put into advancing their positions (Kroszner
& Stratmann, 2005) and how effective they will be
(Miquel & Snyder, 2006). METHODS
Firms also lack information about new politicians’
Research Setting
susceptibility to influence and preferred methods
of interaction. This information is important, because We tested our hypotheses using the political con-
even subtle differences in donors’ communication can tributions made by U.S.-based, publicly traded firms in
affect their access to politicians (Kalla & Broockman, elections for state governors and legislators from 2009
2016). Some politicians and regulators are reluctant to 2016 across the 50 U.S. states. This context is ap-
to engage with certain firms (McDonnell & Werner, propriate for several reasons. First, many large Amer-
2016), especially those they view as illegitimate (Hiatt ican firms operate in multiple states, creating an
& Park, 2013; Werner, 2015). In turn, managers need identifiable and quantifiable set of firm dependencies
time to assess how new politicians view, and desire to on state governments (i.e., jurisdictions). Second, po-
interact with, their firms. litical contributions are an important form of CPA.
2021 Sutton, Devine, Lamont, and Holmes, Jr. 47
Third, contributions can be adjusted in small in- mentioned state-level environmental regulatory
crements, allowing firms to vary their political concerns but did not specify state-level concerns
contributions from one state to another and from one for other types of regulations (Tyson Foods, 2016).
election cycle to the next as the firms and their en- As another example, environmental regulation was
vironments change. Also, in using 2009–2016 election the only state-level regulatory concern mentioned
data, we captured two presidential election cycles in Item 1a of Lockheed Martin’s 10-K (Lockheed
(i.e., 2011–2012 and 2015–2016, with few guberna- Martin, 2013). Finally, because of their environmen-
torial elections) and two midterm election cycles (i.e., tally intensive operations, these firms are consistently
2009–2010 and 2013–2014).4 This setting allows us to scrutinized by media and SMOs (Baron, 2003; Eesley &
assess how managers allocate their firms’ political Lenox, 2006).
contributions to manage the firms’ dependencies and To test the hypotheses, data were structured in a
to deal with the various sources of uncertainty they cross-sectional time-series data set consisting of firm–
face. state observations (e.g., Chevron–Texas, Chevron–
The sample contained the 65 largest, by revenue, Louisiana, etc.). Every firm–state combination in
publicly traded firms that are required to report which a firm had an environmentally intensive
chemical releases to the United States Environmental presence was included for each year (i.e., 65 firms 3
Protection Agency. We used the 65 largest firms be- number of states in which a firm operated 3 8 years),
cause the sample needed to contain firms that operate resulting in 5,990 observations across our observa-
in multiple states, and many smaller firms do not. tion period after accounting for missing data.
Furthermore, reliable state-level political contribution
data were not available for firms smaller than those in
Dependent Variable
our sample, so we restricted our sample to those firms
for which we could gather reliable data. These firms Corporate political activity. Corporate political
operate in the chemical, oil and gas, energy, and activity was measured as the total donations to can-
manufacturing industries. Table 1 presents the distri- didates or political parties in a state by a firm, and
bution of the sample firms across the three-digit North these data were collected from the National Institute
American Industry Classification System (NAICS) on Money in State Politics. We included contribu-
codes associated with these industries (United States tions directly from the firm and contributions from
Environmental Protection Agency, 2013). These in- its political action committee (PAC), which are
dustries were chosen because the firms in them en- commonly used in studies of political action (Hadani
gage in practices that are heavily scrutinized by & Schuler, 2013; Schuler, 1999).
government officials and face similar public policy
challenges and concerns. Specifically, each firm
Independent Variables
must file reports with the Environmental Protection
Agency because the nature of the firms’ operations is Dependence. Managers enact dependence by fo-
environmentally intensive. cusing their attention and actions on aspects of the
We found consistent evidence from the 10-Ks of environment they consider to be the most problematic
firms in our sample industries that environmental reg- (Pfeffer & Salancik, 1978). Therefore, our measure
ulations are the primary source of state government- of dependence on government needed to reflect
related priorities and uncertainty for firms in these enacted dependence, as indicated by the environ-
industries. For example, in their 2016 10-K, Tyson mental elements that capture managers’ attention. As
Foods mentioned environmental regulations before we explained above, managers of firms in our sample
mentioning other types of regulations, and they were focused on environmental laws and regulations
did so in Item 1, which is a general description of as the primary source of government-related uncer-
the firm’s business. In other words, environmental tainty. Toxic releases reflect the environmental in-
regulatory concerns were a major part of their tensity of a firm’s operations in a state and, as a result,
business. Furthermore, Tyson Foods specifically its exposure to government action in that state. Sec-
ond, and similarly, managers pay particular attention
4
Our models predict yearly firm–state contributions in to data on toxic chemical releases, partly because
both election years and non-election years. Since most government regulators, the media, and SMOs attend to
contributions are made in election years, we also tested these data (Toffel & Marshall, 2004).
models that focus only on election-year contributions and To construct the measure, we collected toxic chem-
found results consistent with those reported below. ical release data from the Environmental Protection
48 Academy of Management Journal February
TABLE 1
Sample Firm Distribution across Industries
NAICS Code NAICS Industry Group Definition No. of Firms
Agency’s Toxic Release Inventory database (United helps establish comparability with prior work; (3) a
States Environmental Protection Agency, 2019), single source avoids duplication of articles and is
which is a collection of facility-level chemical re- likely to pick up on events that resonate generally
lease data that all firms in the sample industries are with social concerns (McCarthy, McPhail, & Smith,
required to report annually. We then aggregated the 1996); and (4), because the WSJ focuses on business
data at the firm–state level and calculated depen- and is widely read by executives, it likely has a strong
dence as a firm’s total pounds of toxic chemical re- influence on managers’ impressions of their firm’s
leases in a state divided by the firm’s total pounds of media coverage.
chemical releases in all states. Local media tenor. We used the NewsBank data-
National media tenor. Following prior research base to search a major newspaper for each state (e.g.,
on the media’s role in influencing social evaluations the Tampa Bay Times) that was more state oriented
(Bansal & Clelland, 2004; Pollock & Rindova, 2003), rather than nationally oriented in scope. The local
we used the Janis–Fadner coefficient of imbalance media tenor measure was constructed using the same
(Janis & Fadner, 1965) to measure the national media method that was employed for national media tenor.
