Week 2 - Assignment: Describe the Senior Leadership Team's Role in Managing the Agency's Mission and Week 3 - Assignment: Examine Senior Leadership's Actions Toward Improving Organizational Efficiency
Fila64
week23PublicManagementinTime.pdf
Public Management in Time: A Longitudinal Examination of the Full Range of Leadership Theory Zachary W. Oberfield Haverford College
AbsTRAcT
Though management is theorized as a temporal process, and public organizations are understood as generally inertial, most public management studies rely upon cross-sectional research designs. As such, we have little understanding about how public management matters over time. To fill this gap in the literature, this article conducts a longitudinal analy- sis of one of the most established management theories: Bass’s full range of leadership theory. This theory expects positive outcomes when managers establish patterns of trans- actional and transformational leadership. To examine Bass’s theory, this article studies US federal government subagencies over a 7-year period in the beginning of the 21st century. The findings show that there are remarkably strong intra-organizational patterns over time. Nonetheless, there is evidence that management matters: improvements in leadership are positively associated with follower cooperation, satisfaction, and perceptions of work quality. In comparison, as Bass expects, transformational leadership is a stronger predictor of improvement. Though public management appears to matter over time, the article also shows that cross-sectional examinations may overvalue its effect. As such, the article closes by arguing for further longitudinal public management study.
Public managers are simultaneously depicted as powerful and constrained (Ingraham, Joyce, and Donahue 2003; Moynihan and Pandey 2005; Riccucci 2005; Meier et al. 2007; Reisner 2011). Managers are powerful, in part, because they shape the formal and informal organizational processes that propel a group of people to come together to work for a common goal (Simon 1997). Without their efforts, cooperative behavior would be impossible and organizations would devolve into dysfunction. At the same time, public managers are also constrained: They must reckon with considerable inter- nal inertia due to personnel continuity, strong unions, and difficult-to-alter standard operating procedures, and they have relatively few tools at their disposal to motivate followers (Kaufman 1975; Wilson 1989; O’Toole and Meier 1999).
Although many studies have examined different components of this tension, there are two limitations to our existing understanding of the effect of public management.
The author would like to thank Kenneth Meier, the editors, and three anonymous reviewers for their useful suggestions and feedback. Address correspondence to the author at [email protected].
JPART 24:407–429
doi:10.1093/jopart/mus060 Advance Access publication December 14, 2012 © The Author 2012. Published by Oxford University Press on behalf of the Journal of Public Administration Research and Theory, Inc. All rights reserved. For permissions, please e-mail: [email protected]
Journal of Public Administration Research and Theory 408
First, few studies have focused on the relationship between public managers and followers over time: most use cross-sectional research designs that examine variation across cases at a single moment. This is potentially problematic because management is theorized as a temporal process: Managers implement changes at one time and followers act or think differently at a later time (O’Toole and Meier 1999). Second, those public management studies that do incorporate time tend to focus on a small number of cases or within a sin- gle part of a government. These case studies are vulnerable to concerns about generaliz- ability. In sum, there are few longitudinal, cross-government public management studies.
This article contributes by studying the relationship between managers and follow- ers over time across an entire level of government. In particular, it conducts a longitu- dinal analysis of Bass’s (1996) full range of leadership theory. This theory expects that followers will perform better when managers establish patterns of transactional leader- ship: rational, organized approaches for rewarding and punishing follower behavior. However, according to the theory, transactional leadership has limited potential; to achieve higher levels of success, leaders must adopt transformational leadership that inspires followers and generates high levels of commitment to organization goals. Although Bass’s model has gained widespread attention from scholars in the generic management literature, it has been slower to gain prominence in public administra- tion research (Van Wart 2003). Nonetheless, recently there have been a spate of studies that apply Bass’s theory in public sector settings (Trottier, Van Wart, and Wang 2008; Paarlberg and Lavigna 2010; Wright and Pandey 2010; Moynihan, Pandey, and Wright 2012; Vigoda-Gadot and Beeri 2012; Wright, Moynihan, and Pandey 2012). These stud- ies have contributed much and suggest the utility of Bass’s model in the public sector; however, like most public management studies, nearly all use cross-sectional research designs. By examining individuals and organizations outside of their temporal contexts, these studies may overvalue the importance of this component of management.
As such, this article asks: Are changes in transactional and transformational leadership associated with changes in follower perceptions, experiences, and behavior across a wide array of government agencies? Is transformational leadership, as Bass expects, the more successful leadership strategy? To answer these questions, this article focuses on the US federal government in the beginning of the 21st century; it draws on nearly a million responses from five waves of surveys administered to federal work- ers as part of the Office of Personnel Management’s (OPM) Federal Human Capital Surveys (FHCS) and Federal Employee Viewpoint Surveys (FEVS). For outcomes, the article examines follower cooperation, satisfaction, and perceptions of work qual- ity. By studying management over time, and its association with an array of follower outcomes, this article provides us with a clearer understanding of the potential for Bass’s theory in the public context. Also, more generally, it gives temporal context to our understanding about how public management “matters” over time.
MAnAgEMEnT And bAss’s “FuLL-RAngE” ThEoRy
The prominence of management1 in public administration has risen and fallen over time (Frederickson et al. 2012, 105). Nonetheless, in recent years a consensus has
1 “Management” refers to the approaches used by organization leaders to maintain and advance their organizations using planning, motivation, and communication (Rainey 2003). Leadership, defined below, is understood as one component of management.
Oberfield Public Management in Time 409
developed that, though managers are constrained in important ways, they play a major role in shaping the success of public organizations (Ingraham, Joyce, and Donahue 2003; Riccucci 2005). In part, managers are expected to “matter” because of their formal and informal powers to shape followers’ attitudes and actions. Formally, managers make decisions about follower promotions, determine work flows, establish patterns of communication, and enforce rules. From this perspective, managers help build the systems by which organizations function and, when the systems are rational and fair, it is expected that their followers will thrive (Behn 2006). However, managers also are expected to matter because they set the informal norms for how people within their organizations behave; as they act, they model particular behaviors for followers and set the tone for their organization’s logic of appropriateness (March and Olsen 2010). Although this area of power is more opaque, it is rooted in the understanding that organization leaders have a prominent symbolic status that goes beyond their contracted duties.
One particularly prominent component of management is leadership: the attempt by organization leaders to direct or energize followers to achieve particu- lar goals (Rainey 2003). When effective, leadership is expected to shape followers’ actions and attitudes (Behn 2006). Though there is disagreement about which leader- ship approaches are most useful (Fernandez 2005), Bass’s “full range” of leadership theory is one of the “best-articulated and promising models” (Moynihan, Pandey, and Wright 2012, 1). This theory has been embraced by scholars and practitioners across a range of disciplines and organizations and has achieved a level of public acclaim that is rare for academic theories.2 Like other theorists (Barnard 1968; Burns 1978; Simon 1997), Bass places leaders at the heart of group processes and argues that followers are strongly influenced by their efforts. As they interact with follow- ers, Bass argues that there are eight types of leadership that constitute the full range: laissez-faire, a style in which managers avoid conflict and intervention; passive man- agement by exception, intervention when standards are not met; active management by exception, monitoring follower performance; contingent reward, clarifying what needs to be done and giving rewards; individualized consideration, diagnosing and elevating the needs of each follower; idealized influence, becoming a role model for followers and instilling pride; intellectual stimulation, encouraging followers to see the world differently; and inspirational motivation, articulating an inspiring vision for followers.
From these eight factors, Bass creates three general types of leadership: laissez- faire, transactional (passive management by exception, active management by excep- tion, contingent reward), and transformational (individual consideration, idealized influence, intellectual stimulation, and inspirational motivation). After reviewing evi- dence across a number of nations, Bass (1996, 754) concludes that laissez-faire leader- ship is the least effective and transformational leadership is the most effective. Though transformational leadership is most effective, Bass argues that transactional leader- ship is an important part of the foundation upon which transformational leadership efforts are built.
2 An article in Psychology Today refers to transformational leadership as “the most popular theory of leadership today” (Riggio 2009), there are Web sites devoted to its practice (www.transformationalleadership. net), and a Transformational Leadership Council that aims to connect leaders around the world (http://www. transformationalleadershipcouncil.com/).
