Abstract
In a volatile political climate where agencies often become the stage for partisan battles, career employees pay the price for increased political attention to agencies. This study seeks to understand how contentious political environments contribute to employee turnover intent. It draws on strategic management and systems theories that link resource turbulence to management changes and employee turnover. The question is important to address because career bureaucrats are critical to government performance as both policy makers and implementers. Individual-level data are paired with organizational-level data to assess how political environments influence turnover preferences. Other aspects of an agency’s political environment (age, design, ideology, and political appointees) are also tested. The analysis shows that budget volatility increases the intent to leave, but staffing volatility increases intent to move within government. All political environment variables except ideology affected intent to change jobs.
As political discourse dissolves into partisan rancor and holds administrative agencies hostage with shutdowns and threats, we must look at the fundamental risk to democracy and effective government posed by battering agencies and their employees. Although most of the public’s attention focuses on legislative activities and the news of the day, the battleground for policies and programs is often inside government agencies. Presidents actively seek to mobilize agencies to support their policies. Congress gives agencies the responsibility of implementing programs and policies consistent with congressional intent. Courts step in to serve as the final arbitrators to resolve the differences between the executive and legislative branches.
The central theme of this article is to examine what happens to federal government employees in politically contentious times. Specifically, the question is as follows:
The question is important to address because the career bureaucrats are critical to government performance as both policy makers and implementers. When organizations experience high turnover, they risk losing institutional knowledge and expertise necessary for seamless service to the public. Turnover can be expensive in terms of replacement and training costs, institutional knowledge, and staff morale (Abbassi & Hollman, 2000; Cascio, 2006; Partnership for Public Service & Booz Allen Hamilton, 2010). Expressions of turnover intent require attention because they are symptoms of underlying problems within an organization. Employees’ frustration due to high turnover intent could damage organizational productivity, even if people do not choose to leave (Krueger & Killham, 2006).
Federal agencies can be politically contentious when the preferences of the president are at odds with the preferences of the Congress. Federal agencies can be caught amid the political maelstrom as a result. The U.S. Constitutional separation of powers divides the power between the executive and the legislative for a system of checks and balances. The system divides control of resources, so that Congress has the power to authorize funding and the executive has control over authorizing staffing levels. Although there are instances when the two branches seek to influence the resources provided by the other, the roles of Congress and the executive are well delineated. The mismatched and volatile resource streams provided by the Congress and the executive could affect federal agencies negatively, particularly when the resources are constrained to very low levels by either body.
At the same time, political contentions do not affect all federal agencies in similar ways. Agencies such as the Department of Education are in political maelstrom due to differing ideological beliefs about how to improve education and the appropriate role of government. Some agencies such as the Nuclear Regulatory Commission operate in relative isolation due to their highly technical role, largely outside of public attention. Other agencies have come into scrutiny when their missions are politically salient. For instance, the U.S. Customs and Border Patrol drew little political attention until the Trump administration made border security central to its administrative agenda. Career employees of organizations whose environments are in the midst of increased political attention, whether positive or negative, will have different professional experiences than employees of organizations with less attention where a steady climate is the norm.
Systems theory and strategic management approaches offer guidance on how agencies adjust their structures and behaviors to weather the political maelstrom. Most currently used concepts of how the political world interacts with agencies are based on variables that are largely time invariant or rarely fluctuate much within an organization. This ignores the reality that political status can quickly change along with administrations and political leadership. From an open systems perspective, Pfeffer and Salancik’s (1978) resource dependency theory asserts that organizations are influenced by their environments and work to develop contingencies to protect against problems associated with dependency on external actors for resources. Strategic management approaches posit that managers actively work to reduce environmental uncertainty and dependence (Hillman, Withers, & Collins, 2009). Management changes associated with turbulence could then lead to turnover (Cameron, Kim, & Whetten, 1987; Harrison, Torres, & Kukalis, 1988). Harrison (2004) even suggests that turnover is a critical output of open systems.
Systems theory and strategic management approaches are used in this article to develop a series of hypotheses on how the different dimensions of political turbulence and environments influence career bureaucrats’ intent to stay or leave. Additional hypotheses are developed using the literatures that consider the administrative presidency, congressional control and bureaucratic responsiveness, and agency design. Hypotheses related to six such dimensions of political context are tested. These dimensions relate to agency budget volatility, staffing volatility, agency age, agency level (e.g., cabinet level), agency political ideology, and political appointments in the agency. Control variables include agency size and individual-level factors, such as pay satisfaction, tenure, age, sex, and minority status. Finally, emerging literature in public administration shows that there are multiple forms of voluntary turnover options—moving to a new agency within the federal government, leaving the government for a new job, and leaving the government for other futures such as retirement. Each type of turnover is driven by different factors. Hence, this article also examines whether political turbulence influences people toward some future job preferences over others. The hypotheses are tested in a novel way by attaching organizational-level data to individual responses from the Federal Viewpoint Survey (FEVS).