tenor of each firm in a given year. We coded an article SMO resources. The ability of social movements
as positive (negative) if 80% or more of its content to mobilize and affect change is driven by their ability
was positive (negative). For the content analysis, we to marshal resources, including money (McCarthy &
used the Linguistic Inquiry and Word Count soft- Zald, 1977). Because financial resources signal pub-
ware, which has been extensively used in the liter- lic support and enable SMOs to fund public rela-
ature. For ease of interpretation, we reverse-coded tions campaigns and lobbying, SMO income is a good
this measure so that higher values reflected more indicator of influence. Moreover, since many SMO
negative tenor. effects are inherently local (Schneiberg & Bartley,
Because the sampled firms were concerned with 2001), we focused on environmental nonprofits
environmental legal and regulatory changes, we within each state (Pacheco & Dean, 2015; Pacheco,
searched The Wall Street Journal (WSJ) for articles York, & Hargrave, 2014). To measure SMO resources,
about each firm using keywords such as “environ- we scaled the income received by environmental
ment,” “pollution,” and “habitat destruction.” Reli- SMOs headquartered in a state by the state’s pop-
ance on the WSJ is preferable for four reasons: (1) it is ulation. These data came from the National Center for
a prominent national news source; (2) this approach Charitable Statistics.
2021 Sutton, Devine, Lamont, and Holmes, Jr. 49
TMT turnover. To capture TMT turnover, we di- Standard & Poor’s Compustat database. Since larger
vided the number of executives who left each year by firms may have more resources for CPA (e.g., Schuler
the number of TMT members at the beginning of that et al., 2002), we measured firm size as the log trans-
year (Walsh, 1988). We did not correct for normal formation of firm assets. We controlled for interna-
retirement, because our theory does not differentiate tionalization, measured as the percentage of a firm’s
among types of turnover. TMT turnover is likely to revenues earned outside of the United States, because
produce response uncertainty, regardless of whether firms with a larger international presence might be
the turnover was due to retirement, attrition, or ter- less dependent on domestic operations. Additionally,
mination of employment. We used ExecuComp to because a firm’s performance outlook may shape its
gather these data. political contributions, we controlled for firm per-
Politician turnover. Both governors and legisla- formance using Tobin’s q (Bertrand & Schoar, 2003).
tors exert considerable power over a state’s legal and
regulatory environment, and, in turn, their turnover
Analysis
is noted by managers. To measure politician turn-
over, we created measures for gubernatorial and leg- To test our hypotheses, we used linear regression
islative turnover by dividing the number of newly with fixed effects for firm, state, and time. The re-
elected officials by the total number of each state’s gressions were further specified with robust stan-
offices. Politician turnover was calculated as the dard errors. Taken together, this approach provides a
simple average of gubernatorial and legislative turn- conservative test of the hypotheses. Across our ana-
over. These data were acquired from The Book of the lyses, all independent and control variables pre-
States, published by the Council of State Govern- ceded our dependent variable to establish temporal
ments and available from the Council’s Knowledge spacing.5
Center website (http://knowledgecenter.csg.org/kc/
category/content-type/content-type/book-states).
RESULTS
Control variables. We accounted for the possibility
of alternative influences on CPA in several ways. First, Table 2 provides summary statistics regarding the
we incorporated fixed effects for firm, state, and time, distribution of firms’ presence and CPA across states.
to control for time-invariant firm and state factors as The firms operated in an average of 11.77 states and
well as temporal factors that may influence CPA. they made political contributions in an average of
We also included other control variables. Political 4.03 of the states they operated in each year. Table 3
ideology might affect CPA as the Republican Party is presents the descriptive statistics and correlations
often considered to be more sympathetic to business for each variable. The sampled firms made contri-
than is the Democratic Party (Hersch & McDougall, butions of $11.9 million in the 2010 election cycle
2000). For each state, Republican was a dichotomous (i.e., 2009 and 2010 contributions), $11.1 million in
variable, with a value of 1 if the Republican Party the 2012 cycle, $13.1 million in the 2014 cycle, and
controlled the legislature and governor’s office in a $12.5 million in the 2016 cycle.6
state, and a value of 0 otherwise. Data for this control Table 4 presents the regression results. Baseline
were obtained from The Book of the States. To con- models including only the controls and independent
trol for the possibility that firms may have a unique variables are presented in Models 1 and 2. As ex-
dependence relationship with the state in which pected, the main effect of dependence was positive
they have their headquarters (HQ), we coded HQ and significant in Model 2 and in all subsequent
state with 1 if a firm had its HQ location in a state and models. This finding provides support for our base-
0 otherwise. Since each state does not hold elections line prediction, which is also in line with the re-
each year, we included a binary variable that indi- source dependence literature.
cated if it was an election year. Models 3 to 8 add the interaction terms to test
Because media coverage of industry events might Hypotheses 1 to 4. To improve interpretation, we
also be important, we controlled for industry media
tenor. To do so, we searched the WSJ for articles 5
Except for the election cycle independent variables (e.g.,
about the 50 largest firms by revenue for each of the politician turnover) that were lagged to the prior election
three-digit NAICS codes in Table 1. The industry cycle and the election year indicator, the independent and
media tenor measure was constructed using the control variables were lagged by one year (t 2 1).
same method described for national media tenor. We 6
These totals include only states in which firms had an
also controlled for other firm characteristics using environmentally intensive presence.