Journal of Public Administration Research and Theory 410
Transactional Leadership
The three components of transactional leadership are passive management by excep- tion, active management by exception, and contingent reward. These components share a principal–agent logic about the leader–follower relationship: Managers are responsible for communicating goals and instructions to workers, observing follower behavior, and responding in kind to the actions of followers. As such, they tap into longstanding concerns about organizational coordination, shirking, and control (Brehm and Gates 1997); thus, although not identical, transactional leadership shares much in common with other managerial strategies, like “management for results,” that emphasize monitoring and rewards (Moynihan 2007). Bass sees these approaches to management as “lower order” in that they are fundamental to organization function- ing: They enable followers to get recognition for their work and make it possible for managers to detect when a follower is not producing work. Without transactional leadership, Bass argues, it is not possible to develop “higher order” leadership strate- gies, which have the potential to unlock employee potential and creativity.
Empirical studies that examine transactional leadership approaches show mixed results. Lowe et al.’s (1996) meta-analysis of the generic management literature shows that contingent reward, a component of transactional leadership, is associated with effectiveness but management by exception, another component, is generally not. Taken as a whole, the study suggests that transformational leadership has a stronger relationship with effectiveness than transactional leadership. Another study looks across a diverse array of organizations at frontline workers’ decisions to work, shirk, or sabotage in the context of superior monitoring, an important component of trans- actional leadership (Brehm and Gates 1997). It finds that supervision is less useful for explaining behavior than employees’ solidary preferences.
Other studies provide a more positive portrait of transactional leadership. Choi and Rainey (2010) study perceptions of performance among US federal employees. Although their primary interest is diversity management, they find a strong, statistically significant relationship between “results-oriented culture” and perceived performance. They measure results-oriented culture with questions that ask followers about manag- ers’ use of rewards and sanctions; in this way, the study shows support for transactional leadership. Moynihan and Pandey (2005) examine the environmental and organizational influences on perceptions of organization effectiveness. One of their organizational influences is the extent to which workers have a clear idea about their organization’s main goals. Although this is not the entirety of transactional leadership, it is an impor- tant part: Without clear goals, it is difficult for managers to fairly reward and punish workers. Their findings show that when workers have a clear idea of their main goals, they are much more likely to perceive their organizations as effective. Trottier et al. (2008) examine transformational and transactional leadership and employee satisfac- tion. They find that two components of transactional leadership, contingent reward and management by exception, are positively associated with satisfaction.
Based on this review, and the theoretical expectations driven by Bass’s theory, this article tests the following hypothesis:
H1: Increases in transactional leadership will be positively associated with follower cooperation, satisfaction, and perceptions of work quality.
Oberfield Public Management in Time 411
Transformational Leadership
Though Bass is clear that transactional leadership is an important component of organization functioning, recall that he labels it “lower order.” In this way, he indi- cates that transactional leadership is a necessary component of management but that it is not enough for an organization to achieve its full potential. For that, Bass argues, leaders must employ transformational leadership—they need to inspire followers and unlock their potential for creativity. In other words, transformational leadership “aug- ments transactional leadership to achieve higher levels of subordinate performance with the primary difference residing in the process by which the leader motivates sub- ordinates and in the types of goals set” (Lowe, Kroeck, and Sivasubramaniam 1996, 387). Another way of conceiving transformational leadership is by analogizing it to classic psychological expectations about how people satisfy their needs. First, leaders need to satisfy followers’ lower-level needs, like fairness and stability, via transactional leadership; then they can move on to satisfy followers’ higher level needs, like self- actualization, via transformational leadership (Trottier, Van Wart, and Wang 2008). In this way, leaders can fundamentally alter how their followers behave and improve how their organizations function.
Empirical work shows that transformational leadership is associated with an array of positive organizational outcomes. For example, the meta-analysis by Lowe et al. (1996), which examines studies in the generic management literature, finds that transformational leadership is more highly associated with organization effectiveness than transactional leadership. Although scholars have been slower to examine transformational leadership in public settings (Van Wart 2003), recently there has been a spate of such studies. Trottier et al. (2008) examine Bass’s leader- ship styles and show that a component of transformational leadership, idealized influence, is the strongest predictor of satisfaction. Moynihan et al. (2012) show that transformational leadership is positively associated with public employee goal clarity and an organization’s developmental culture. Via these two outcomes, they argue, transformational leadership shapes an important component of organization performance—the availability and use of information. In a public health setting, Vigoda-Gadot and Beeri (2012) show a positive relationship between transforma- tional leadership and organization citizenship behavior. Wright et al. (2012) find that transformational leadership is associated with high levels of follower public service motivation and goal clarity.
These studies suggest that transformational leadership has positive effects on how employees act and organizations function. However, they tend to rely upon cross-sec- tional research designs and/or one part of a government. This article seeks to deepen our understanding of transformational leadership in two ways. First, as mentioned above, it studies leadership over time across an entire level of government; therefore, it can show how changes in leadership are associated with follower changes in a diverse array of public organizations. Second, by studying follower perceptions, experiences, and behavior, it will show how transformational leadership is associated with various follower outcomes. As such, the article tests the following hypothesis:
H2: Increases in transformational leadership will be positively associated with follower cooperation, satisfaction, and perceptions of work quality.
Journal of Public Administration Research and Theory 412
The Full Range of Leadership
Although Bass sees transactional and transformational leadership as distinct strate- gies, he also notes that they are not mutually exclusive; rather, he expects that organiza- tion leaders can generate the best outcomes when they combine attention to followers’ “lower level” and “higher level” needs. Improvement is expected because followers understand that there will be a base level of fairness in how decisions are handled via transactional leadership; however, they will also feel inspired and empowered via transformational leadership. As such, Bass sees these leadership styles as complemen- tary. To test how followers experience combined leadership, this article tests the fol- lowing hypothesis:
H3: Increases in the full range of leadership will be positively associated with follower cooperation, satisfaction, and perceptions of work quality.
ModELing LEAdERshiP’s EFFEcT
To examine its hypotheses, this article utilizes an open systems model that conceives of individuals as situated in organizations that are themselves embedded in larger insti- tutional and material environments; people, ideas, and information come in and leave organizations in predictable, patterned ways (Scott 2003, 29).3 Many of the factors that shape organizational functioning are conceived of as continuous over time. This last point is important because public organizations are understood as useful, at least in part, because they bring stability and consistency to the disorder of the political world. These advantages, of course, may push these organizations away from respon- siveness and change and toward sclerosis (Merton 1940; Weber 1947; Kaufman 1975; O’Toole and Meier 1999; Scott 2003). Rigidity may result because many of the forces that shape how public organizations function—from standard operating procedures to organization cultures to personnel continuity—exert stabilizing pressures. Thus, as seen in figure 1, although public managers have prominent organizational positions, they are one among many factors that influence followers.
Macro-level forces, like culture, history, and the structure of government, shape the basic traditions and institutions in which public employees operate. These macro forces matter because they establish expectations in organization members about the nature of the citizen–government relationship, bureaucratic capacity, and the poten- tial for government success (Mead 2004). In comparison with the other levels of input, macro-level influences are the most stable—political institutions and regional cultures, for example, change very slowly. Meso-level factors include environmental influences like political support and task expectations. Compared to macro-level forces, there is greater flux here: Social developments and political trends may lead to increased
3 Compared to prior organizational theorizing, an open systems approach directs attention to the environment in which an organization is situated; however, it does not trade “inside” for “outside.” Rather, it broadens the definition of what an organization is and, as a result, how we explain its functioning. As Hall (1977, 58) writes, according to an open systems approach, “Organizations are affected by what comes into them in the form of input, by what transpires inside the organization, and by the nature of the environmental acceptance of the organization and its output.”
Oberfield Public Management in Time 413
authority and responsibility for one agency at the expense of another. For example, in the wake of the 2008 financial crisis, US elected officials restructured the responsibili- ties of the various agencies tasked with regulating the financial industry (Nicholson 2010). A similar pattern occurred after the 9/11 terrorist attacks in the intelligence community (Posner 2005). Although there is greater volatility at the meso level relative to the macro level, there is also considerable continuity—change, when it comes, tends to be precipitated by major focusing events (Kingdon 2002).
At the micro level are influences like management, organization technologies, work- force characteristics, and individual-level characteristics. Management includes lead- ership, goal setting, strategic planning, and environmental navigation. Organization technologies are the tools used by members of the organization to achieve the organi- zation’s goals. Since organizations are deeply tied to the technologies that they use, technological innovation offers an important route to organizational change (Rainey 2003). The characteristics of the workforce include readily measured human resources components, like expertise and diversity, as well as more-difficult-to-measure charac- teristics, like culture. Finally, understanding public employees requires reckoning with individual-level factors like pre-organizational experiences and education levels.
