Taking Initiative: Proactive Management and Organizational Performance in Networked Environments

Abstract

INTRODUCTION

For more than a century, the classical intraorganizational management perspective has guided studies in public administration. Unfortunately, this orientation is increasingly inapplicable as agencies become characterized as multiorganizational, multigovernmental, and multisectoral (Agranoff and McGuire 1999). Put simply, organizations are oftentimes operating in networked environments, where isolated organizations of old have been “significantly replaced by networks of actors who must be taken into account in the management of public policy” (Hall and O'Toole 2004, 186).

A network, defined by Meier and O'Toole (2001, 272), is a “pattern of two or more units, in which not all major components are encompassed within a single hierarchical array.” Networks are also classically defined as the totality of all units connected by a certain type of relationship (Aldrich and Whetten 1981). Conceptual dimensions of network management, from these points, have burgeoned to include analyses of critical functional equivalents to traditional management processes (the POSDCORB of network management) (Agranoff and McGuire 2001). However, scholars who are attempting to explore the nuances of network management often find the literature wanting for broadly applied theories, meaningful, large-N studies, and succinct explanatory research. These deficiencies have motivated scholars to promote the development of a network management knowledge base equivalent to the hierarchical organizational paradigm of bureaucratic public management (Agranoff and McGuire 1999; Gill and Meier 1999; O'Toole 1997; Weiner 1990).

While the aforementioned literature is working to establish, empirically, that network management contributes positively to organizational performance, theoretical work remains on to answer how network management induces positive organizational outcomes. Similarly, although the classical intraorganizational management perspective may seem unsuitable for today's multiorganizational environment, researchers should not abandon what classic organizational theory can offer as the perspective shifts. For example, concepts of cooperation, uncertainty avoidance, credibility, and coalitional behavior may help further explain why managers behave as they do in managing interactions with network actors.¹

The story along these lines is not about the personal attributes of a manager, that is, personality. Instead, it explores how managers induce network cooperation by establishing proactive management behaviors, while also utilizing their function as boundary-spanning communicators to reduce environmental uncertainties (Cherrington 1994; Gibson, Ivancevich, and Donnelly 1994; Miller 1992).

This article represents a first step toward bringing a proactive management perspective to bear on the empirical analysis of network management organizational performance. The goal is to enable researchers to see a clearer picture of how network management influences program performance. Public education provides the context for the investigation. The coming analyses show statistically significant and substantively interesting relationships among measures of managerial behaviors (specifically proactive and reactive management), resource variables, and environmental constraint variables. Findings are both descriptive and explanatory. They constitute an important contribution to how management behavior influences organizational performance on a set of programmatic indicators.

NETWORK MANAGEMENT

In an effort to see how general networking management influences organizational performance, Meier and O'Toole (2003, 692) characterize and measure the concept as “greater interaction between managers and environmental actors” who are not direct-line subordinates or superiors. The authors concede that this narrow definition ignores skill, reputation, and strategic consideration. Fortunately, this characterization can be improved upon by delineating particular managerial behaviors from the general concept, like proactive management versus reactive management, and assessing their independent effects on performance.²

PROACTIVE MANAGEMENT AND NETWORK CAPITAL³

A successful hierarchy consists of actors who find numerous convincing ways to demonstrate their continued commitment to cooperation (Miller 1992). The same logic is followed for networked environments, though networks are functionally situated in single hierarchical arrays. For example, managers proactively interact with actors in their external networks in order to reduce uncertainty in their environments and strengthen their credibility as boundary-spanners. Boundary-spanners, in this case, represent those who have access to ideas and information that is transmitted throughout the network. Similarly, adopting a strategy of high uncertainty avoidance is evidence that managers are motivated to some degree by fear of the unknown. Thus, their behavior reflects their desire to avoid uncertain situations that may exist in the network. Put simply, proactive management facilitates exchange, communication, interaction, coordination, and control in organizations, as well as among networks. These activities generate a type of network capital that can be translated into potential performance gains for the organization (for other sources of network capital, see Meier, O'Toole and Goerdel 2005).