50 Academy of Management Journal February
TABLE 2
Distribution of Firms’ Environmentally Intensive Presence and CPA across States
Mean Median Min. Max. 1st Quartile 3rd Quartile
include graphs of the hypothesized interactions in (b 5 20.093, p , .001). This relationship is illus-
Figure 1, which displays the effect of dependence at trated in Figure 1c. The results reveal that TMT
low and high levels (21 and 11 standard deviation, turnover weakens the effect of dependence on CPA.
hereafter “SD”) of the respective moderators. The marginal effect of dependence on CPA is 71%
Hypothesis 1a predicted that negative national less when TMT turnover is high (11 SD) compared
media tenor would strengthen the relationship be- to when it is low (21 SD). Therefore, the results
tween firms’ dependence on a state and their CPA in support Hypothesis 3.
that state. The coefficient was positive and signifi- Finally, Hypothesis 4 predicted that politician
cant (b 5 0.037, p , .01) in Model 3 and also Model 8 turnover would weaken the dependence–CPA rela-
(b 5 0.025, p , .05). Figure 1a demonstrates that the tionship. The relevant interaction terms in Model 7
effect of dependence on CPA is stronger when na- (b 5 20.064, p , .05) and Model 8 (b 5 20.072,
tional media tenor is more negative than when it is p , .05) are negative and significant. The graph of
less negative. For example, the marginal effect of this effect is depicted in Figure 1d, in which the
dependence on CPA is 70% higher when negative relationship between dependence and CPA is shown
national media is high (11 SD) compared to when to weaken as politician turnover increased. In par-
negative national media is low (21 SD). Together, ticular, the marginal effect of dependence on CPA is
these results support Hypothesis 1a. 63% less when politician turnover is high (11 SD)
Hypothesis 1b predicted that negative local media rather than low (21 SD). Thus, the results lend sup-
tenor would strengthen the relationship between a port to Hypothesis 4.
firm’s dependence and its use of CPA. However, the
coefficients for the interaction terms were non-
Manager Interviews
significant in both Model 4 (b 5 20.081, p 5 .161)
and Model 8 (b 5 20.081, p 5 .140). As such, Hy- To further understand the relationship between
pothesis 1b is not supported. dependence, uncertainty, and CPA, we gathered
Hypothesis 2 predicted that the relationship be- qualitative data through a series of interviews with
tween dependence and CPA would strengthen as individuals directly involved in CPA. The inter-
oppositional SMO resources in a state increased. The viewees were contacted through a combination of
coefficients for the interaction terms were positive “cold contact” emails sent to TMT members of our
and significant in both Model 5 (b 5 0.166, p , .05) sample firms, as well as contacts made through the
and Model 8 (b 5 0.190, p , .05). In addition, authors’ professional networks. We spoke with 16
Figure 1b shows that the relationship between de- interviewees in total. Eight of the interviewees were
pendence on a state and the firm’s CPA in the state is senior executives in highly regulated firms, four of
stronger when SMO resources are larger. For exam- which were firms in our sample. Six of the inter-
ple, the marginal effect of dependence on CPA is viewees were political consultants, and two were
102% larger when SMO resources are high (11 SD) current or former government officials. To under-
rather than at the mean. We state the effect relative to stand the sources and role of state uncertainty, we
the mean since the same increase in dependence asked interviewees what factors affect the environ-
results in a near zero decrease in CPA when SMO mental legal and regulatory uncertainty within a
resources are low (21 SD). Accordingly, the results state, as well as whether they had dedicated per-
support Hypothesis 2. sonnel or systems in place to monitor those factors.
Hypothesis 3 predicted that TMT turnover weakens To understand the sources and role of response un-
the relationship between dependence and CPA. The certainty, we asked how TMT turnover and politi-
relevant interaction terms are negative and signifi- cian turnover affected uncertainty and the use of
cant in Model 6 (b 5 20.078, p , .01) and Model 8 political activity.
2021
TABLE 3
Descriptive Statistics and Correlationsa
Mean SD 1 2 3 4 5 6 7 8 9 10 11 12 13
14 Republican 0.46 0.50 2.03 .04 .03 .02 2.30 .02 .00 .01 2.01 2.02 .04 2.01 .00
Note: Standard errors in parentheses. Hypothesis tests are one tailed. Firm, state, and year fixed effects are included in all models.
†
p , .10
*p , .05
February
**p , .01
***p , .001
2021
FIGURE 1
The Effects of Dependence on Corporate Political Activity at Low and High Values of (a) Media Tenor, (b) SMO Resources, (c) TMT Turnover,
and (d) Politician Turnover
1.4
A B
1.2
1.5
1
CPA
CPA
1
0.8
0.6
0.5
0 +1 SD +2 SD +3 SD +4 SD 0 +1 SD +2 SD +3 SD +4 SD
Dependence Dependence
C Media Tenor = –1 SD Media Tenor = +1 SD
D SMO Resources = –1 SD SMO Resources = +1 SD
1.4
1.4
1.2
1.2
Sutton, Devine, Lamont, and Holmes, Jr.
1
1
CPA
CPA
0.8
0.8
0.6
0.6
0 +1 SD +2 SD +3 SD +4 SD 0 +1 SD +2 SD +3 SD +4 SD
Dependence Dependence
TMT Turnover = –1 SD TMT Turnover = +1 SD Politician Turnover = –1 SD Politician Turnover = +1 SD
53
54 Academy of Management Journal February
The data from the interviews enliven and illustrate whether firms’ state location decisions were endog-
the hypothesized relationships. First, one interviewee enous. First, we estimated a Heckman selection
claimed that heavily regulated firms that fail to engage model wherein state presence was predicted that in-
in CPA are committing “corporate malpractice.” The cluded all of our variables as well as firm, state, and
hyperbole notwithstanding, this individual clearly time fixed effects (Heckman, 1976). As an instrument,
viewed CPA as a critical aspect of strategy in firms that we used the industry average of total pounds of toxic
are more dependent on government. In addition, the chemical releases in a state, since it is likely to predict
interviewees confirmed that they are highly concerned firm dependence but it is not endogenous to firm out-
about media coverage and influential SMOs and that comes.7 Consistent results were found. Second, we
they have mechanisms to monitor both. One TMT used a control function approach whereby we esti-
member said that some of the monitoring is performed mated a first-stage linear regression (predicting de-
by firm employees and some is outsourced. Indeed, pendence) using the same instrumental variable(s) and
there is a small industry of professional services firms, then implemented the residuals into the second-stage
including some public relations firms, that provide this models as a control variable (Wooldridge, 2010). In line
type of monitoring, as evidenced by the information with best practices, all stages were bootstrapped (1,000
publicly available on these firms’ websites. We con- replications) and consistent results were obtained.