Compared with the other levels, influences at the micro level are most easily changed. Managers can adopt a new managing approach; employees with new types of skills can be hired; and technological developments can alter how employees understand their work. Nonetheless, there is considerable inertia at this level as well: Organizations have considerable personnel continuity over time and, perhaps as a result, cultures are difficult to change. In addition, the actions of organization mem- bers are often prescribed by longstanding standard operating procedures and norms of appropriateness (March and Olsen 1984; March and Olsen 2010). These factors do not guarantee stability but they dampen the prospects of change.
With this model, we have a better appreciation of the environment in which public management operates and a clearer sense of why leadership might or might not matter over time. Since influences from all three levels exert stabilizing forces on employees,
Figure 1 Situating Public Management: An Open Systems Model
:secneulfnilevel-orcaM cihpargoeg;yrotsiH tnemnrevog;erutluc snoitutitsnidnaerutcurts
:secneulfnilevel-oseM fotroppuslacitiloP noitazinagro;noitazinagro
snoitatcepxeksat
:secneulfnilevel-orciM ;tnemeganaM
;seigolonhcetnoitazinagro ;scitsiretcarahcecrofkrow
level-laudividni dnasecneirepxe scitsiretcarahc
’seeyolpmecilbuP ,snoitpecrep dna,secneirepxe
snoitca
Journal of Public Administration Research and Theory 414
some level of inertia is expected. Nonetheless, leadership is one of the most proximate and fungible influences on followers; as such, it may have a considerable effect on followers’ actions and perceptions. Bass’s theory, which has strong theoretical and empirical support, offers a useful case for examining how changes in leadership matter across a broad array of organizations.
REsEARch dEsign
This article studies Bass’s theory by analyzing survey data drawn from federal employ- ees as part of the FHCS and the FEVS. In particular, it focuses on follower changes from 2004–2011. The survey was originally administered every 2 years (2002–2010), but recently it was administered in successive years (2010 and 2011).4 The number of respondents and government-wide response rates were: 140,984 and 54% (2004); 221,479 and 57% (2006); 212,223 and 51% (2008); 263,475 and 52% (2010); and 266,376 and 49% (2011). In each year, the survey was administered electronically to a sample of full-time permanent federal employees. The sample was selected to ensure the validity of findings across the government, at the agency level, at the subagency level, and at the supervisory status level (OPM 2012). The sample was drawn from employees in agencies represented on the President’s Management Council as well as from employees employed by small independent agencies.5 Over the years studied here, the population covered by eligible respondents ranged from 93% to 97% of the executive branch workforce.
The article pools and tests four change periods: 2004–2006, 2006–2008, 2008– 2010, and 2010–2011.6 Since the FHCS/FEVS surveys are anonymous, individual- level panel research is impossible. As a result, this article focuses on subagencies, or “bureaus.” It averages non-supervisor employee responses at the subagency level. As such, the unit of analysis is a subagency-year. In addition to these survey data, the article uses personnel data drawn from OPM’s Central Personnel Data File (2004– 2009) and Enterprise Human Resources Integration Statistical Data Mart (2010– 2011) to build demographic and human resources variables. In all, the article analyzes 131 discrete subagencies in 276 pooled subagency-years. The subagencies included in this analysis are drawn from across the federal government and include subagencies from the Departments of Defense, Agriculture, Commerce, Justice, Labor, Education, Health and Human Services, Homeland Security, Interior, State, Transportation,
4 In its findings section, the article discusses efforts that were made to detect whether these temporal differences affected its findings. 5 In the 2010 and 2011 version of the survey, eligible respondents were drawn from the pool of “full-time, permanent employees of Departments and large agencies and the small/independent agencies” (OPM 2012). 6 Since these years include a period of considerable economic and governmental volatility, it is worth considering what effect this might have had on the data examined here. Some observers predicted that the symbolism of the election of President Obama, and his call to public service, would lead to a renewed interest in public sector work (Perry and Buckwalter 2010). However, coincident with his election was a major national recession and the rise of a strong anti-government populist movement (Williamson, Skocpol, and Coggin 2011). These events may have dampened enthusiasm for public service. Although it is interesting to speculate, to date there has been little investigation about whether these occurrences had any effect on trends in public sector human capital or bureaucratic thought and behavior. Nonetheless, in an effort to control for government and economy-wide temporal trends, the article includes a dummy variable for each year.
Oberfield Public Management in Time 415
Treasury, and Veterans Affairs. They also include subagencies from the Executive Office of the Presidency and independent executive branch organizations. The sub- agencies included in the study are not perfectly representative of the entire govern- ment7; nonetheless, as the above list suggests, they are drawn from a wide array of agencies across the US federal government. In an effort to limit the effect of outliers, this article excludes subagencies in which fewer than 100 employees completed the sur- vey. Excluding subagencies with less than 100 respondents reduced the total number of possible subagencies by 16%. In terms of policy areas and parent agencies, there were no patterns between the excluded and included subagencies. Because the article’s outcomes are continuous variables, and because it analyzes pooled change data, it uses a fixed-effects OLS regression with standard errors clustered by subagency.
dependent Variables
As noted above, transactional and transformational leadership are theorized as influ- encing organizations by affecting followers’ attitudes, experiences, and actions. Thus, to examine these leadership strategies over time, this article focuses on three outcomes. First, it examines cooperation—the ability of individuals to work together for mutual benefit. Organizations are, to a considerable extent, “systems of cooperative behavior” (Simon 1997, 81); when they have high levels of cooperation, they are more likely to achieve their goals (Smith, Carroll, and Ashford 1995; Marks, Mathieu, and Zaccaro 2001; Milton and Westphal 2005). Though cooperation is essential, organizations have entropic tendencies (Hirschman 1970); as such, successful management is theorized as necessary for cooperation. Due to the importance of cooperation as an organizational outcome, and its dependence on management, it is a useful outcome for examining Bass’s theory.
To measure cooperation, the article creates an index based on employees’ responses to two questions: “The people I work with cooperate to get the job done” and “Employees in my work unit share job knowledge with each other.” For both questions, respondents answered on a scale of 1 “Strongly Disagree” to 5 “Strongly Agree.” This index, which has been used in other studies (Choi and Rainey 2010), has sufficient content validity since it asks employees about general levels of cooperation as well as about the sharing of information, a crucial component of organization functioning (Adcock and Collier 2001); it is highly reliable as per the average scale reliability coefficient (Cronbach’s alpha) across the 5 years (0.86). Since the article is concerned with change over time, it uses dynamic measures of employee cooperation: It averages perceptions at the subagency level at time 1 and subtracts them from aver- ages at the subagency level at time 2.
Second, the article focuses on follower satisfaction. There is a long-running debate about the importance of employee satisfaction in organization settings (Kim 2002). In the literature, there are mixed findings about the connection between satisfaction and productivity. However, there is agreement that employee satisfaction is associated
7 Since the article does not seek a perfectly representative sample, FHCS/FEVS sampling weights were not used.
Journal of Public Administration Research and Theory 416
with an array of important organization outcomes, like motivation, absenteeism, and turnover (Rainey 2003; Wright and Davis 2003). Also, since organizations cannot con- tract every action that employees take, there is interest in fostering good “organization citizenship,” an outcome that is strongly linked to employee satisfaction (Organ and Lingl 1995). Prior studies suggest that employee satisfaction is tied to how organiza- tions are managed (Trottier, Van Wart, and Wang 2008); as such, it is a useful outcome for examining the effect of Bass’s theory.
To measure satisfaction, the article creates an index based on employees’ responses to two questions: “Considering everything, how satisfied are you with your job?” and “Considering everything, how satisfied are you with your organization?” For both questions, respondents answered on a scale of 1 “Very Dissatisfied” to 5 “Very Satisfied.” These questions have been used as outcome measures by prior satisfaction studies (Trottier, Van Wart, and Wang 2008). The index has high levels of content validity since it includes two components that, taken together, provide a broad per- spective of employee satisfaction; it is highly reliable as per the average scale reliability coefficient (Cronbach’s alpha) across the 5 years (0.93). Since the article is concerned with change over time, it uses dynamic measures of employee satisfaction: It averages perceptions at the subagency level at time 1 and subtracts them from averages at the subagency level at time 2.