Literature Review
Turnover intent is generally viewed through an intraorganizational lens, where individual demographic traits are paired with organizational behaviors and satisfaction levels. The motivation literature provides numerous elements that influence turnover choices. These factors find substantial support in the literature, but still only explain a piece of the job change decision. The list of variables and conceptualizations is long; it includes job satisfaction, pay satisfaction, opportunities for advancement, fit with various people in the organization and the organization itself, goal ambiguity, intrinsic motivation, and connection with an organization’s mission (for examples, see Bertelli, 2007; Kristof, 1996; Rainey & Steinbauer, 1999; Romzek, 1990; Ting, 1997; Wright & David, 2003). By examining the political environment as influencing organization structure and behavior and, thus, influencing employee career choices, this study offers a more expansive perspective on turnover intent.
There are multiple ways that scholars assess the political environment of federal agencies. Many of these studies propose that the political environment influences organizational behavior and management. Some studies directly study careerist reactions and responses to elected and appointed officials, but few explore whether the impacts of the political environment contribute to careerist turnover decisions. Currently used ideas about political environments are largely time invariant or have little change from one year to another such as when an agency was founded, agency placement in the president’s cabinet, and agency ideological leanings. The other widely considered factor is the use of appointees to politicize an agency and achieve presidential goals. None of these is thought to be a single proxy for the political nature of an agency. Instead, they are pieces of the puzzle that individually tell parts of the story of what it means for career employees to work in political organizations. What is missing is that agencies are often caught in a tug-of-war between the administration and Congress, and that this struggle manifests in volatility or turbulence. There is a growing recognition in the governance literature that public organizations experience turbulence as a result of highly partisan and contentious political environments that offer diminished support for agencies (Ansell, Trondal, & Øgård, 2017). Building on the strategic management literature from business and Oxford English Dictionary definition of turbulence as “a state of conflict or confusion,” this study incorporates environmental turbulence from the strategic management and open systems literatures to say that political environments affect careerists where people working in agencies that are in a state of conflict or confusion caused by resource volatility will make different career choices than those in stable organizations. The source of this resource volatility is the political attention and action of the White House and Congress through their respective allocations of funding and staffing.
Assessing Political Environments by Turbulence
A key element of this research is to link the study of turnover intent to the political environment. Propositions in the systems literature provide the nexus between environments, organizations, and employees. The open systems approach was first presented in biology by Von Bertalanffy (1950) and later extended to organizations (Burns & Stalker, 1961/1994; Katz & Kahn, 1966; Pfeffer & Salancik, 1978; Scott, 1961). It conceptualizes organizations as nested in a broader open system where there is an endogenous relationship between organizations and environments.
A central premise of systems theory is that organizational form, function, and structure are influenced by the stability or lack of stability in the system (Burns & Stalker, 1961/1994; Hannan & Freeman, 1977). Burns and Stalker (1961/1994) view organizations as landing on a continuum that moves from mechanistic to organic operations where placement is dictated by the stability of the system. Where conditions are stable, a mechanistic organizational form emphasizing standard operating procedures and traditional hierarchy will emerge. On the opposite end, conditions that are volatile or unstable call for organic systems with less rigidity and more employee participation. This flexibility becomes necessary for the adaptation required in the environment.
In Minztberg’s (1979) terms, the form of the bureaucracy is highly dependent on environments. He describes five types of bureaucracies that emerge from environmental demands. Each type differs based on factors including level of centralization, preferred management styles, how jobs are structured, and which part of the organization is strongest. It is here that we make the jump from the environment to internal operations. Mintzberg continues, explaining that the internal operations then directly alter the individual employee’s experience by changing communications, job structures, and resource availability. If turnover is an expression of seeking one’s preferred work, individuals will want to change jobs when the internal operations become frustrating.
As the political environment that agencies function within varies from one agency to another, it is possible that the environment requires agencies to adjust their working styles/structures and, thus, influences individuals. This, in turn, may contribute to differences in turnover intent rates. In 1989, Wilson offered an application of systems theory to public organizations in his four-part typology. He asserted that a public agency’s relationship to elected officials is defined by its ability to clearly articulate and report both outcomes and outputs. Where it is difficult to report these outcomes and outputs or where they are controversial, the agency is subject to greater political scrutiny. Where tasks are more controversial, Wilson proposes that increased political scrutiny will create a turbulent and complex political environment. However, agencies with clearly measurable outputs and outcomes lend themselves to more standardized, routine, and stable environments. Wilson continues to propose that agency management will adapt to the demands of the political environment and thereby alter internal structures, behaviors, and management approaches accordingly. When viewing Wilson’s typology alongside Minztberg’s (1979) assertions about how organizational structures change in response to turbulence, we see a direct link from volatile or stable political environments to differences in management. The final link in the chain from turbulent environments to turnover is then offered by private sector literature that finds increased turnover in more turbulent environments (Cameron et al., 1987; Harrison et al., 1988).