sistently heard that the reason for this monitoring and We also conducted several additional analyses
concern about the media and social movements is that to further explore our theory and assess the robust-
they can trigger legal and regulatory change. ness of the results. First, we considered an alterna-
We also heard from our interviewees that turnover tive operationalization of the dependent variable.
among TMT members and politicians can disrupt the Although our theory is about how managers perceive
political engagement process. An interviewee said and respond to uncertainty, political scientists have
that CPA is disrupted following TMT turnover be- suggested that employees are important components
cause the “new players are often reluctant to see the of firms’ political influence (Fellowes & Wolfe, 2004;
need for political engagement.” This lack of under- Sorauf, 1992). Thus, we tested the hypotheses with a
standing by new members of the TMT was repeated by political contributions measure that also included
several interviewees who described TMT disruption contributions made by a firm’s employees. The find-
as an event that required significant education. As ings were consistent with those above, as displayed in
one interviewee said, “management turnover causes Table 5.
companies to have to reinvigorate the TMT about why Second, following recent research (Hubbard, Pollock,
to pursue political engagement.” Likewise, turnover in Pfarrer, & Rindova, 2018), we tested the sensitivity
the legislature led to what interviewees described as “a of our results for Hypotheses 1a and 1b to different
recalibration,” or a time to gather information before operationalizations (i.e., at 70%, 75%, 85%, and 90%
engaging new legislators. For instance, one inter- thresholds to define negative, positive, and neutral
viewee said that “firms are confused about how to articles) of media tenor. For national media tenor,
engage after political turnover.” These comments thresholds starting at 75% and greater were consistent
suggest that turnover among TMT members and pol- with reported results. For local media tenor, the re-
iticians affects dependence management not just with spective thresholds were non-significant, with the
donations, but also with political engagement more exception of the 75%, which was negative and mar-
broadly. In each case, the interviewees were speaking ginally significant in the model that included firm
about those contexts in which their firms were de- and employee contributions in the dependent vari-
pendent on a state. In sum, the interviewees’ responses able. Accordingly, results were largely consistent
consistently suggested that negative media coverage with those reported in primary analyses.
and SMOs create a type of uncertainty that must be
countered with action in those states where the firm
has dependencies, whereas TMT turnover and politi- 7
The use of industry averages as instruments is common
cian turnover create a type of uncertainty that inhibits
practice in finance and economics (Ferrell, Liang, &
the dependence management via CPA.
Renneboog, 2016; Lin, Ma, Malatesta, & Xuan, 2011), and
has also been used extensively in management (Shi,
Stability and Robustness of Results Connelly, & Sanders, 2016; Su & Tsang, 2015). As an al-
ternative, we used industry dependence on a state as a
Although the fixed effects greatly mitigate en- percentage of its dependence across states—and found
dogeneity-related concerns, we further considered similar results.
TABLE 5
2021
Regressions Predicting CPA (Firm 1 Employee Contributions)
Variable Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8
Note: Standard errors in parentheses. Hypothesis tests are one tailed. Firm, state, and year fixed effects are included in all models.
†
p , .10
*p , .05
55
**p , .01
***p , .001
56 Academy of Management Journal February
Third, we considered several alternative mea- size, performance, and Republicanism. Because us-
sures of SMO resources. We first controlled for SMO ing the various sets of alternatives and controls left
concentration in each state, using a Herfindahl– the conclusions unchanged, they were excluded
Hirschman index based on SMO income. In addi- from primary analyses due to either non-significance
tion, we scaled our SMO measure by the number of or collinearity with the fixed effects. We also esti-
manufacturing sites in a state. We also created a mated models that included state-by-time dummy
measure of SMO resources using SMO assets instead variable interactions and found consistent support.
of income. Lastly, we scaled the count of environ- Finally, we ran regressions excluding all control
mental non-profits in a state-by-state population. variables and obtained consistent findings.
The results were all consistent and offered further
support for Hypothesis 2.
DISCUSSION
Fourth, we considered alternative measures of
TMT turnover. Specifically, we examined TMT change This study examined how firms manage their re-
(i.e., sum of members who left and new members source dependencies, using the context of firm de-
who were added, with this sum divided by the pendence on U.S. state governments and drawing on
number of TMT members at the beginning of each RDT. The study included two rarely used aspects
period; Cho & Shen, 2007), whether a company’s of RDT—Pfeffer and Salancik’s (1978) distinction
current or former CEO left the firm, and whether a new between dependence and uncertainty and their view
CEO took the helm. Except for the new CEO variable, of the environmental enactment process—and one
which was non-significant, these measures produced less commonly researched dependence management
consistent results and provided additional support for strategy (CPA). While ample research has documented
Hypothesis 3. the positive relationship between dependence and
Fifth, we probed the stability of the results for pol- CPA (Hillman et al., 2009; Lux at al., 2011), we used
itician turnover (Hypothesis 4). To do so, we created a the distinction between dependence and uncertainty
politician turnover measure that gave more weight to to posit that uncertainty moderates this relationship.
gubernatorial turnover (2/3 weight) and another And, we used the enactment process to identify im-
measure that gave more weight to legislative turnover portant types and sources of uncertainty confronting
(2/3). We also constructed a proportional measure of the firms in our research context.
legislature turnover as well as a binary indicator for Beyond reinvigorating nuance to the typical view of
gubernatorial turnover. The results were consistent dependence and its management in the current RDT
across these measures. We also considered whether literature, however, we added a novel extension to
firms directed contributions to parties (i.e., not can- RDT by drawing on the work of Milliken (1987). We
didates) when there was politician turnover. Ac- argued that two types of uncertainty have opposite ef-
cordingly, we tested an alternative measure of the fects on the actions firms use to manage their depen-
dependent variable that included only contributions dence on government. First, we argued that state
to parties and found consistent results. uncertainty increases managers’ concerns about their
Lastly, we controlled for the following state char- firm’s dependence, thus strengthening the relationship
acteristics: political contribution limits, population, between resource dependence and CPA. Because
state-on-industry dependence, regulatory intensity, managers and their firms cannot attend to and manage
election competitiveness,8 legislature swing, Re- every aspect of the environment, they place more em-
publican advantage, and controlling party years in phasis on those factors that have historically caused
power. We also considered other time-varying fac- more significant disruptions during the enactment
tors, such as presidential election year and revolving process. In this regard, the national media and oppo-
door lobbyists. Furthermore, we explored alternative sitional SMOs play a central role in influencing state
measures for some of the controls, including firm uncertainty as it relates to dependence on political ju-
risdictions for firms in environmentally intensive
8 industries.