Third, it measures employees’ perceptions of work quality; work quality is an outcome that is described as strongly connected to, if not definitive of, organization performance (Rago 1994; Scott 2003; Swiss 1992). There are varied challenges to focusing on quality in public organizations; nonetheless, it is a prominent concern among public officials around the globe (Boyne and Walker 2002). Employee percep- tions are not, of course, the only way to measure work quality. However, theorists have long argued that operative employees offer a unique and useful vantage point for understanding organization functioning (Lipsky 1980; Simon 1997; Brewer 2006). Outsiders can observe functioning from some remove but insiders are most intimately aware with how their organizations function. As with the other two outcomes, there is evidence that perceptions of work quality are tied to management (Pitts 2009); there- fore, it is a useful outcome for examining Bass’s theory.
To study perceptions of work quality, this article examines responses to the ques- tion: “How would you rate the overall quality of work done by your work group?”8 Respondents answered this question on a scale of 1 “Strongly Disagree” to 5 “Strongly Agree.” This question has been used as an outcome measure by prior cross-sectional studies in the public administration literature (Pitts 2009; Choi and Rainey 2010); it has sufficient content validity since it asks employees directly about the quality of the group’s work. By asking about the group as opposed to the respondent, the question avoids the problem of overly positive self-reporting. Since this article is concerned with change over time, it uses a dynamic measure of employees’ views of group work quality: It averages perceptions at the subagency level at time 1 and subtracts them from averages at the subagency level at time 2.
8 In 2010, OPM replaced the word “group” with “unit.” For the purposes of this article, it is assumed that this change has little to no substantive difference on employees’ responses. The placement of this question within the survey varied over the years; it was question 10 from 2004–2008 and question 28 in 2010 and 2011.
Oberfield Public Management in Time 417
independent Variables
As discussed above, Bass’s “full range” of leadership theory is one of the best artic- ulated and most tested management theories. Bass developed the concepts under- girding his theory using a factor analysis of 70 survey responses to the “Multifactor Leadership Questionnaire” (Bass 1996). Since few surveys can devote space to 70 questions about leadership, most studies of Bass’s model rely upon a truncated meas- urement system, and some examine only one type of leadership (Lowe, Kroeck, and Sivasubramaniam 1996; Trottier, Van Wart, and Wang 2008; Moynihan, Pandey, and Wright 2012). For example, Moynihan et al. (2012) measure transformational leader- ship using a five-question index that includes questions from three of the concept’s four subcomponents: inspirational motivation, idealized influence, and intellectual stimulation.9 Trottier et al. (2008) measure Bass’s theory in three ways. First, they measure six of Bass’s subcomponents individually: management by exception10, con- tingent reward, individualized consideration, idealized influence, inspirational moti- vation, and intellectual stimulation. Second, they collapse management by exception and contingent reward into a transactional leadership index and individualized con- sideration, idealized influence, inspirational motivation, and intellectual stimulation into a transformational leadership index. Third, they collapse all variables into a sin- gle “combined leadership” variable.
The approach used here most closely resembles Trottier et al.’s (2008): It exam- ines transactional and transformational leadership separately, using an index for each, and examines the two types of leadership together, using a single index. Since Bass’s theory suggests that these two approaches to leadership are distinct, but not mutually exclusive, this approach enables the article to investigate the two types of leadership separately and together. Nonetheless, it is important to examine the content valid- ity and convergent/discriminant validity of these indexes when separated (Adcock and Collier 2001). To achieve sufficient content validity, the measures used here must faithfully capture the entirety of Bass’s theory. The indexes meet this threshold since they include at least one question from each of the subcomponents of Bass’s leader- ship types. Transactional leadership is measured using an index of four questions: “In my work unit, steps are taken to deal with a poor performer who cannot or will not improve”11 (passive management by exception); “Employees are rewarded for providing high quality products and services to customers”12 (active management by exception); “Awards in my work unit depend on how well employees perform their
9 The authors omit individualized consideration due to debate in the literature about its inclusion in the concept. Although acknowledging the validity of this debate, individualized consideration is included here in an effort to fully test Bass’s theory. 10 The authors combine passive management by exception and active management by exception into a single index. 11 The placement of this question in the survey varied over the years: It was question 24 in 2004 and question 23 from 2006–2011. 12 In 2010, OPM substituted the word “recognized” for “rewarded.” For the purposes of this article, it is assumed that this change has no substantive difference on employees’ responses. The placement of this question within the survey varied over the years; it was question 28 in 2004, question 25 in 2006–2008, and question 31 in 2010–2011.
Journal of Public Administration Research and Theory 418
jobs”13 (contingent reward); and “In my work unit, differences in performance are recognized in a meaningful way”14 (passive management by exception, active man- agement by exception, and contingent reward). Transformational leadership is exam- ined using an index that combines answers from four questions: “My organization’s leaders maintain high standards of honesty and integrity”15 (idealized influence); “In my organization, leaders generate high levels of motivation and commitment in the workforce” (inspirational motivation); “Employees have a feeling of personal empow- erment and ownership of work processes” (individualized consideration); and “I feel encouraged to come up with new and better ways of doing things” (intellectual stimu- lation). Respondents answered these questions on a scale of 1 “Strongly Disagree” to 5 “Strongly Agree.” To examine the “full range” of leadership, the article examines a variable that combines the transactional and transformational leadership indexes.
To determine the extent to which questions within each index converged on their respective latent variables, and the extent to which each index represented a distinct latent variable, a confirmatory factor analysis (CFA) was conducted of each of the leadership indexes.16 In general, it is expected that the questions that make up each of these indexes will converge; however, since Bass does not see transactional and transformational leadership as mutually exclusive, we would not necessarily expect that they are distinct. In fact, the results from three goodness-of-fit tests show some ambiguity about the variables’ convergent/divergent validity.17 Convergent and dis- criminant validity were also assessed using the formulas developed by Fornell and Larcker (1981); according to these tests, the variables have sufficient convergent and discriminant validity.18 Although there are some mixed results here, since the measures have sufficient content validity, and the article uses an analytic approach that exam- ines the leadership types separately and together, they are useful measures for testing Bass’s theory.
To study Bass’s theory over time, the article calculates the average leadership scores for each subagency at time 1 and subtracts them from the subagency’s average scores at time 2. The scale reliability coefficients (Cronbach’s alpha) for index variables, presented in table 1, show that the variables have high levels of reliability. In addition, table 1 shows that a considerable proportion of the variance in each measure’s latent construct is explained by the individual factors; thus, the variables are internally valid.
13 The placement of this question in the survey varied over the years; in 2004 it was question 30, in 2006– 2008 it was question 28, and in 2010–2011 it was question 25. 14 The placement of this question in the survey varied over the years; in 2004 it was question 31, in 2006– 2008 it was question 29, and in 2010–2011 it was question 24. 15 The placement of this question in the survey varied over the years; in 2004 it was question 41, in 2006 it was question 38, in 2008 it was question 39, and in 2010–2011 it was question 54. 16 There is debate about which goodness-of-fit tests should be reported from a CFA (Curran, West, and Finch 1996; Hooper, Coughlan, and Mullen 2008). Nonetheless, the tests that are generally considered the most useful are: the root mean square error of approximation (RMSEA), the comparative fit index (CFI), and the root mean square residual (RMSR). As such, this article reports the findings from these three tests. 17 At the subagency level, across the 6 years examined here, the RMSR, which should generally be lower than 0.08, was 0.00; the CFI, which should generally be above 0.90, was 0.80; the RMSEA, which should generally be lower than 0.08, was 0.30. 18 The Average Variance Extraction (AVE) scores, which should generally be greater than 0.50 to indicate convergent validity, were 0.99; these figures were well above the squared between-factor correlation values in each year, indicating sufficient discriminant validity.
Oberfield Public Management in Time 419
Ta b
le 1
Su m
m ar
y St
at is
ti cs
a n d I n d ex
D es
cr ip
ti o n s
N M
ea n
S D
M in
M ax
C o
o p
er at
io n
27 6
–0 .0
59 0.
09 7
–0 .3
16 0.
14 8
S at
is fa
ct io
n 27
6 0.
03 7
0. 10
1 –0
.2 49
0. 34
6 P
er ce
p ti
o n
s o
f w
o rk
27 6
0. 00
1 0.
06 3
–0 .2
17 0.
23 2
T ra
n sa
ct io
n al
l ea
d er
sh ip
27 6
0. 00
8 0.