Although the systems literature offers pathways connecting environments to organizations and individuals, there is direct evidence of how individuals and organizations respond to political environments in the bureaucratic responsiveness and congressional control literatures. A key indicator of awareness of the demands placed on agencies by the political world is agency responsiveness to elected officials. In general, the literature on bureaucratic responsiveness to elected officials shows that bureaucrats are responsive, but that their responses can be tempered by other factors. This awareness of political demands and responsiveness suggests that agencies do adjust to their political environment as suggested by Wilson (1989) and other systems theorists. Headrick, Serra, and Twombly (2002) find that bureaucrats recognize that Congress actively uses the budget as a control measure and that budgetary decreases remove resources needed for enforcement and increases serve as recognition that an agency’s work is important. Bureaucrats from the street-level up have an interest in placating congressional committee members to gain resources and avoid negative repercussions. Golden (2000) found that bureaucrats were responsive to the political world, even in extreme situations. In her case study of federal agencies under the Reagan administration including the Environmental Protection Agency, she discusses how careerists were responsive to appointees and the administration when they sought to completely reverse the course of the agency.
These perspectives paint a picture of elected and appointed officials placing the bureaucracy in the position of being pulled in multiple directions by various actors. When agencies become the focus of political actors, increased attention will result in variation in resources compared with the stability that occurs when a lack of attention maintains the status quo as that norm. The literatures from the administrative presidency and bureaucratic responsiveness show that careerists are aware of the desires of various political actors, and respond accordingly. Consistent with research on strategic management and open systems, these pushes and pulls require agencies to adapt their internal structures including centralization, management approaches, and job structures, so too will they require career staff to adjust to environmental demands. Thus, a highly politicized and volatile environment will create a less stable and predictable workplace that will then influence careerist employment decisions.
Two measures of volatility used here are budget and staffing changes. Budget volatility is a measure of congressional attention as the budget is set by Congress. Agency budget can be regarded as a measure of agency support from Congress as the literature tells us that the budget is used as one of the key methods of congressional support and control (see Furlong, 1998; Headrick et al., 2002; May, Workman, & Jones, 2008). On its own, budget gives a sense of agency size, but little else. However, when one examines year-over-year changes for any agency, it tells a different story of the agency’s political environment. One expects that less politicized agencies will have more stable funding streams as the budget will be used less frequently as a means of control through financial rewards or punishments. Less politicized agencies will also have more stable budgets because they are less likely to be implementing new statutes or large new programs requiring bursts of funding. Agency budgetary stability acts as an indicator of a less politicized environment for employees. As stability reduces stress, employees may be less likely to express turnover intent in organizations with steady funding streams.
Although the budget serves as a measure of congressional support, the ongoing process with continuing resolutions and midyear changes creates a level of uncertainty for actual budget results. A better measure of political support of agencies in the eyes of careerists may be changes in staffing levels. Therefore, this study includes a measure of the year-over-year percent change in the number of employees. Employees will probably be more aware of personnel changes such as hiring freezes or large numbers of new incoming employees. Employee change serves as a measure of presidential attention because employee levels are primarily set by the executive branch with rare interventions by Congress. As with budget stability, fluctuations in staffing levels serve as a point of stress where decreased staffing levels means work is distributed to existing staff and increased staffing levels requires training, rearranging responsibilities, and social adjustments to work teams.
Agency Age
Many have proposed that agencies have a life cycle that directly affects the operations of an agency (Bernstein, 1955; Downs, 1966; Lowi, 1979; Moe, 1989; Olson, 1982). Young organizations are thought to be less stable as they are building their structures and programs while responding to new and emerging demands. As they move to being “middle-aged,” the stability increases with established cultures and standard operating procedures. Finally, an older organization becomes the home of conserving established interests and cumbersome bureaucratic processes. For example, Downs (1966) sees young organizations as fighting to establish external legitimacy and stable resource streams. They then move to a stage of rapid growth where they value innovation and expansion. As organizations get older, they slow their growth favoring procedures and stability. Minztberg (1979) made similar propositions that agencies have different stages that alter their internal management systems.