One way managers can reduce uncertainty (within the organization and in the network) is by utilizing their boundary-spanning function to actively initiate interactions between themselves and network actors. Public managers, as boundary-spanners, facilitate communication between network nodes and their particular organization (Cherrington 1994). The purpose is to gain knowledge (while also disseminating it) about the operation of the network and simultaneously to obtain accurate information in order to make important decisions and judgments concerning the organization (Simon 1997).

Similarly, managers engage in proactive behavior in order to maximize program benefits and minimize future losses, in terms of program outputs. This is achieved, in part, by attempting to avoid ambiguous situations (that is, by adopting uncertainty avoidance). More specifically, proactive management reflects efforts by managers to maximize potential benefits (organizational or network), while at the same time reducing uncertainty about the future of those benefits.⁴

The argument thus follows that managers reduce uncertainty by proactively managing network actors. Considering the point more broadly, proactive management behavior is also characterized as part of a larger power structure erected by the manager in order to (1) induce strategic collaboration among network actors, and (2) to control, to some degree, their environments (Kickert and Koppenjan 1997). Decisions emerging from such activities are motivated by a utilitarian rationale in order to gain access to resources and minimize negative environmental inputs (Raab 1999).

Furthermore, by initiating interaction with network actors, proactive managers create favorable environments where the possibility of controlling the (general) agenda is higher than if they engaged in an interaction they did not initiate.⁵ To demonstrate the multifaceted nature of proactive managerial behavior, attention is turned toward the concepts of framing and synthesizing. When mapped onto the common network management sequence of framing, proactive management is a tool used when network (interactions with network actors) effectiveness is suboptimal. For example, proactive managers utilize framing to give shape to specific purposes and influence the alignment of various forms of engagement (Agranoff and McGuire 2001; O'Toole 1997; Stone. Doherty, Jones, and Ross 1999).

Similarly, proactive management is also considered a synthesizing mechanism. Managers who initiate contact with network actors are proactively working to create environments where they can strategically steer network interactions in the direction of achieving meaningful cooperation. In a sense, proactive management, used as a synthesizing mechanism, potentially lowers the cost of network interaction and possibly increases organizational performance across programmatic indicators (Agranoff and McGuire 1999; Kickert and Koppenjan 1997).

These two facets help us conceive of how proactive management operates, in part, within networked environments. Moreover, these characterizations conceptually follow those that are similarly advanced in game-management research, which evaluates outcomes based on strategic management behavior (Klijn and Teisman 1997).

NETWORKS IN PUBLIC EDUCATION

In this study public education provides the context for investigating the influence of proactive management on organizational performance within networked environments. In the United States the public educational function is “conducted by locally managed school districts, typically designed as separate, special purpose governments, and not formally interdependent with other production units” (Meier and O'Toole 2001, 274). Despite this, school superintendents—who are public managers—are often found managing networked relationships with actors in their external environments, such as local business leaders, other superintendents, relevant state legislators, formal education agencies, and parent groups.⁶ These actors are also known as network nodes, or nodal actors (Aldrich and Whetten 1981; O'Toole 1997).

Numerous policy disputes, like those surrounding funding, high-stakes testing, and teacher certification policies, are often fought on school battlegrounds. In this modern era of networked environments, however, school districts and superintendents are no longer the only voice of reason and decision on such matters. Instead, they are one voice in the cacophony of many that make up the public education network. As such, superintendents should view network development as an opportunity to recognize their interdependence with network actors. Furthermore, they should try to manage their networks actively (Meier and O'Toole 2001).

MANAGEMENT AND PUBLIC EDUCATION

Managers of organizations that are situated within larger networks often interact with nodal actors. This simple point alone testifies to the potentially chaotic nature of management processes under such conditions. Even so, it cannot be denied that “managers do and should engage in purposeful, goal-oriented actions” (Rainey 1997, 159). From this, one can evaluate a number of prescriptive frameworks for strategic management (Bryson and Einsweiler 1995; Porter 1985). These illustrate the general point from traditional organization theory that managers often choose approaches that best match their particular situation and goals. This article explores another complementary prescriptive framework: proactive management when operating in a network.⁷

In the context of public education, superintendents (managers) employ one of three modes of network interaction when coming into contact with external network actors: a mandated interaction, an actor-requested interaction, or a manager-initiated interaction.