09 0
–0 .3
04 0.
33 0
T ra
n sf
o rm
at io
n al
l ea
d er
sh ip
27 6
0. 01
9 0.
08 7
–0 .2
11 0.
35 4
F u
ll r
an ge
o f
le ad
er sh
ip 27
6 0.
01 4
0. 08
2 –0
.2 47
0. 34
2 P
er ce
n t
p ro
fe ss
io n
al 27
6 0.
00 2
0. 01
6 –0
.0 91
0. 13
6 P
er ce
n t
cl er
ic al
27 6
–0 .0
04 0.
02 0
–0 .1
47 0.
13 3
P er
ce n
t te
ch n
ic al
27 6
–0 .0
03 0.
01 6
–0 .0
95 0.
12 0
R es
o u
rc es
27 6
–0 .0
34 0.
14 4
–0 .6
05 0.
71 0
R ac
ia l
d iv
er si
ty 27
6 0.
00 4
0. 01
4 –0
.0 67
0. 05
3 S
ex d
iv er
si ty
27 6
0. 00
0 0.
00 6
–0 .0
37 0.
02 1
T o
ta l
w o
rk fo
rc e
27 6
60 1
27 71
–1 13
54 30
58 8
P ct
t u
rn ov
er 27
6 –0
.0 01
0. 05
5 –0
.5 34
0. 39
1
In d
ep en
d en
t V
ar ia
b le
I n
d ex
D
es cr
ip ti
o n
s T
ra n
sf o
rm at
io n
al L
ea d
er sh
ip T
ra n
sa ct
io n
al L
ea d
er sh
ip F
u ll
R an
ge o
f L
ea d
er sh
ip
Y ea
r 20
04 20
06 20
08 20
10 20
11 20
04 20
06 20
08 20
10 20
11 20
04 20
06 20
08 20
10 20
11 R
el ia
b il
it y
(C ro
n b
ac h
’s
al p
h a)
0. 93
0. 94
0. 94
0. 95
0. 94
0. 91
0. 92
0. 94
0. 94
0. 94
0. 95
0. 95
0. 96
0. 96
0. 96
F ac
to r
lo ad
in g
av er
ag e
0. 90
0. 90
0. 91
0. 90
0. 89
0. 88
0. 88
0. 90
0. 90
0. 89
0. 85
0. 87
0. 88
0. 88
0. 87
E ig
en va
lu e
3. 22
3. 20
3. 29
3. 27
3. 21
3. 17
3. 16
3. 29
3. 27
3. 20
5. 92
6. 04
6. 26
6. 24
6. 13
In te
rn al
v al
id it
y (e
ig en
va lu
e/ n
o.
o f
fa ct
o rs
) 0.
84 0.
80 0.
82 0.
82 0.
80 0.
79 0.
79 0.
82 0.
82 0.
80 0.
74 0.
76 0.
78 0.
78 0.
77
Journal of Public Administration Research and Theory 420
control Variables
As this article assesses its hypotheses, it controls for a variety of organizational char- acteristics. As discussed above, there are strong reasons to expect some level of inertia over time. As such, the article includes a lagged measure of the dependent variables in each model. These variables represent the amount of change in a given dependent variable in the previous period. Since these variables are theorized as being influenced by factors from all three levels of the open systems model, they enable the article to control for change in a vast array of unobserved, and difficult-to-observe, variables (e.g., political support). In all, the article studies four instances of change (2004–2006, 2006–2008, 2008–2010, and 2010–2011); the first three instances will serve as controls, respectively, for the second three instances.
There is theoretical and empirical support for the view that, all else being equal, diversity increases some organizational difficulties (Choi 2009; Choi and Rainey 2010). As such, it is important to control for changes in various elements of agency diversity. To do so, this article includes measures of changing race and sex diversity. It measures race and sex diversity using the Blau index of variability.19 To study the effect of changes in diversity, current measures of subagency diversity are subtracted from past measures.20
Another factor that is important to control for is changes in personnel (Kaufman 1975; O’Toole and Meier 2003). This article controls for this factor by including a variable in its analysis that measures the percentage change in personnel turnover from time 1 to time 2.21 In addition, the article measures the change in personnel size by subtracting the current size of the workforce from the previous workforce size.
One factor that may be associated with changes in employees’ actions and attitudes is the amount of resources that organizations have. As political winds shift, a subagency may lose favored status and funding; without these resources, it may not be possible to achieve its goals and to function properly (Ingraham, Joyce, and Donahue 2003). To control for this, the article includes respondents’ agreement with a single statement from the FHCS and FEVS: “I have sufficient resources (e.g., people, materials, budget) to get my job done.” Respondents answered this question on a scale of 1 “Strongly Disagree” to 5 “Strongly Agree.” By subtracting respondents’ answers at time 1 from their answers at time 2, the article determines the extent to which employees perceived the resource environment as changing over time. Finally, it is also important to control for the type of personnel employed in an agency. To do so, the article measures and controls for the change in the percentage of employees in three different types of employment: profes- sional, technical, and clerical (as classified by OPM [2009]).
19 In its general form, the Blau index of variability is calculated with the following formula: 1 2−∑ pi , where p is the proportion of group members in a population and i is the number of different categories. For measuring racial diversity, this article examines five categories: white, black, Hispanic, Asian/Pacific Islander, and Native Alaskan/American Indian. 20 Although the results of a change in a Blau score do not indicate why diversity changes— i.e., did more minorities enter the workforce—it is useful as an aggregate measure of how diversity changes. Because diversity is generally expected to create difficulties, it is less important why the Blau changes than that it does (or does not) over time. 21 OPM classifies exits into seven categories: quits, discharges, reductions in force, miscellaneous terminations, retirements, and deaths. The measurement used in this article adds each of these categories and calculates the percentage of the workforce that turned over during a specific year.
Oberfield Public Management in Time 421
Findings
The open systems model developed above suggests that employee perceptions, experi- ences, and actions are likely to be strongly inertial. Thus, to begin, it is useful to con- sider the relationship between present and past across each of the outcomes examined here. Correlation analysis shows that the relationship is quite strong for cooperation (r = 0.76), satisfaction (r = 0.82), and perceptions of work quality (r = 0.85). These figures suggest that, for example, subagencies with high levels of satisfaction at time 1 are likely to have high levels of satisfaction at time 2. Similarly, we see a negative relationship between past and present change for cooperation (r = −0.67), satisfaction (r = −0.31), and perceptions of work quality (r = −0.42). These figures show evidence of regression to the organizational mean: Trends from a prior period are reversed in a subsequent period. For example, organizations that increase cooperation from time 1 to time 2 are likely to return to prior levels of cooperation at time 3.
Further evidence of inertia is found in the average subagency change scores in table 1: The average change for each outcome was less than one tenth of a point out of a possible 4 points.22 Even the minimum and maximum changes are small: During the time period covered by this study, the most that any subagency changed in follower cooperation, satisfaction, and perceptions of work quality was less than a single point out of a possible 4 points.
Although the volatility over time was generally low across all three outcomes, there was more volatility in follower satisfaction and cooperation than in perceptions of work quality: The average change in perceptions of work quality was one-thou- sandth of a point (out of 4 points), the average change in cooperation was six-hun- dredths of a point, and the average change in satisfaction was four-hundredths of a point. Taken together, these results suggest that, all else being equal, it may be more difficult for managers to create a sense of work quality improvement over time than it is for them to improve follower cooperation and satisfaction.
As with the dependent variables, correlation analysis shows that followers’ views of management were inertial: The present values of all leadership indexes were strongly correlated with their lagged values (transactional leadership, r = 0.87; transformational leadership, r = 0.91; full range of leadership, r = 0.90). Also, there was a negative relationship between past and present change for all independent var- iables (transactional leadership, r = −0.17; transformational leadership, r = −0.22; and full range of leadership r = −0.17); thus, again we see evidence of regression to the mean after some year-to-year fluctuation. Further evidence of management inertia is found in table 1; the mean subagency change score for each management variable was quite low: eight-thousandths of a point for transactional leadership, two-hundredths of a point for transformational leadership, and one-hundredth of a point for the full range of leadership (all out of a possible 4 points). Although all are low, these figures suggest that there is greater volatility in transformational leadership over time.