Although agencies may go through these stages, embedded in agency age is the political and historical context of the initial agency design. The decisions made and circumstances at agency conception become engrained in agency culture and persist over time (Niklasson & Pierre, 2012; Verhoest, Verschurere, Rubecksen, & MacCarthaigh, 2010). These insights into the importance of agency age/stage can be at odds with the political issues that arise at any given time. Although many factors may move an issue or program to the top of the agenda of the White House and Congress, younger agencies may absorb this outside attention differently than older agencies where operations are more set. Where a young organization may respond with shifts in structure and staffing to accommodate new demands from the political environment, an older organization is likely to have an established structure designed to address new issues while protecting its core functions (Minztberg, 1979). In this way, older organizations may better insulate their employees from the demands of changes in political attention. These organizational behaviors alter agency culture, thus affecting employees and their career choices.
Although this proposition is not new, it is rarely empirically tested. For this study, agency age is captured in three categories that mark distinct phases in the growth of the federal government. The youngest organizations are those that were founded after President Kennedy took office in 1961 through 2015. This captures agencies created in response to the Civil Rights Movement through today. Younger organizations include some that are considered the most politicized or controversial including the Department of Education and the Environmental Protection Agency. They also include large agencies that are the result of reorganizations including the Department of Homeland Security and the Department of Health and Human Services. The “middle-aged” organizations are those founded between 1933 and 1960. Middle-aged organizations capture agencies developed out of New Deal Program responses to the Great Depression and a wave of new agencies that emerged at the end of World War II. These organizations include the Social Security Administration and the National Aeronautics and Space Administration. The oldest group of organizations are those that existed before the 1933 such as the Department of Justice, the Department of Agriculture, and the Treasury Department.
Agency Design and Cabinet-Level Placement
Moe (1989) states, “American public bureaucracy is not designed to be effective. The bureaucracy rises out of politics, and its design reflects the interests, strategies, and compromises of those who exercise political power” (p. 267). The nature of our separation of powers system is evident in agency design, as the conflict between Congress and the president shows up in the battle over structures. Lewis (2003) discusses how the structural choices made while establishing an agency dictate who has control of the agency—the White House or Congress—and how much control they have. In his words, and building on the classic work of Harold Seidman (1970), “Agency design is fundamentally and inescapably political” (Lewis, 2003), and has substantive impacts on policy (Canes-Wrone, 2009; Lewis, 2003).
Presidents view agency design in terms of how to increase executive control, whereas members of Congress are looking to respond to the short-term desires of their constituents (Lewis, 2003). All politicians wish to see their policy goals and accomplishments survive political fluctuations. When designing agencies, Congress often seeks to insulate them from political shifts by creating administrative procedures, statutory approaches including extended reauthorization bills and mandatory spending accounts, and structural devises such as creating a commission with limited numbers of political appointees. Presidents seek control by placing agencies within the traditional executive-branch department structure, thereby allowing for political appointments and information to move through a clear hierarchy. Presidents oppose congressional efforts to insulate agencies from presidential influence (Lewis, 2003).
In considering agency design, politicians thus consider how to minimize the damage that an agency can do in implementing statutes from their original goals when it falls into the hands of the opposite party. When congressional and presidential policy preferences match, agencies are designed to allow greater amounts of executive influence. Similarly, when Congress is substantially divided and cannot muster cohesive opposition to the president, resulting agencies have strong executive influence (Lewis, 2003).
Although this study does not incorporate all elements of agency design, it incorporates presidential cabinet level as a variable that captures political standing. As the leaders of cabinet-level agencies have more direct contact with the White House and often address the most critical and salient political issues, it is important to consider their political standing in the study of the impacts of the political environment on employee career change decisions. Their position as central to the political dialog is mirrored in their structure. Recognizing their importance, Congress also takes an active role in vetting appointments to these agencies and conducting oversight. These agencies are regularly asked to report directly to the president and congressional committees. As such, it is likely that their efforts will become central to conflicts between the White House and Capitol Hill. Their structures also differ from independent agencies that often have protections that insulate them from politicization such as term appointment and reporting structures that go directly to Congress instead of through the executive branch.
Ideology
Agencies are thought to have an ideological leaning based on the nature of their work and self-selection of employees to work on agency programs and policies (Clinton, Bertelli, Grose, Lewis, & Nixon, 2012; Clinton & Lewis, 2007; Edwards, 2001; Richardson, Clinton, & Lewis, 2018). A question in literatures on bureaucratic responsiveness is whether agencies act based on the will of political figures, the will of the people, or agency preferences (see Carpenter, 2001; McCubbins, Noll, & Weingast, 1987; Whitford, 2005). Although ideological matching of agencies and administrations is widely thought of in terms of policy implementation and performance, it is also relevant to the management of organizations by appointees. Where the ideology of the administration is at odds with the general leanings of an agency, there may be tensions between appointees and careerists as presidents and appointees fear that career bureaucrats will either actively oppose or be indifferent to the presidential agenda (Durant & Resh, 2011). Although presidents use appointees to alter the course of an agency (Lewis, 2008; Maranto & Hult, 2004), there are cases where an appointee is placed in an organization to undo or minimize programs instead of advancing the mission of the agency (Auer, 2008).