Mandated Interaction

The first mode is characterized as a required interaction between manager and network actor. Practical examples include a mandated campus visit from the state education agency, as well as a monthly scheduled meeting with a parent group. Conceptually, this mode of interaction is part of the formal network structure that emerges in networks where resource dependencies and the value of network stability are high (Aldrich and Whetten 1981; Landau 1969).

Actor-Requested Interaction

In this mode, a network actor requests interaction with the manager. Examples of this include a concerned parent asking to meet with a superintendent or a local business leader calling to talk about a recent bond issue. These actions demonstrate the reality of coalitional behavior that exists between managers and network actors. Simply put, network actors purposefully initiate interaction with superintendents in order to share information, garner support, and/or highlight incentives available to the manager in exchange for favorable responsiveness. Pfeffer and Salancik (1977, 1978) and Cobb (1991) highlight from this general concept the essential motive that drives coalitional activities: to attract and secure resources or support for differentiated purposes, while operating within a set of organizations or network. The key point to notice, as it concerns the manager, is that he or she is now in a reactive position of interaction in relation to the nodal actor.

Manager-Initiated Interaction

In the final mode the manager initiates contact with a network actor. This mode includes, for example, a superintendent meeting with a state legislator to discuss a policy concern or a manager initiating contact with another superintendent to brainstorm innovative ideas for improving learning environments. Alongside the conceptualization of organizational coalition behavior from above, this managerial action is portrayed as management being (proactively) responsive to external constraints and network dependencies. The concept of “managerial responsiveness” here and throughout the article means being aware of various demands and constraints within the network and the organization itself, and then molding actions to take account of the most critical constraints and demands made by those on whom the organization is most dependent for its future success, even survival. This construction elucidates components of the rational organizational model, which are contingent partly upon the manager-initiating network interaction (Pfeffer and Salancik 1977). What the superintendent is trying to do through proactive management is to increase his or her organization's (school district) overall performance on a set of programmatic indicators.

Across all three modes of interaction, it is plausible that managers are exploiting and/or buffering their environments. The interesting question is, do managers who initiate contact with network actors (proactive management) experience more success than managers who simply engage in mandated, or actor-requested (reactive management) interactions? To answer this, more operational distinction is made between the two management behaviors.

In the present case proactive management exists when a manager initiates contact with network actors.⁸ Conversely, a superintendent whose network interaction emerges from the first two modes can be thought of as a reactive manager. Suffice it to say, not all managers who initiate have higher-quality networking skills. Along the same lines, not all managers who simply interact when they are asked have lower-quality networking skills. The object is not to measure subjective networking quality. Instead, the objective is to explore and somehow quantify managerial behavior, specifically, proactive and reactive management, in order to determine their influence on organizational performance when operating in a networked environment.

MODELING THE EFFECTS OF MANAGERIAL BEHAVIOR

The network management portion of public management is concerned with the use of managerial tools and strategies that have demonstrable effects on program performance. This is the impetus behind studying managerial behaviors as strategic tools. Due to this fact, it is important that scholars spend time grappling with questions of measurement and estimation that are unique to the field as they begin to think about modeling meaningful relationships. This study follows the advice of Gill and Meier (1999) and focuses on the relative magnitude of program effects (for example, program success). More specifically, it analyzes program performance as a function of managerial behaviors (proactive/reactive), program resources, and task difficulties (or constraints).⁹

METHODS

Unit of Analysis

The data used in this study are drawn from independent school districts in Texas. School districts represent the unit of analysis. It is maintained in this study that “independent” means the school district is not subordinate to another unit such as a city. Independent districts have their own elected board, have the ability to tax and set budgets, and acquire bonding authority by a vote of the residents (Meier and O'Toole 2001, 2003). The survey portion of these data was collected using mail questionnaires. All district superintendents were sent mail questionnaires asking for responses to a number of questions concerning management style, goals, and how they allocated their time within and without the district. Additional nonsurvey data are from the Texas Education Agency.¹⁴ Nonsurvey data are pooled for six school years (1996–2002).