Putting this evidence together, we begin to develop a picture of employee percep- tions and actions over time: There is some evidence of change but the scope of change
22 Since responses for all outcomes spanned from 1 to 5, the maximum change is 4 points.
Journal of Public Administration Research and Theory 422
is modest. Thus, these descriptive findings suggest that the theorized influences from figure 1 exert strong inertial pressures on employees. Nonetheless, as outlined above, there are good reasons to expect that management change is associated with changes in followers’ actions and experiences. As such, table 2 presents findings from a sta- tistical analysis that measures the relationships between changes in leadership and changes in the article’s three dependent variables. On the left side of the table (columns 1–3), the models include the transactional and transformational leadership variables separately; on the right side of the table (columns 4–6), the models test the effects of the combined full range of leadership variable.
The results in table 2 show support for the hypotheses drawn above. On the left side of the table, we see that, when separated (columns 1–3), increases in transforma- tional and transactional leadership are associated with improvements in each out- come. In comparison, transformational leadership is more powerful: An increase of 1 point in transformational leadership is associated with approximately 0.22 of a point in cooperation, 0.25 of a point in improvement in perceptions of work quality, and 0.64 of a point in improvement in satisfaction. In contrast, an increase of 1 point in transactional leadership is estimated to be associated with around 0.14 of a point in improvement in perceptions of work quality, 0.15 of a point in improved cooperation, and 0.20 of a point in increased satisfaction. On the right side of the table, we see evi- dence that increases in the full range of leadership are associated with improvement in follower cooperation, satisfaction, and perceptions of work quality. An increase of 1 point in the full range of leadership is associated with 0.42 of a point in cooperation, 0.77 of a point in satisfaction, and 0.39 of a point in perceptions of work quality.
Since one of the rationales for this article is understanding management in time, it is also interesting to compare the above results with results from a cross-sectional analysis. In this way, we can gauge the extent to which atemporal models overvalue the effect of management in comparison with the models used here. The cross-sectional analyses, which are not presented here due to space limitations, suggest that excluding time generally inflates the effect of management.23 For example, in the cross-sectional analysis, a 1-point increase in transformational leadership was associated with a 0.32- point increase in perceptions of work quality; in the longitudinal change analysis, a 1-point increase in this type of leadership was associated with a 0.25-point increase in perceived work quality. Although this example highlights the general trend, in two of nine instances, the cross-sectional models undervalued the effect of leadership. Though the findings are not uniform, together they suggest that analyzing manage- ment outside of a temporal context can alter its perceived effect.
Because the 2010 dummy variables in table 2 are statistically significant in all models, and because the last change period was 1 year instead of 2, the results were checked for temporal robustness: unpooled year-by-year change analyses were under- taken for the years in which there were sufficient observations. In each of these analyses, the same models from table 2 were employed. In each year, all independent variables were positive, significant predictors of all three dependent variables at p < .10 level. In the separated models, improved transformational leadership was typically the
23 The results of the cross-sectional analyses are available from the author upon request.
Oberfield Public Management in Time 423
Ta b
le 2
P re
d ic
ti n g
C h an
ge s
in F
o llo
w er
C o o p er
at io
n , S
at is
fa ct
io n , a
n d W
o rk
Q u al
it y
1 2
3 4
5 6
C o
o p
er at
io n
S at
is fa
ct io
n W
o rk
Q u
al it
y C
o o
p er
at io
n S
at is
fa ct
io n
W o
rk Q
u al
it y
In d
ep en
d en
t va
ri ab
le s
T
ra n
sf o
rm at
io n
al
l ea
d er
sh ip
0.
22 1*
** (
0. 05
7) 0.
63 7*
** (
0. 07
2) 0.
25 3*
** (
0. 07
8) —
— —
T
ra n
sa ct
io n
al
l ea
d er
sh ip
0.
20 2*
** (
0. 04
7) 0.
15 3*
* (0
.0 59
) 0.
13 9*
** (
0. 05
3) —
— —
F
u ll
r an
ge o
f
l ea
d er
sh ip
— —
— 0.
42 2*
** (
0. 04
1) 0.
76 6*
** (
0. 05
7) 0.
38 6*
** (
0. 05
0)
C o
n tr
o l
va ri
ab le
s
D ep
en d
en t
va ri
ab le
,
l ag
− 0.
29 7*
** (
0. 05
1) −
0. 12
7* **
( 0.
03 8)
− 0.
39 3*
** (
0. 06
1) −
0. 29
7* **
( 0.
05 1)
− 0.
12 3*
* (0
.0 41
) −
0. 39
5* **
( 0.
06 1)
P
er ce
n t
p ro
fe ss
io n
al
0. 25
9 (0
.2 31
) 0.
10 8
(0 .3
53 )
0. 23
7 (0
.3 87
) 0.
25 6
(0 .2
31 )
0. 03
6 (0
.3 46
) 0.
22 0
(0 .3
80 )
P
er ce
n t
cl er
ic al
− 0.
14 1
(0 .1
60 )
− 0.
09 7
(0 .1
54 )
− 0.
12 8
(0 .1
90 )
− 0.
13 7
(0 .1
57 )
0. 01
1 (0
.1 77
) −
0. 10
4 (0
.1 85
)
P er
ce n
t te
ch n
ic al
− 0.
22 9
(0 .2
18 )
0. 17
1 (0
.3 06
) −
0. 24
6 (0
.3 13
) −
0. 22
5 (0
.2 14
) 0.
28 3
(0 .2
93 )
− 0.
22 1
(0 .3
06 )
R
es o
u rc
es −
0. 01
4 (0
.0 22
) 0.
05 7*
( 0.
02 7)
− 0.
06 9*
* (0
.0 27
) −
0. 01
3 (0
.0 21
) 0.
08 7*
* (0
.0 29
) −
0. 06
2* (
0. 02
6)
R ac
ia l
d iv
er si
ty
0. 10
7 (0
.2 97
) 0.
24 4
(0 .4
36 )
− 0.
02 2
(0 .3
55 )
0. 10
9 (0
.2 94
) 0.
30 3
(0 .4
93 )
− 0.
00 9
(0 .3
65 )
S
ex d
iv er
si ty
− 0.
79 4
(0 .6
29 )
0. 59
1 (0
.8 21
) −
0. 47
7 (0
.6 96
) −
0. 75
8 (0
.6 01
) 1.
50 7
(0 .8
21 )
− 0.
26 3
(0 .6
70 )
T
o ta
l w
o rk
fo rc
e
0. 00
0* **
( 0.
00 0)
− 0.
00 0
(0 .0
00 )
0. 00
0* *
(0 .0
00 )
0. 00
0* **
( 0.
00 0)
− 0.
00 0
(0 .0
00 )
0. 00
0* (
0. 00
0)
P er
ce n
t tu
rn ov
er
0. 05
9 (0
.0 45
) −
0. 06
4 (0
.0 57
) 0.
03 3
(0 .0
70 )
0. 05
9 (0
.0 46
) −
0. 06
6 (0
.0 73
) 0.
03 3
(0 .0
67 )
20
08
0. 01
1 (0
.0 11
) 0.
02 0
(0 .0
14 )
0. 01
6 (0
.0 17
) 0.
01 1
(0 .0
12 )
0. 01
3 (0
.0 15
) 0.
01 4
(0 .0
17 )
20
10 −
0. 12
5* **
( 0.
01 1)
0. 07
5* **
( 0.
00 6)
0. 03
5* **
( 0.
00 8)
− 0.
12 5*
** (
0. 01
0) 0.
08 1*
** (
0. 00
6) 0.
03 7*
** (
0. 00
7)
C o
n st
an t
− 0.
03 2*
** (
0. 00
8) −
0. 00
3 (0
.0 04
) −
0. 02
3* **
( 0.
00 5)
− 0.
03 2*
** (
0. 00
8) −
0. 00
1 (0
.0 05
) −
0. 02
2* **
( 0.
00 5)
R
2 (w
it h
in )
0.
89 9
0. 85
5 0.
66 0
0. 89
9 0.
83 7
0. 65
7 N
ot e:
N =
2 76
s u
b ag
en cy
-y ea
rs d
ra w
n f
ro m
1 31
d is
ti n
ct s
u b
ag en
ci es
. T
h e
d ep
en d
en t
va ri
ab le
s ar
e ch
an ge
s in
a ve
ra ge
c o
o p
er at
io n
, sa
ti sf
ac ti
o n
, an
d p
er ce
p ti
o n
s o
f w
o rk
q u
al it
y at
t h
e su
b ag
en cy
le
ve l f
ro m
t im
e 1
to t
im e
2. O
L S
r eg
re ss
io n
c o
ef fi
ci en
ts g
en er
at ed
f ro
m fi
xe d
-e ff
ec ts
m o
d el
s in
w h
ic h
s ta
n d
ar d
e rr
o rs
( in
p ar
en th
es es
) ar
e cl
u st
er ed
b y
su b
ag en
cy . S
ta ti
st ic
al s
ig n
ifi ca
n ce
d en
o ta
ti o
n
is :
** *p
< .