In addition to the frictions created for careerists when there are ideological differences between agencies and administrations, there are certain circumstances that punctuate these differences. Michaels (1995) found that appointee–careerist relationships experience stress when there is a change in party control of the White House as new appointees view careerists as loyal to the previous administration. However, impacts of ideological divides may be overstated. Golden (1992) found that ideology was less of a question of partisan value matching, but rather that bureaucrats expressed that it was extreme ideology in appointees that affected appointee–careerist relationships.
When an agency’s work is ideologically supported by an administration, employees may feel motivated by the prospect of progress toward achieving the agency’s mission. This is consistent with Rainey and Steinbauer (1999) and other scholarship that finds that public sector employees are motivated by their connection to agency missions. In contrast, agencies whose work is not favored by an administration are likely to face either stagnation or active efforts to dismantle programs and initiatives that the administration dislikes. Given that in 2015, the president was a Democrat and Congress was controlled by the Republican Party, ideology can be viewed as congruence with either the White House or the Congress making it important to view ideology in conjunction with other variables that explore executive or congressional attention.
Politicization of Agencies Through Appointees
Presidents use the selection and placement of appointees as a powerful means of exerting control over agencies (Lewis, 2003). The choice of appointees can serve as a signal to agencies about presidential priorities and preferences (Wood & Waterman, 1994). Although this is effective in cases where signals are clear, the signals are often more ambiguous or conflicting (Aberbach & Rockman, 2009; for an example, see Durant, 2009). A potential danger of using appointees to politicize the bureaucracy is that it can undermine agency performance (Lewis, 2008; but see Durant & Warber, 2001, for a different perspective and Durant, 2010, for a critique of the Program Assessment Rating Tool (PART) score evidence Lewis marshals). Lewis (2008) points out that appointees often are poorly prepared for their jobs and stay for short tenures, making teamwork challenging. Appointees avoid long-term planning and career staff respond slowly after watching a series of appointees leave as quickly as they arrive (Ali, 2019; Lewis, 2008). Lewis (2008) finds that highly politicized agencies do not perform as well as agencies with less appointee influence. He points out that this creates a challenge in a democratic system where some sheltering from the political system is necessary for bureaucratic agencies to run efficiently by supporting institutional knowledge and allowing for long-term planning, but that responsiveness to political officials is also necessary (Lewis, 2008). This creates two types of effects of politicizing the bureaucracy: first, by influencing the type of people selected for appointments, and second, by discreetly affecting morale, tenure, and incentives of career managers (Lewis 2008, 2009).
The traditional measure of politicization of agencies through appointees looks at the ratio of appointed positions to career positions, but it does not differentiate among types of appointments. Given that the overall ratio of appointees to careerists is always low, comparing agencies only on these ratios does not give a complete picture of how presidents strategically use appointments. Recent scholarship offers an alternative measure of agency politicization through appointments is to examine whether the appointments are held by career members of the Senior Executive Service (SES) who are often in their positions because of managerial and technical expertise versus political appointments (Dull & Roberts, 2009; Lewis, 2008; Resh, 2015). Where more positions are held by SES, the agency is less politicized than those dominated by political appointments.
This alters the political makeup of those in leadership positions, and could then filter into the experience of careerists affecting their satisfaction and career change decisions. Political appointees are more likely to have shorter tenures causing frequent changes in leadership that can be disruptive to staff and performance. The lack of expertise by political appointees compared with SES members may also cause friction for employees who are motivated to fulfill their agency’s mission rather than supporting policies based on political calculations.
Future Job Preference
The traditional approach of viewing turnover only as a question of leaving the organization altogether ignores that large organizations offer movement within. Taking a job at a different agency within the federal government is an attractive alternative for employees who want to change jobs without the risk of entirely leaving government. These may be lateral transfers or advancements, but all involve changing agencies without losing the security and benefits associated with longer tenure in the federal government.
A handful of studies differentiate between types of turnover. They consider whether the same factors drive moving within government as leaving government or retiring. These studies find that there are different factors that contribute to each turnover choice. Ali (2017) used this approach to test whether turnover intent varied by agency as well as measures of satisfaction and motivation. Likewise, Kim and Fernandez (2017) explore whether employee empowerment is correlated with certain types of turnover, and Whitford and Lee (2014) see that employee loyalty and ability to voice their views influence preferred types of turnover differently. Ertas (2015) focuses on demographic factors, and finds differences in the drivers of future job preference in a study of whether millennials are more likely to want to change jobs compared with older employees. Wynen, Op de Beeck, and Hondeghem (2013) also find that intent to leave versus moving within government is driven by distinct factors.