Measures

Measures used in this study are discussed in terms of equations [3] and [4]: proactive management (M_P) and reactive management (M_R); the vector of environmental forces (X); and current performance (O_t). The following discussion addresses these items in order.¹⁵

Proactive and Reactive Management

The two new managerial behavior variables, M_P and M_R, are terms simply derived from the original M₂ term from equation [1]. In the survey, superintendents were asked to respond to (1) how frequently they interacted with individuals in the aforementioned network nodes,¹⁶ and (2) if they initiated the last contact with the network actor.¹⁷ In order to measure proactive management (M_P), frequency of contact with network actors is interacted with the occurrence of initiated contacts. The same process is used to measure reactive management, except frequency of contact with network actors is interacted with the occurrence of non-initiated contacts.¹⁸

Proactive superintendents should be found to initiate contact and interact more frequently with all five network nodes than superintendents who reactively manage. A composite network management scale is created for both proactive and reactive management using factor analysis, considering all five main nodes. Each item included in the proactive and reactive analyses positively loaded on the first factor, producing eigenvalues of 1.49 and 1.32, respectively (table 1). These factor scores are then used as measures of proactive and reactive management.

These measures work well because they consider the interactive relationship between general strategic management and the network. Although they only consider whether a manager initiated the last contact with a particular network node, they create a good starting point to investigate how those who initiate differ from those who do not. At a minimum, if we find that initiation (proactive management) matters in this limited sense, future results using improved initiation measures could only prove more compelling.

Environmental Variables

Superintendents often find themselves working in volatile environments. More specifically, they face daily challenges to their ability to effectively manage. These challenges come in the form of environmental constraints, as well as opportunities (Meier and O'Toole 2001). Sorting out all the environmental factors pressuring superintendents is a difficult, but not impossible, task. Fortunately, this process can be simplified by looking to recent literature on public education policy and public school management. Scholars who focus on this task suggest that environmental variables be categorized as either task difficulties or program resources (Hedges and Greenwald 1996; Meier and O'Toole 2001).

We learn from Jencks and Phillips (1998), for example, that racial inequalities and income disparities are negatively correlated with educational performance, particularly when focusing on standardized testing. Evidence like this leads scholars to believe that measures of race and poverty should be included in models concerned with constraints from the environment and their influence on program performance. Not including them would be misspecifiying the model. Taking this into account, three measures of race and poverty, namely, the percentages of African American, Latino, and poor students in a given school district, are considered. As such, it is expected that race and poverty will be negatively related to organizational performance.

In terms of resources, a growing literature (based primarily on longitudinal studies) is beginning to confirm the basic tenet that schools with more resources generally fare better (Wenglinsky 1997). Three measures of resources are evaluated in this study. Average teacher salary, total instructional expenditures per student, and average years of teaching experience are related to both financial and human resources of the school district.

Teachers, principals, and administrators (that is, superintendents) are in constant search of resources in order to meet the diverse needs of students in their districts. More specifically, administrators are concerned with securing fiscal resources to update structural and technological facilities, provide competitive teaching salaries, and provide special needs services. However, since most spending in education pays for salaries of teachers and other staff (Meier, Wrinkle, and Polinard 1999), the environment can be described as personnel-intensive. As such, it can be exploited as a resource. For example, (Hanushek and Pace (1995) evaluate how economic incentives, like higher salaries, attract better-qualified people to a profession. Superintendents and principals also value the resource of an experienced workforce of teachers and often work to obtain employees with such credentials (Meier, Wrinkle, and Polinard 1999; Meier and O'Toole 2001).

Along with these, school size, as a structural component, is also important. School size can act as a constraint or resource. School size is typically measured as average daily student attendance (Bidwell and Kasarda 1975). Generally speaking, the structural dimensions of size have an inverse relationship with performance (Indik and Seashore 1961; Katzell, Barrett, and Parker 1961; Marriot 1949; Thomas 1959). Coleman and Hoffer (1987) find that smaller schools promote social interaction, which creates a form of social capital that facilitates the work of the school.

Taking all these factors into consideration, teaching experience, total instructional spending per student, and teacher salary are all expected to positively contribute to organizational performance. Conversely, school size is expected to have a negative relationship with organizational performance.