For each election, we computed the inverse of the
We also explained how response uncertainty
Herfindahl index from the votes cast for each candidate
in a single election. We evaluated the index at different weakens the relationship between a firm’s depen-
thresholds and assessed competitive elections in absolute dence on a political jurisdiction and its CPA in that
terms and as a proportion of a state’s total elections. These jurisdiction. More precisely, we argued that turnover
data were obtained from the Harvard Dataverse repository among TMT members (i.e., the subjects of enactment,
(Klarner, 2018). those engaged in the enactment process) or among
2021 Sutton, Devine, Lamont, and Holmes, Jr. 57
state politicians (i.e., the objects of enactment, those uncertainty moderates the effects of dependence on
whom the subjects of enactment are looking to influ- response. These results extend those of Bode, Wagner,
ence) increases response uncertainty by disrupting Petersen, and Ellram (2011), who found similar results
firms’ enactment processes and disconnecting enact- for response uncertainty in a very different setting, but
ment from response. did not examine state uncertainty. In making this
We tested our hypotheses using state-level politi- distinction, we also explain how the enactment pro-
cal contributions over eight years in all 50 U.S. states. cess, which was central to the original articulation of
Consistent with prior work (Hillman et al., 2004; Lux RDT but has received relatively little attention, affects
et al., 2011), we found support for our baseline pre- how managers experience state and response uncer-
diction of a positive relationship between the mag- tainty and how they manage dependencies.
nitude of a firm’s dependence on a jurisdiction and Second, we add value and nuance to RDT by estab-
its CPA in the jurisdiction. Importantly, the results lishing theoretically and empirically that the factors
also provide strong support for the hypothesized driving managerial and firm attention and action are
moderation effects of the different types of uncer- both local (e.g., SMO opposition and politician turn-
tainty. Regarding state uncertainty, negative national over) and more firm-wide or global (e.g., national me-
media coverage and oppositional SMO resources dia tenor and TMT turnover). These local and global
strengthen the relationship between dependence factors, which shape the enactment process, are not
and CPA. For response uncertainty, TMT turnover unique to our setting. Identifying and distinguishing
and politician turnover weakened the relationship these factors adds novel nuance to RDT’s concept of
between dependence and CPA. enactment and can inform future work.
Notably, however, the hypothesis that local media Third, our focus on CPA, uncertainty, and enact-
tenor would moderate the relationship between de- ment advances the RDT literature if not the theory
pendence and CPA was not supported. There are at itself. Hillman et al. (2009: 1411) noted that Pfeffer
least two likely reasons for this. First, the TMT is cen- and Salancik’s (1978) discussion of political action
trally involved in the CPA decision-making process is critically important yet has been the “the most
(Hadani et al., 2015; Ozer, 2010), and, due to their commonly overlooked” aspect of RDT, both theoret-
limited time, they are more likely to read national ically and empirically. Our explanation of situational
media outlets like the WSJ than numerous local papers. factors that shape managerial and firm attention and
Therefore, the TMT is more likely to be influenced by action, therefore, advances RDT research on this
national media. Second, because the national media rather neglected strategy for managing dependence.
tends to have a larger and more geographically diverse We also make two contributions to the CPA liter-
readership than local papers, negative coverage in ature. With few exceptions (e.g., Holburn & Vanden
national publications like the WSJ poses a threat of a Bergh, 2014), this literature has focused on federal or
political contagion that local media outlets do not international political activity to the exclusion of
pose. The firm’s misbehavior in one state may prompt state-level activity (e.g., Hansen & Mitchell, 2000;
government action in other states if national media Hart, 2001; Schuler et al., 2002). We are among the
covers the story, as the DuPont example (see above) first to examine (1) the antecedents of firm political
illustrates. contributions across multiple jurisdictions and (2)
the relationship between state-level dependencies
and contributions to state politicians. Because state-
Contributions
level politics have important effects on firms, as
This paper makes three contributions to the litera- Tesla’s challenges with dealer franchise laws illus-
ture on RDT. Our primary theoretical contribution lies trate (“Electric car manufacturer,” 2016), more at-
in integrating Milliken’s (1987) uncertainty typology tention to state-level political strategies is needed.
with RDT to explain how state uncertainty and re- Lastly, we explain the relationship between TMT
sponse uncertainty interact with dependence but have turnover and CPA. Dependence on government
opposite effects on dependence management. Uncer- leads firms to adopt long-term relational approaches
tainty is a central element in RDT, but much prior work to CPA (Hillman & Hitt, 1999), which suggests that
has assumed that dependence and uncertainty covary stable political bodies are important (Lux et al.,
and that responses increase with the magnitude of the 2011). However, extending this idea, we explain how
dependence (Hillman et al., 2009). Drawing on a subtle TMT stability also matters (but for different reasons).
insight of Pfeffer and Salancik (1978), we distinguish In turn, we provide a more nuanced description of
between dependence and uncertainty and argue that why and how some of the critical players in CPA
58 Academy of Management Journal February
(TMT members and politicians) shape the relation- TMT change within either firm, but especially the buyer
ship between dependence and CPA. in this case, could increase both state and response un-
Our results also hold an important implication for certainty. Conversely, the diffuse control in government
managers. It seems that response uncertainty, partic- makes this sort of relationship between state and re-
ularly associated with TMT turnover, may need to be sponse uncertainty less likely in that context.
explicitly managed. If CPA is useful but firms do not Scholars also should consider other forms of CPA.
use it due to response uncertainty, then response un- Whereas we examined political contributions, lob-
certainty could compromise firms’ success in those bying also is important. Schuler et al. (2002) revealed
jurisdictions where they are dependent. Those firms that political contributions and lobbying at the federal
would be more vulnerable to government action and to level are used together much, but not all, of the time.
competitors with stable TMTs that use CPA more ag- Thus, future research should consider how our results
gressively. Managers would be advised to quickly re- generalize to other political activities. This could be
solve response uncertainty within their firms to ensure empirically challenging, however, because states vary
clarity during managerial and political upheaval. considerably in their lobbying disclosure require-
ments. Nonetheless, contributions remain an impor-
tant form of political influence (Holburn & Vanden
Limitations and Future Research
Bergh, 2014), perhaps especially so at the state level.