00 1,
* *p
< .
01 ,
*p <
. 05
.
Journal of Public Administration Research and Theory 424
stronger predictor of improvement. As such, the article’s main findings are robust across the times examined here.
In addition to evaluating the article’s hypotheses, it is useful to also discuss two control variables in table 2. First, it is interesting to see that, as in the change corre- lation analysis discussed above, the lagged dependent variables had strong, negative associations with the dependent variables. This suggests that, even when controlling for changes in leadership, there is a tendency for followers to perceive regression to a prior mean. The power of the lagged dependent variables appeared to vary across the three outcomes in a pattern that matches the descriptive findings above: It was most powerful for perceptions of work quality, somewhat less for cooperation, and considerably less for satisfaction. It is also interesting to compare the coefficients of the lagged change variables with the management coefficients. On the right side of the table, we see that prior cooperation change and prior satisfaction change were weaker predictors than the full range of leadership variable. However, the lagged work qual- ity perception change variable was a slightly stronger predictor of work quality per- ception change than the leadership coefficient. These findings suggest that it may be easier for leadership to shift employee cooperation and satisfaction than it is to shift employee perceptions of work quality.
Second, it is interesting to note that followers’ perceptions of resources had a mixed association with the three outcomes examined here. As might be expected, when followers perceived an increase in resources, there was a corresponding increase in satisfaction. However, when followers in a subagency perceived an increase in the amount of resources, they were also likely to perceive a decrease in work quality. Also, resource perceptions were unassociated with follower cooperation. It is not clear how to interpret these findings since one would expect them to be unidirec- tional. However, these findings suggest that organization resources, which are gener- ally understood as having a positive effect on organization functioning, may have a complex relationship with how followers think and act. For instance, we may specu- late that greater resources come with “strings attached” like higher expectations or increased task requests. Thus, employees may benefit from increased resources but also see a decline in output quality. Also, it is important to note that, in comparison to management, resource perceptions are relatively weak predictors of the three out- comes examined here. For example, on the left side of the table, we see that a 1 unit increase in resources was associated with around a tenth of a point in satisfaction change (positive) and perceived work quality change (negative). However, in these same models, improvements in leadership were associated with between 0.14 and 0.63 of a point in change. This suggests that management matters even in the context of a changing resource environment.
Taken together, the findings here suggest that follower actions, experiences, and perceptions are strongly inertial over time. Nonetheless, there is also strong support for Bass’s model: “lower order” transactional leadership strategies, which show fol- lowers that managers are paying attention and responding in kind, and “higher order” transformational strategies, which inspire followers and stimulate them, appear to matter for followers. In comparison, as Bass predicts, transformational leadership appears to be the more potent management strategy.
Oberfield Public Management in Time 425
concLusion
Do public managers play a central role in shaping the actions and attitudes of fol- lowers? When they adopt transformational leadership strategies, can they expect to see changes? In the public administration literature, there is a general consensus that, though public managers are constrained, the answer to both these questions is “yes.” However, the evidence in support of these answers is somewhat limited: Though man- agement is theorized as a temporal process, it is typically studied at a single moment in time or, when studied over time, within one part of a government. Thus, there is little evidence linking management change and follower change across a government over time. To study management temporally, this article focuses on one of the most established and popular leadership theories, Bass’s full range theory; to ensure a broad scope, the article examines a diverse array of subagencies across the US federal gov- ernment. To frame its analysis, the article uses an open systems model that accounts for a variety of unobserved, time-invariant influences on followers. The model permits space for leadership to matter but also seeks to control for the inertial components of organizational life.
Before discussing the article’s findings and implications, it is important to convey its context. This study focuses on a particular time period at one level of government in one nation. Though the findings are robust across the time period studied here, and the subagencies are drawn from a wide spectrum of government agencies, it is possible that its findings are not generalizable to other time periods, levels of govern- ment, and nations. Also, though Bass’s theory is well developed and supported by considerable empirical evidence, it is obviously not all-encompassing. There are many other management strategies that are theorized as important for organization func- tioning; future work should seek to explore these theories temporally across a large assortment of public organizations. Additionally, this article focuses on three discrete follower outcomes. However, future public management studies should examine other outcomes to see how they change in response to management change. Finally, the change periods examined here (which range from 1 to 2 years) are relatively short. Future studies should examine change over a longer period of time.
Nonetheless, within the boundaries of its empirical investigation, the article has a variety of intriguing findings. To begin, it shows that there are strong patterns of employee actions, experiences, and perceptions over time. At the level of theory, these findings portray organizations, and the employees who make them up, as embedded in strong, particular traditions. These traditions are not unalterable, but they are likely to be serious forces shaping how individuals think and act. Such a finding, in one sense, is not particularly surprising: Organizations have long been seen as useful due to their ability to bring stability to the vicissitudes of the social world (Weber 1947; Scott 2003). Also, theorists have argued that public management must be understood in its temporal context (O’Toole and Meier 1999). Nonetheless, relatively few public man- agement studies take time into account. As such, the findings here are a reminder that management should be understood as one of many competing forces that shape how a public organization functions. As the results here show, when scholars examine public management over time, and control for a broad array of unobserved, time-invariant characteristics, they may develop a different picture of its power.
Journal of Public Administration Research and Theory 426
In addition, the findings here suggest that leadership, at least from the perspec- tive of followers, tends to stay the same over time: Organizations with leaders who are seen as transformational at one time are very likely to be seen that way at another time. Thus, even if leadership has the potential to shape followers, it may not, due to inertia. From a theoretical perspective, this finding implies that, although leadership is typically understood as a lever for creating change, it may not be as nimble a tool as it seems. In fact, leadership may serve to inhibit change. This finding begs two further questions: What explains the genesis of followers’ perceptions of leadership within an organization? Also, once established, what actions do leaders take to maintain or break this momentum? Perhaps leadership traditions emerge from agency tasks or entering cohorts of managers. Perhaps they begin based on agency reputation. Whatever the reasons, it would be useful for future studies to ask about the origins of followers’ perceptions of leadership and the reasons why they might change.
Despite the strong tendency toward continuity, the article presents evidence sup- porting Bass’s theory: Improvements in transformational and transactional leadership, separately and together, are positive predictors of follower cooperation, satisfaction, and perceptions of work quality. Thus, even in the face of considerable inertia, it is possible for management strategies to shape followers. More specifically, the article shows that improvement in transformational leadership is, when controlling for a vari- ety of other aspects of management and agency characteristics, a stronger predictor across each outcome examined here. This suggests that public managers, after estab- lishing traditions of transactional leadership, should consider how they might satisfy their followers’ “higher order” needs via inspiration, individual attention, and intel- lectual stimulation. In these ways, public managers may be able to influence followers and shape how they act. Aggregated at the organizational level, as Bass expects, these strategies may improve how public organizations function.
Finally, although this article shows that leadership generally matters across the three outcomes examined in this article, there were important differences. Subagencies in which managers shifted toward the full range of leadership saw much larger gains in employee satisfaction than in cooperation and perceptions of work quality. This finding suggests that, although it is generally expected that the full range of leadership is associated with positive follower results, it may shape some follower outcomes more strongly than others. Perhaps, as other recent work suggests (Moynihan, Pandey, and Wright 2012), leadership has an indirect relationship with organization behavior: It may shape employee behavior by first affecting how they experience their organiza- tions. Whatever the pattern, future work should temporally investigate leadership’s relationship with a diverse array of behavioral and perceptual outcomes. In this way, we can develop a better understanding of how public managers influence followers and, perhaps, their organizations over time.
REFEREncEs
Adcock, R., and D. Collier. 2001. Measurement validity: A shared standard for qualitative and quantitative research. American Political Science Review 95:529–46.
Barnard, C. 1968. The functions of the executive. Cambridge, MA: Harvard Univ. Press. Bass, B. 1996. Is there universality in the full range model of leadership? International Journal of
Public Administration 19:731–61.
Oberfield Public Management in Time 427
Behn, R. 2006. Performance leadership: 11 better practices that can ratchet up performance. Washington, DC: IBM Center for the Business of Government.