Dependent Variables

Several performance indicators are used as dependent variables. The dependent variable of interest, however, is the percentage of overall students in each school district who pass all parts of the statewide standardized test each year—the Texas Assessment of Academic Skill test, commonly known as the TAAS test. This test, given to all students in grades three through eight and eleven, covers topics in reading, writing, and mathematics. The TAAS exam is a high-stakes test (in grade eleven) that receives regular media attention. The Texas Education Agency also publicly scrutinizes (and sanctions against) districts that perform poorly on this exam and praises districts that perform satisfactorily. This makes it a satisfying measure of overall organizational performance. Other dependent variables in the analysis, however, include:

• Attendance

• Dropout rate

• Average SAT scores

• Average ACT scores

• Percentage of students who score above 1110 on the SAT

• Percentage of African American students who pass the TAAS exam

• Percentage of Latino students who pass the TAAS exam

• Percentage of low-income students who pass the TAAS exam

HYPOTHESES

As posited earlier, proactive and reactive management reflect modes of initiation (or non-initiation) with network actors, as well as frequency of interaction. Consequences associated with employing these interaction mechanisms (management behaviors) provide us with expectations as to how organizational performance is affected across a number of programmatic indicators.

• H₁ When controlling for resources and constraints in the environment, as proactive management increases in the network (M_P), the percentage of students passing the TAAS increases.

• H₂ When controlling for resources and constraints in the environment, as reactive management increases in the network (M_R), the percentage of students passing the TAAS does not increase.¹⁹

Null Hypothesis Testing

Two hypotheses are posited above: a null hypothesis (H₂), competing with an alternative research hypothesis (H₁), each describing complementary notions about some management phenomenon, namely, the influence of managerial behavior on organizational performance. Given that public administrators are experts in decision making, the two hypotheses can be considered two decisions, or actions, that are allowable (Gill and Meier 1999). That is, a manager is either likely to experience increased or stagnant program performance, depending on whether he or she decides to proactively or reactively manage interactions with network actors.²⁰ The evidence helps disentangle the general effects of network management by examining two types of managerial behavior.

TESTING THE MODELS

The first set of models is drawn from simplified equations [3] and [4] as additive and non-autoregressive. These models demonstrate that specific management behaviors matter to organizational performance, especially proactive management. Results are presented in table 2. The key coefficient for proactive management is positive and significant, indicating that proactive management matters to organizational performance, when it comes to the main programmatic indicator, the TAAS test. When controlling for resources and constraints in the environment, as proactive management increases, the percentage of students passing the TAAS exam increases, supporting alternative hypothesis (H₁).

Other findings in table 2 suggest that average teacher salary, instructional spending per student, and teacher experience all contribute positively to performance, as expected. Moreover, school size is negatively related to organizational performance. As school size increases year to year, the percentage of students passing the TAAS decreases. Similarly, all three environmental constraint variables, percentage of African American, Latino, and low income students in a district, are significant and in the expected direction (negative). Finally, the second column on table 2 reveals no significant relationship between reactive management and program performance, all else being equal.

PROACTIVE MANAGEMENT: THE AUTOREGRESSIVE COMPONENT

This is appropriate given the pressure placed on superintendents to stress standard operating procedures, specialization, and consistency within the organization (Meier and O'Toole 2001). The results are reported in table 3. Proactive management continues to be positive and significant, even when controlling for organizational inertia. Turning attention toward the autoregressive term, while the lagged dependent variable dominates both equations, resource and constraint variables continue to significantly influence program performance, with few exceptions.

Instructional spending per student and teacher experience remain positive and significant. Despite being insignificant, average teacher salary remains signed in the expected direction (positive). Additionally, after respecifying the model to include the necessary autoregressive component, the percentages of low-income, African American, and Latino students in a district remain consistently significant and in the expected direction (negative). Most interesting, however, is that the key coefficient for proactive management, .124, remains positive and significant—indicating that, even after incorporating an autoregressive term into the public management/performance model, proactive management still matters to performance.

Results for reactive management follow the additive non-autoregressive results in table 2. When controlling for environmental forces and organizational inertia, reactive management does not significantly contribute to program performance.