There is no reason to believe that our results are State legislative campaigns typically receive less
restricted to CPA, as the Bode et al. (2011) findings funding than campaigns for federal office (Campaign
tentatively support. High levels of state uncertainty, Finance Institute, 2016; National Institute on Money
where the future of the resource exchange is unpre- in State Politics, 2012), so state-level political contri-
dictable, are likely to amplify the necessity of using butions might have greater electoral influence than
mergers, joint ventures, board member selection, and those at the federal level.
other strategies to manage dependencies. The mod-
erating effect of state uncertainty might be relevant in CONCLUSION
contexts in which different firms’ management teams
interact frequently over long periods of time to man- Since the 1960s, CPA scholars have justified their
age dependencies, likely driving down state uncer- work, in part, by explaining how pervasive CPA has
tainty as the interactions bring familiarity. In these become (Epstein, 1969; Pittman, 1977). The present
contexts, partners’ decision-making power is concen- study makes theoretical and empirical contributions
trated in a relatively small group, compared with polit- to understanding where, why, and when firms use
ical dependencies wherein decision-making power political contributions. The theory we advance holds
tends to be more diffuse. Therefore, relationships may the potential to serve as a springboard for future CPA
serve a more important role in driving down state un- studies that model the determinants of CPA and
certainty in interfirm dependencies. Additionally, re- perhaps the consequences of CPA. These steps are
sponse uncertainty, whereby firms cannot decipher how critical, especially in light of relatively recent devel-
to respond effectively, is likely to curtail the use of other opments in campaign finance policy and the corre-
strategies to manage dependencies as well. The sources sponding increase in the scale and scope of CPA.
of uncertainty also are likely to vary across contexts, and
considering the enactment process should help identify
the important sources of uncertainty in each context. REFERENCES
The research opportunities appear numerous.
It is also worth considering how state and response Aldrich, H. E., & Pfeffer, J. 1976. Environments of organi-
uncertainty may be related. The two might be con- zations. Annual Review of Sociology, 2: 79–105.
nected in contexts where interpersonal relationships Bansal, P., & Clelland, I. 2004. Talking trash: Legitimacy,
serve as both monitoring mechanisms and control impression management, and unsystematic risk in
mechanisms. For example, the close relationship the context of the natural environment. Academy of
between the CEOs of a buyer–supplier dyad may act Management Journal, 47: 93–103.
as both the monitoring mechanism for state uncer- Baron, D. P. 2003. Private politics. Journal of Economics &
tainty (e.g., monitoring a buyer’s commitment to the Management Strategy, 12: 31–66.
supplier) and the mechanism for responding to an in- Bertrand, M., & Schoar, A. 2003. Managing with style: The
creasingly important dependency (e.g., using the close effect of managers on firm policies. Quarterly Journal
relationship to influence the buyer’s commitment). A of Economics, 118: 1169–1208.
2021 Sutton, Devine, Lamont, and Holmes, Jr. 59
Hersch, P. L., & McDougall, G. S. 2000. Determinants of au- Kroszner, R. S., & Stratmann, T. 2005. Corporate campaign
tomobile PAC contributions to House incumbents: Own contributions, repeat giving, and the rewards to legislator
versus rival effects. Public Choice, 104: 329–343. reputation. Journal of Law & Economics, 48: 41–71.
Hess, D. J. 2014. Sustainability transitions: A political co- Langbein, L. I. 1986. Money and access: Some empirical
alition perspective. Research Policy, 43: 278–283. evidence. Journal of Politics, 48: 1052–1062.
Hiatt, S. R., & Park, S. 2013. Lords of the harvest: Third-party Lant, T. K., Milliken, F. J., & Batra, B. 1992. The role of
influence and regulatory approval of genetically mod- managerial learning and interpretation in strategic
ified organisms. Academy of Management Journal, persistence and reorientation: An empirical explora-
56: 923–944. tion. Strategic Management Journal, 13: 585–608.
Hiatt, S. R., Sine, W. D., & Tolbert, P. S. 2009. From Pabst to Lin, C., Ma, Y., Malatesta, P., & Xuan, Y. 2011. Ownership
Pepsi: The deinstitutionalization of social practices structure and the cost of corporate borrowing. Journal
and the creation of entrepreneurial opportunities. of Financial Economics, 100: 1–23.
Administrative Science Quarterly, 54: 635–667. Lockheed Martin. 2013. 2012 annual report. Retrieved
Hillman, A. J., & Hitt, M. A. 1999. Corporate political from https://www.lockheedmartin.com/content/dam/
strategy formulation: A model of approach, partici- lockheed-martin/eo/documents/annual-reports/2012-
pation, and strategy decisions. Academy of Man- annual-report.pdf
agement Review, 24: 825–842. Lux, S., Crook, T. R., & Woehr, D. J. 2011. Mixing business
Hillman, A. J., Keim, G. D., & Schuler, D. 2004. Corporate with politics: A meta-analysis of the antecedents and
political activity: A review and research agenda. outcomes of corporate political activity. Journal of
Journal of Management, 30: 837–857. Management, 37: 223–247.
Hillman, A. J., Withers, M. C., & Collins, B. J. 2009. Re- Magee, C. 2002. Do political action committees give money
to candidates for electoral or influence motives? Public
source dependence theory: A review. Journal of
Choice, 112: 373–399.
Management, 35: 1404–1427.