Although intragovernment movement is attractive to employees, organizations face challenges when an employee moves within government as they risk losing institutional knowledge from the departing employee and need to hire and train a new employee. Inclusion of movement within government may give a more accurate picture of staffing challenges facing organizations and yield more useful retention strategies. Where agencies differ in political climate and organizational culture, employee movement within government becomes an important issue if certain agencies consistently attract and retain the strongest employees whereas others suffer from poor retention and less motivated workforces.
Summary of Hypotheses and Literatures.
New variable/concept created for this research.
Method
Data and Variables
Data for this study were compiled from a variety of sources including the 2015 FEVS and FedScope data from the U.S. Office of Personnel Management, the 2012 Plum Book, and agency budgets. Characteristics of each agency were gathered, and then agency data were assigned to individuals in the FEVS data. This allows for testing the impacts of both individual and organizational factors on turnover intent.
Individual-Level Variables
The 2015 FEVS is used for the dependent variables and individual-level variables including pay satisfaction, sex, age, tenure, and whether one is a minority. The survey was conducted electronically by the U.S. Office of Personnel Management over a 6-week period beginning in April 2015. It included employees from 82 federal agencies encompassing 97% of the federal workforce. The overall response rate was 46.8% yielding a final sample of more than 400,000 respondents. The FEVS survey asks 84 questions about people’s jobs and agencies, and demographic questions. As with all voluntary surveys, FEVS is limited by potential self-selection bias. For a full discussion of the limitations of FEVS as a research tool, see Fernandez, Resh, Moldogaziev, and Oberfield (2015).
Turnover intent
The dependent variable is constructed from a question in the FEVS that asks, “Are you considering leaving your organization in the next year, and if so why?” The possible responses include “No”; “Yes, to take another job inside the federal government”; “Yes, to take another job outside the federal government”; and “Yes, other.” Among respondents, 66.49% of people did not want to leave their current organization, 18.54% wanted to move within government, 4.00% wanted to leave government, and 10.97% wanted to leave for other futures including retiring, taking care of family, or returning to school.
Control variables
Five control variables on the individual level are also taken from the FEVS. These are included to capture frequently used determinants and controls used in turnover studies, and include pay satisfaction, tenure, age, sex, and minority status. Pay satisfaction is measured on a 1 to 5 Likert-type scale ranging from very dissatisfied to very satisfied. Tenure is measured in three categories: 0 to 5 years, 6 to 14 years, and 15 or more years. Age is also a categorical variable including those below 40 years old, 40 to 49 years, 50 to 59 years, and those who are 60 years and older. Each of these categorical variables was broken into multiple dummy variables. Both sex and minority status are dummy variables (see Table 2).
Descriptive Statistics for Individual-Level Variables.
Agency-Level Variables
Budget and staffing volatility
The next two variables test the turbulence concept presented here with measures of budgetary and staffing volatility. For each agency, the year-over-year change is calculated. These changes are then used to determine the standard deviation for annual change by agency. The mean is calculated for individual agency percent change. The yearly percent changes for that agency and the agency mean are then used to calculate the standard deviation for that agency. The same process is repeated for each agency. The years included are 2010 through 2014 for each agency standard deviation calculation. The calculations were also generated using the data from 2010 through 2015 to test whether including the current year altered findings. As the findings were all consistent, 2010 through 2014 was used to allow for any time lags between changes and employee reactions. Stability in the budget and staffing are the equivalent of predictability of resources for federal agencies. For these variables, small standard deviations indicate predictable, stable levels over time. However, large standard deviations indicate more dramatic movements meaning that the organization must adjust to changing levels.
The budget data were constructed by gathering actual budget data for previous years from each agency’s annual budget request submission to Congress. In the president’s annual budget request, agencies provide historical data about spending that include the requested budget for the next year, the approved budget for the current year, and the actual budget for 2 years prior. For example, the 2017 budget request is used to gather the 2015 actual budget data. For the full-time employee equivalents (FTE) volatility measure, employment data were collected from the Office of Personnel Management’s FedScope data.
Agency age
The founding year for each agency was used to divide agencies into the three agency age categories: those founded before the Great Depression, those founded from the Great Depression to the beginning of the Kennedy administration, and those founded after the start of the Kennedy administration. These dates were chosen both for their historical significance in the growth of the U.S. administrative state. An analysis of the data shows that these are the three points where the data naturally cluster.
Cabinet level
To isolate the impact of agency political positioning, a dummy variable identifies whether an agency is cabinet level or not.
Ideology
To measure ideology, this study uses Richardson et al.’s (2018) measure constructed using data from the 2014 Survey on the Future of Government Service that captured the perceptions of political appointees and career member of the SES. Responses were used to construct ideology estimates ranging from −2 for the most liberal agency to +2 for the most conservative.