PROACTIVE MANAGEMENT: ADDITIONAL PERFORMANCE INDICATORS

Next, tables 4 and 5 evaluate the influence of proactive and reactive management on various indicators of organizational performance. Models used in these analyses are from equations [5] and [6]. Table 4 looks specifically at the impact of proactive management on organizational performance.

The key coefficient—proactive management—continues to positively and significantly contribute to three additional performance indicators—attendance rate, low-income TAAS pass rate, and the Latino pass rate. Likewise, the slope of the coefficient is appropriately signed across all of the additional indicators, except average ACT score. One can speculate why managerial proactivity does not significantly influence SAT and ACT scores, in general. It could be that there exists a trade-off between focusing on baseline goals in the organization, such as success on standardized tests, attendance, and performance of traditionally difficult-to-educate groups, on the one hand and on elite goals associated with college aspirations on the other. Or, it could also be that there are diminishing returns on proactive management for those who are already performing at standard or (well) beyond expectations.

Table 5 offers different results. It is both statistically and substantively compelling that reactive management contributes negatively to programmatic outcomes dealing with SAT.

At the same time, reactive management does not significantly affect the attendance rate, the dropout rate, average ACT scores, or the TAAS pass rates of Latino, African American, and low-income students. At best, these mixed findings tell us that managers who manage reactively do not necessarily experience an increase or decrease in most performance outcomes.

Finally, both proactive and reactive management measures are evaluated within a single model. The results are presented in table 6. In the base model proactive management remains positive and significant, even when controlling for reactive management and environmental forces.

Similarly, after incorporating the autoregressive component, proactive management, again, continues to contribute positively and significantly to program performance. In both models environmental constraints and resources contributed to performance similarly to earlier analyses.

DISCUSSION

This study attempts to evaluate the influence of specific management behaviors on organizational performance, when operating within a networked environment. The evidence is compelling enough to assert there is something different about managers who initiate contact with network actors (proactive management) versus those who do not, in terms of performance. Previous analyses focused on statistical results from various models that incorporated measures of proactive and reactive management.

While the results are compelling, what story are they telling? Do proactive managers positively influence organizational performance? Do proactive managers differ from reactive managers? The answers to these questions are the same: Yes, within limits. Evidence indicates that proactive management contributes positively to higher performance outcomes, especially when it comes to average overall TAAS pass rates and the TAAS pass rates for low-income and minority students. Put simply, there is added value in proactively managing interactions with network actors.

On the other hand, how is this different from reactive managers? First of all, the evidence is stacked against reactive managers. Across six performance indicators, reactive managers did not contribute significantly to positive performance. In fact, table 5 demonstrates that reactive management contributes negatively to performance when it comes to SAT scores. This leads to the conclusion that reactive management needs to be tested as a separate alternative hypothesis, instead of simply as a null hypothesis to proactive management. This is one area available for future work in this research series.

Along these lines, this research can benefit from further quantitative analyses, particularly tests establishing whether nonlinear relationships exist. More specifically, since the Meier/O'Toole model of management is autoregressive, nonlinear, and contingent, the next phase of analyses concentrates on integrating proactive and reactive management more thoroughly into equation [1].

CONCLUSION

Proactive management is one strategic management tool used to achieve organizational success, when operating within a networked environment. Oftentimes, managers are motivated to adopt contact-initiation strategies so as to reduce uncertainty and maximize network interactions. Superintendents attempt to increase organizational performance by exploiting resources and buffering constraints (Meier and O'Toole 2001). Another critical management task for superintendents is translating network capital into program resources (monetary and otherwise), which are used to increase programmatic outcomes.

One fact is clear, proactive management matters to organizational performance. Managers who initiate contact with network actors more frequently experience organizational success (evidenced across a number of programmatic indicators) than those who do not initiate such contact. In light of such evidence, superintendents should work to incorporate contact-initiation strategies as they operate in their networked environments. That is, they should focus on establishing behaviors of proactive management when interacting with network actors.

While managers who do not initiate contact do not necessarily experience a decrease in performance, there is no evidence that reactive management increases organizational performance. Given this information, if managers know that proactive management is generally positively associated with increased performance, then perhaps they will proactively engage their respective networks.

References