Maguire, S., & Hardy, C. 2009. Discourse and deinstitu-
Holburn, G. L. F., & Vanden Bergh, R. G. 2014. Integrated market
tionalization: The decline of DDT. Academy of
and nonmarket strategies: Political campaign contribu-
Management Journal, 52: 148–178.
tions around merger and acquisition events in the energy
sector. Strategic Management Journal, 35: 450–460. Maher, K., & McWhirter, C. 2017, February 13. DuPont
settlement of chemical exposure case seen as “shot in
Hubbard, T. D., Pollock, T. G., Pfarrer, M. D., & Rindova,
the arm” for other suits. Wall Street Journal. Retrieved
V. P. 2018. Safe bets or hot hands? How status and
from https://www.wsj.com/articles/dupont-chemours-
celebrity influence strategic alliance formations by
settle-teflon-chemical-exposure-case-for-671-million-
newly-public firms. Academy of Management Jour- 1486987602
nal, 61: 1951–1975.
March, J. G. 1978. Bounded rationality, ambiguity, and the
Janis, I. L., & Fadner, R. 1965. The coefficient of imbalance. engineering of choice. Bell Journal of Economics, 9:
In H. D. Lasswell & N. C. Leites (Eds.), Language of 587–608.
politics: Studies in quantitative semantics: 153–169.
Cambridge, MA: MIT Press. March, J. G., & Simon, H. A. 1958. Organizations. Oxford,
U.K.: Wiley.
Kalla, J. L., & Broockman, D. E. 2016. Campaign contribu-
McCarthy, J. D., & Zald, M. N. 1977. Resource mobilization
tions facilitate access to congressional officials: A
and social movements: A partial theory. American
randomized field experiment. American Journal of
Journal of Sociology, 82: 1212–1241.
Political Science, 60: 545–558.
McCarthy, J. D., McPhail, C., & Smith, J. 1996. Images
Kassinis, G., & Vafeas, N. 2006. Stakeholder pressures and
of protest: Dimensions of selection bias in media
environmental performance. Academy of Manage-
coverage of Washington demonstrations, 1982 and
ment Journal, 49: 145–159.
1991. American Sociological Review, 61: 478–499.
King, M. D., & Haveman, H. A. 2008. Antislavery in America:
McDonnell, M.-H., & Werner, T. 2016. Blacklisted busi-
The press, the pulpit, and the rise of antislavery socie-
nesses: Social activists’ challenges and the disruption
ties. Administrative Science Quarterly, 53: 492–528.
of corporate political activity. Administrative Sci-
Klarner, C. 2018. State legislative election returns, 1967–2016. ence Quarterly, 61: 584–620.
Harvard Dataverse, V3. doi: 10.7910/DVN/3WZFK9 Melone, N. P. 1994. Reasoning in the executive suite: The
Krasno, J. S., Green, D. P., & Cowden, J. A. 1994. The dy- influence of role experience-based expertise on deci-
namics of campaign fundraising in house elections. sion processes of corporate executives. Organization
Journal of Politics, 56: 459–474. Science, 5: 438–455.
2021 Sutton, Devine, Lamont, and Holmes, Jr. 61
Meyer, D. S., & Whittier, N. 1994. Social movement spill- determinants of fire insurance regulation. American
over. Social Problems, 41: 277–298. Journal of Sociology, 107: 101–146.
Milliken, F. J. 1987. Three types of perceived uncertainty about Schuler, D. A. 1999. Corporate political action: Rethinking
the environment: State, effect, and response uncertainty. the economic and organizational influences. Business
Academy of Management Review, 12: 133–143. and Politics, 1: 83–97.
Miquel, G. P. I., & Snyder, J. M., Jr. 2006. Legislative ef- Schuler, D. A., Rehbein, K., & Cramer, R. D. 2002. Pursuing
fectiveness and legislative careers. Legislative Stud- strategic advantage through political means: A multi-
ies Quarterly, 31: 347–381. variate approach. Academy of Management Journal,
Mitchell, N. J., Hansen, W. L., & Jepsen, E. M. 1997. The 45: 659–672.
determinants of domestic and foreign corporate po- Schwenk, C. R. 1984. Cognitive simplification processes
litical activity. Journal of Politics, 59: 1096–1113. in strategic decision‐making. Strategic Management
National Institute on Money in State Politics. 2012, April 12. Journal, 5: 111–128.
An overview of campaign finances, 2009–2010 elec- Shi, W., Connelly, B. L., & Sanders, W. G. 2016. Buying bad
tions. Retrieved from https://www.followthemoney. behavior: Tournament incentives and securities class
org/research/institute-reports/an-overview-of-campaign- action lawsuits. Strategic Management Journal, 37:
finances-2009-2010-elections 1354–1378.
National Institute on Money in State Politics. 2019. Custom Smircich, L., & Stubbart, C. 1985. Strategic management in
query: Exxon Mobil [Results]. Follow The Money. Re- an enacted world. Academy of Management Review,
trieved from https://www.followthemoney.org/entity- 10: 724–736.
details?eid5946.
Sorauf, F. J. 1992. Politics and money. American Behav-
North, D. C. 1990. Institutions, institutional change and ioral Scientist, 35: 725–734.
economic performance. Cambridge, U.K.: Cambridge
Starbuck, W. H. 1975. Information systems for organiza-
University Press.
tions of the future. In E. Grochla & N. Szyperski (Eds.),
Nystrom, P., & Starbuck, W. 1984. To avoid crises, un learn. Information systems and organizational structure:
Organizational Dynamics, 12: 53–66. 217–229. New York, NY: de Gruyter.
Ozer, M. 2010. Top management teams and corporate Stratmann, T. 1991. What do campaign contributions buy?
political activity: Do top management teams have Deciphering causal effects of money and votes.
influence on corporate political activity? Journal of Southern Economic Journal, 57: 606–620.
Business Research, 63: 1196–1201.
Su, W., & Tsang, E. W. 2015. Product diversification and fi-
Pacheco, D. F., & Dean, T. J. 2015. Firm responses to social nancial performance: The moderating role of secondary
movement pressures: A competitive dynamics perspec- stakeholders. Academy of Management Journal, 58:
tive. Strategic Management Journal, 36: 1093–1104. 1128–1148.
Pacheco, D. F., York, J. G., & Hargrave, T. J. 2014. The co-
Sutcliffe, K. M., & Huber, G. P. 1998. Firm and industry as
evolution of industries, social movements, and institu-
determinants of executive perceptions of the environ-
tions: Wind power in the United States. Organization
ment. Strategic Management Journal, 19: 793–807.