Boyne, G., and R. Walker. 2002. Total quality management and performance: An evaluation of the evidence and lessons for research on public organizations. Public Performance & Management Review 26:111–31.
Brehm, J., and S. Gates. 1997. Working, shirking, and sabotage: Bureaucratic response to a democratic public. Ann Arbor: Univ. of Michigan Press.
Brewer, G. 2006. All measures of performance are subjective: More evidence on U.S. federal agencies. In Public service performance: Perspectives on measurement and management, eds. G. Boyne, K. Meier, L. O’Toole, and R. Walker. New York: Cambridge Univ. Press.
Burns, J. 1978. Leadership. New York: Harper Collins. Choi, S. 2009. Diversity in the US federal government: Diversity management and employee turno-
ver in federal agencies. Journal of Public Administration Research and Theory 19:603–30. Choi, S., and H. Rainey. 2010. Managing diversity in U.S. federal agencies: Effects of diversity
and diversity management on employee perceptions of organizational performance. Public Administration Review 70:109–21.
Curran, P., S. West, and J. Finch. 1996. The robustness of test statistics to nonnormality and speci- fication error in CFA. Psychological Methods 1:16–29.
Fernandez, S. 2005. Developing and testing an integrative framework of public sector leadership: Evidence from the public education arena. Journal of Public Administration Research and Theory 15:197–217.
Fornell, C., and D. Larcker. 1981. Evaluating structural equation models with unobservable vari- ables and measurement error. Journal of Marketing Research 18:39–50.
Frederickson, H., K. Smith, C. Larimer, and M. Licari. 2012. The public administration theory primer. Boulder, CO: Westview Press.
Hall, Richard H. 1977. Organizations: Structure and process, 2nd ed. Englewood Cliffs, NJ: Prentice-Hall.
Hirschman, A. 1970. Exit, voice, and loyalty: Responses to decline in firms, organizations, and states. Cambridge, MA: Harvard Univ. Press.
Hooper, D., J. Coughlan, and M. Mullen. 2008. Structural equation modeling. The Electronic Journal of Business Research Methods 6:53–60.
Ingraham, P., P. Joyce, and A. Donahue. 2003. Government performance: Why management matters. Baltimore, MD: Johns Hopkins Univ. Press.
Kaufman, H. 1975. The limits of organizational change. Tuscaloosa, AL: Univ. of Alabama Press. Kim, S. 2002. Participative management and job satisfaction: Lessons for management leadership.
Public Administration Review 62:231–41. Kingdon, J. 2002. Agendas, alternatives, and public policies. New York: Longman Publishing Group. Lipsky, M. 1980. Street-level bureaucracy: Dilemmas of the individual in public service. New York:
Russell Sage Foundation. Lowe, K., K. Kroeck, and N. Sivasubramaniam. 1996. Effectiveness correlates of transformational
and transactional leadership: A meta-analytic review of the MLQ literature. The Leadership Quarterly 7:385–425.
March, J., and J. Olsen. 1984. The new institutionalism: Organizational factors in political life. American Political Science Review 78:734–49.
———. 2010. Logic of appropriateness. In The Oxford handbook of American bureaucracy, ed. R. Durant. New York: Oxford Univ. Press.
Marks, M., J. Mathieu, and S. Zaccaro. 2001. A temporally based framework and taxonomy of team processes. The Academy of Management Review 26:356–76.
Mead, L. 2004. Government matters: Welfare reform in Wisconsin. Princeton, NJ: Princeton Univ. Press.
Meier, K., L. O’Toole, G. Boyne, and R. Walker. 2007. Strategic management and the performance of public organizations: Testing venerable ideas against recent theories. Journal of Public Administration Research and Theory 17:357–77.
Journal of Public Administration Research and Theory 428
Merton, R. 1940. Bureaucratic structure and personality. Social Forces 18:560–8. Milton, L., and J. Westphal. 2005. Identity confirmation networks and cooperation in work groups.
Academy of Management Journal 48:191–212. Moynihan, D. 2007. The reality of results: Managing for results in state and local government. In
In pursuit of performance: Management systems in state and local government, ed. P. Ingraham. Baltimore, MD: Johns Hopkins Univ. Press.
Moynihan, D., and S. Pandey. 2005. Testing how management matters in an era of government by performance management. Journal of Public Administration Research and Theory 15:421–39.
Moynihan, D., S. Pandey, and B. Wright. 2012. Setting the table: How transformational leadership fosters performance information use. Journal of Public Administration Research and Theory 22:143–64.
Nicholson, C. 2010. The Dodd-Frank Bill up close. The New York Times. http://dealbook.nytimes. com/2010/06/28/the-dodd-frank-bill-up-close/.
O’Toole, L., and K. Meier. 1999. Modeling the impact of public management: Implications of struc- tural context. Journal of Public Administration Research and Theory 9:505–26.
———. 2003. Plus ça change: Public management, personnel stability, and organizational perfor- mance. Journal of Public Administration Research and Theory 13:43–64.
Organ, D., and A. Lingl. 1995. Personality, satisfaction, and organizational citizenship behavior. The Journal of Social Psychology 135:339–50.
Paarlberg, L., and B. Lavigna. 2010. Transformational leadership and public service motivation: Driving individual and organizational performance. Public Administration Review 70:710–8.
Perry, J., and N. Buckwalter. 2010. The public service of the future. Public Administration Review 70:s238–45.
Pitts, D. 2009. Diversity management, job satisfaction, and performance: Evidence from U.S. federal agencies. Public Administration Review 69:328–38.
Posner, R. 2005. Preventing surprise attacks: Intelligence reform in the wake of 9/11. New York: Rowman & Littlefield.
Rago, W. 1994. Adapting total quality management (TQM) to government: Another point of view. Public Administration Review 54:61–64.
Rainey, H. 2003. Understanding and managing public organizations, 3rd ed. New York: Jossey-Bass. Reisner, R. 2011. A leader’s guide to transformation: Developing a playbook for successful change
initiatives. IBM Center for the Business of Government. Riccucci, N. 2005. How management matters: Street-level bureaucrats and welfare reform. Washington,
DC: Georgetown Univ. Press. Riggio, R. 2009. Are you a transformational leader? Psychology Today. http://www.psychologyto-
day.com/. Scott, W. 2003. Organizations: Rational, natural, and open systems. Saddle River, NJ: Prentice Hall. Simon, H. 1997. Administrative behavior: A study of decision-making processes in administrative
organizations. New York: Free Press. Smith, K., S. Carroll, and S. Ashford. 1995. Intra-and interorganizational cooperation: Toward a
research agenda. Academy of Management Journal 38:7–23. Swiss, J. 1992. Adapting total quality management (TQM) to government. Public Administration
Review 52:356–62. Trottier, T., M. Van Wart, and X. Wang. 2008. Examining the nature and significance of leadership
in government organizations. Public Administration Review 68:319–33. US Office of Personnel Management (OPM). 2009. Introduction to the position classification stand-
ards. Washington, DC: OPM. ———. 2012. Federal employee viewpoint survey: What is the FedView survey? http://www.fedview.
opm.gov/2011/. Van Wart, M. 2003. Public-sector leadership theory: An assessment. Public Administration Review
63:214–28.
Oberfield Public Management in Time 429
Vigoda-Gadot, E., and I. Beeri. 2012. Change-oriented organizational citizenship behavior in public administration: The power of leadership and the cost of organizational politics. Journal of Public Administration Research and Theory 22:573–96.
Weber, M., trans. T. Parsons. 1947. The theory of social and economic organization. New York: Free Press.
Williamson, V., T. Skocpol, and J. Coggin. 2011. The tea party and the remaking of republican con- servatism. Perspectives on Politics 9:25–43.
Wilson, J. 1989. Bureaucracy: What government agencies do and why they do it. New York: Basic Books.
Wright, B., and B. Davis. 2003. Job satisfaction in the public sector. The American Review of Public Administration 33:70–90.
Wright, B., D. Moynihan, and S. Pandey. 2012. Pulling the levers: Transformational leadership, pub- lic service motivation, and mission valence. Public Administration Review 72:206–15.
Wright, B., and S. Pandey. 2010. Transformational leadership in the public sector: Does structure matter? Journal of Public Administration Research and Theory 20:75–89.
Copyright of Journal of Public Administration Research & Theory is the property of Oxford University Press / USA 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.