Political appointees
As presidents use appointees to influence policy either through partisan appointees or technical experts, the measure of appointees is the percent of appointees who are political versus career SES members. These data were collected from the 2012 Plum Book. Agencies where all appointees are political would be 100%, whereas those with only SES appointees would be 0%.
Control variable, agency staff size
As larger agencies may offer employees more options for internal transfers, agency staffing levels are used as a control. Although the data initially count staff size by individuals, the measure was collapsed into units of 10,000 for purposes of interpretation. At the individual count level, measures were all statistically significant, but had coefficients of zero. These data were collected from the Office of Personnel Management’s FedScope data (Tables 3 and 4).
Descriptive Statistics for Agency-Level Variables.
Descriptive Statistics for Agency-Level Variables.
Analysis
Four binary logit models were used to assess the political environment’s impact on turnover intent. In the first model, the dependent variable is whether people want to change jobs regardless of future job preference. This captures the first stage of decision making where individuals make the initial choice to leave their organization. The next three binary logit models test the impact of the political environment on individual expressions of turnover intent. These models differentiate among future job preferences including people who wish to stay in their current jobs as the base group, those who want a new federal job (Model 2), those who wish to leave the federal government for another job (Model 3), and those who want to leave for other futures including retirement, taking care of family, and pursuing different educational or career paths (Model 4). Models 2 through 4 were initially run together as a multinomial logit model, but were separated for simplicity of interpretation using odds ratios. The Hausman test was applied to this model to confirm that there is not an independence of irrelevant alternatives (IIA) violation. The results from the multinomial logit model were the same as for the three binary logit models. When the correlations were examined to test for multicollinearity, all correlations were below the .70 threshold.
The models were also tested by individually adding each independent variable as well as testing other possible measures of political attention including Congressional Record documents and Federal Register documents. These variables were not included because they could represent various forms of politicization and bureaucratic discretion. Although not utilized in this study, an in-depth analysis of these documents offers a rich source for future research related to this subject.
Findings—Testing Political Environments
As asserted by systems scholars, this research finds that an organization’s position in the environment and the stability or volatility of an agency’s resources filter into the organization and influence employees (Burns & Stalker, 1961/1994; Hannan & Freeman, 1977; Wilson, 1989). This study extends the proposition that turbulence, in addition to the position in the political environment, influence employee preferences and choices. The six variables that capture the complex political environment for agencies and their employees are budget volatility, staffing volatility, agency age, cabinet-level reporting structure, agency ideology, and political appointees (Table 5).
Logit Results—Compared Those Who Want to Stay With Those Who Want to Change Jobs.
Reference group is those who are very dissatisfied with their pay.
Reference group is those whose tenure is less than 6 years.
Reference group is those who are younger than 40.
p < .1. **p < .05. ***p < .01.
The two measures of the volatility of agency environments (budget and staffing) yielded interesting findings. Contrary to H1, budget volatility reduced turnover intent for the full sample, transfers, and those who wanted other futures, but increased the probability of people wanting to leave government. This suggests some self-selection where budget volatility encourages some people to stay, but those who are frustrated choose to leave government completely rather than transferring to a less volatile agency. A 1-point increase in the standard deviation of budget changes (volatility) decreased the odds of people expressing turnover intent by 0.5%, of wanting to transfer by 0.9%, of wanting to leave for other futures by 0.3%, but increased the odds of wanting to leave government by 0.6%. Noting that the range for these standard deviations is 0.79 to 45.32, the impact across the full range becomes more meaningful.
H2 stating that increased staffing volatility would correlate with higher turnover intent was supported for all future job preferences except for those who wanted a new job outside of the federal government. The impacts of staffing volatility were slightly stronger than the impacts of budget volatility. With a 1-point increase in the standard deviation (volatility) of staff size changes, the odds of expressing any turnover intent increased by 2.7%, of wanting to transfer by 5.0%, of wanting to leave for other futures 2.0%, but decreased the odds of wanting to leave government by 3.1%. The range for these standard deviations is 0.52 to 16.83.
The proposition in H3 that employees of young agencies are more likely to want to change jobs is confirmed. This supports scholarship that asserts that organizations have a life cycle, where early on there is less stability as the organization solidifies its path (Bernstein, 1955; Downs, 1966; Lowi, 1979; Moe, 1989; Olson, 1982). This finding is consistent across all preferred future job categories. For instance, employees working for an agency founded before 1933 are 25.5% less likely to want to change jobs compared with employees working at an agency founded after 1960. The models were also run with the oldest agencies as the base group, and found that there is little difference in turnover intent between the oldest agencies and those founded between 1933 and 1960.
H4, that people who work for cabinet-level agencies are more likely to want to change jobs, was also confirmed for all types of turnover intent. Overall, the odds of employees who work for cabinet-level agencies expressing turnover intent were 1.411 times higher than their counterparts in agencies that are not cabinet level. The greatest impacts were among those who wished to transfer within government, where the odds were 1.798 times. The lowest difference was for those who wanted to leave for a job outside of the federal government as the odds of expressing turnover intent increased by 1.076.