Science, 25: 1609–1632.
Toffel, M. W., & Marshall, J. D. 2004. Improving environ-
Pfeffer, J., & Salancik, G. R. 1978. The external control of
mental performance assessment: A comparative analysis
organizations: A resource dependence perspective.
of weighting methods used to evaluate chemical release
New York, NY: Harper & Row.
inventories. Journal of Industrial Ecology, 8: 143–172.
Pittman, R. 1977. Market structure and campaign contri-
Tyson Foods. 2016. 2016 annual report. Retrieved from
butions. Public Choice, 31: 37–52.
https://www.sec.gov/Archives/edgar/data/100493/
Pollock, T. G., & Rindova, V. P. 2003. Media legitimation 000010049315000109/tsn201510kq4.htm
effects in the market for initial public offerings.
United States Environmental Protection Agency. 2013. TRI-
Academy of Management Journal, 46: 631–642.
covered industry sectors. Retrieved from https://www.
Porter, M. E. 2008. The five competitive forces that shape epa.gov/toxics-release-inventory-tri-program/my-
strategy. Harvard Business Review, 86(1): 78–93. facilitys-six-digit-naics-code-tri-covered-industry
Rowley, T. J. 1997. Moving beyond dyadic ties: A network United States Environmental Protection Agency. 2019.
theory of stakeholder influences. Academy of Man- Toxic Release Inventory (TRI) program [Data set].
agement Review, 22: 887–910. Retrieved from https://www.epa.gov/toxics-release-
Schneiberg, M., & Bartley, T. 2001. Regulating American inventory-tri-program/my-facilitys-six-digit-naics-code-
industries: Markets, politics, and the institutional tri-covered-industry
62 Academy of Management Journal February
Walsh, J. P. 1988. Selectivity and selective perception: An inves- Richard A. Devine ([email protected]) is an assistant
tigation of managers’ belief structures and information pro- professor of management in the Driehaus College of Busi-
cessing. Academy of Management Journal, 31: 873–896. ness at DePaul University. He received his PhD from
Weber, S., & Fong, L. 2017, December 17. How an unregu- Florida State University. His research interests are in the
lated chemical entered a North Carolina community’s areas of non-market strategy, leadership and individual
drinking water. Retrieved from https://www.pbs.org/ differences, and entrepreneurship.
newshour/show/how-an-unregulated-chemical-entered- Bruce T. Lamont ([email protected]) is the Jim Moran
a-north-carolina-communitys-drinking-water Eminent Scholar of Business Administration at Florida
Weick, K. E. 1969. Social psychology in an era of social State University. He received his PhD from the University
change. American Psychologist, 24: 990–998. of North Carolina at Chapel Hill. His primary research in-
terests encompass the effective management of acquisi-
Weick, K. E. 1978. The spines of leaders. Durham, NC:
tions and contextual contributions to theory, particularly
Duke University Press.
involving the African context.
Weick, K. E. 1988. Enacted sensemaking in crisis situa-
tions. Journal of Management Studies, 25: 305–317. R. Michael Holmes, Jr. ([email protected]) is the Jim
Moran Associate Professor of Strategic Management at
Weick, K. E., Sutcliffe, K. M., & Obstfeld, D. 2005. Orga- Florida State University. He also has a visiting appoint-
nizing and the process of sensemaking. Organization ment at the University of Johannesburg. Michael received
Science, 16: 409–421. his PhD from Texas A&M University. His research interests
Werner, T. 2015. Gaining access by doing good: The effect of include institutions and the microfoundations of strategic
sociopolitical reputation on firm participation in public management and entrepreneurship.
policy making. Management Science, 61: 1989–2011.
Wiersema, M. F., & Bantel, K. A. 1992. Top management
team demography and corporate strategic change.
Academy of Management Journal, 35: 91–121.
Wolfe, R. A., & Putler, D. S. 2002. How tight are the ties that bind APPENDIX A
stakeholder groups? Organization Science, 13: 64–80.
DESCRIPTION OF INTERVIEWEES
Wooldridge, J. M. 2010. Econometric analysis of cross section
and panel data. Cambridge, MA: MIT Press. TABLE A1
Wry, T. J., Cobb, A., & Aldrich, H. E. 2013. More than a Description of Interviewees
metaphor: Assessing the historical legacy of resource Executive Interviewee Description
dependence and its contemporary promise as a theory
of environmental complexity. Academy of Manage- A TMT member at sample firm,
ment Annals, 7: 439–486. Alpha
York, J. G., Hargrave, T. J., & Pacheco, D. F. 2016. Converging B Public affairs senior manager at
sample firm, Beta
winds: Logic hybridization in the Colorado wind energy
C Government affairs senior
field. Academy of Management Journal, 59: 579–610. manager at sample firm,
Zald, M. N., & Ash, R. 1966. Social movement organizations: Gamma
Growth, decay and change. Social Forces, 44: 327–341. D Senior manager at sample firm,
Delta
Zald, M. N., & McCarthy, J. D. 1979. Social movement E Senior manager at publicly
industries: Competition and cooperation among traded firm, Epsilon
movement organizations (Center for Research on F Senior manager at publicly
Social Organization working paper 201). Retrieved traded firm, Zeta
from http://deepblue.lib.umich.edu/bitstream/2027.42/ G Senior manager at privately held
50975/1/201.pdf firm, Eta
H Senior manager at privately held
firm, Theta
I Political consultant and lobbyist
J Political consultant and lobbyist
Trey Sutton ([email protected]) is an assistant pro- K Political consultant and lobbyist
L Political consultant and lobbyist
fessor of management in the Robins School of Business at
M Political consultant and lobbyist
the University of Richmond. He received his PhD from
N Political consultant and attorney
Florida State University. His research interests concern the O Current government official
interaction of firms and nonmarket stakeholders, as well as P Former government official
the effects of these interactions on stakeholder outcomes.
Copyright of Academy of Management Journal is the property of Academy of Management
and its content may not be copied or emailed to multiple sites or posted to a listserv without
the copyright holder's express written permission. However, users may print, download, or
email articles for individual use.