Agency ideology (H5) was not a significant determinant of turnover intent for any form of turnover intent except for transferring within government. Here, the findings were in the opposite direction from what was expected. People working for more conservative agencies were less likely to want to transfer.
Politicizing agencies by relying on political appointees rather than career SES appointees increased the probability of expressing turnover intent in agencies (H6). Overall, a 1 percentage point increase in appointees who are political correlated with the odds of expressing turnover intent increasing by 0.1%. Although the increments are small, the overall impact is worth noting. Of note, the impact was greatest on the decision to leave government for another job (odds ratio = 1.005), but not significant for the choice to leave for “other futures” including retirement, education, and family needs. This may capture a self-selection process where employees who are most frustrated with the direct politicization implemented through appointees choose to leave government.
Finally, H7, that political environments will have a greater impact on the choice to transfer within government than on the choice to leave government for another job or other futures, is supported for all measures of political environments except for political appointees. In each case, the odds ratios show the largest impacts for the choice to transfer. This is expected as the choice to leave government involves more risk and other factors compared with staying with the federal government as an employer. If one is frustrated with his or her organization, the simplest choice is to transfer within government. (See Table 6 for the Hypotheses Decision Table).
Hypotheses Decision Table.
Note. Opposite Findings were statistically significant in the opposite direction. ✓: reject the null hypothesis; X: fail to reject the null hypothesis.
Discussion
With the exception of ideology, politicization of agencies appears to have an impact on both the decision to move within government and the choice to leave government for either another job or other futures. This study uses six different ways that agencies are influenced by political environments. Each offers insights on components that influence organizations and careerists. Rather than attempting to find one measure of political environments, this study finds that each operates independently. Yet, all measures except for budget volatility and ideology demonstrate that more turbulent political environments increase turnover intent.
Volatility in budgets and staffing levels are the measures of political support that directly reflect agency resources. Volatility in staffing increased turnover intent overall, transferring and other futures, but decreased turnover intent for leaving government. Staffing levels are the clearest signal to employees as they are often asked to absorb the work of unfilled positions and train new employees. Where staffing is stable, baseline workloads are easier to manage. Budget volatility functioned in the opposite way, where volatility decreased turnover intent overall, for transferring and other futures, but increased turnover intent for leaving government. Budget volatility had the weakest impact on turnover intent. This is likely because the current approach to budget via continuing resolution means that staff are perpetually unclear about the actual budget without information about authorizations compared with previous years. The difference in findings for intent to leave government for new jobs versus those who wish to transfer or leave for other futures may capture employee self-selection where those who are most frustrated with the impacts of politicization respond differently than others. For instance, volatile budgets may signify political relevance for most employees, thus decreasing turnover intent, but for those who are most frustrated by resource instability, they prefer to leave government.
Two components of political environments are structural—agency age and cabinet-level status. In both, agency placement in the environment increases politicization and perhaps volatility. The youngest agencies show higher probabilities of turnover intent compared with the oldest agencies. However, there was no substantive difference between middle-aged organizations and the oldest organizations. This offers support for organizational life cycle propositions that differentiate young organizations (Bernstein, 1955; Downs, 1966; Lowi, 1979; Moe, 1989; Olson, 1982). An opportunity for future research is to better understand how agency age translates into agency operations. These findings indicate that people who worked for the youngest agencies were the most likely to express either form of turnover intent. This may reflect Downs’ (1966) “climbers” who are attracted to younger organizations where they can build their careers by making a substantive impact on the direction of a younger organization. People working in cabinet-level agencies are also more likely to want to change jobs. These agencies have more direct contact with both the White House and Congress, including more actively monitoring and policy efforts. They are often addressing the most salient political issues. These agencies also lack some of the insulation provided to smaller, independent agencies.
Favoring political appointees over career SES appointees also creates volatility within organizations. As political appointees are often in their positions for shorter times and are motivated by fulfilling White House initiatives, working under political appointees compared with the technical experts from the SES creates stress for agency employees. These findings support research by Lewis (2008), Dull and Roberts (2009), and Resh (2015) that measuring politicization through appointees is best done by examining the ratio of political appointees to SES appointees. Here, people who work for agencies that have more political appointees are more likely to want to leave their current position. The impacts were greatest on the preference for leaving government for another job, still significant but weaker for transferring within government, and not significant for leaving for other futures.
An important finding is that agency ideology is not a driver of turnover intent. This is consistent with Golden’s (1992) finding that employees respond to extreme ideologies, but are tolerant of a wide range of ideological differences. Although agency ideological differences with administrations may affect agency action, it does not result in differences in turnover intent. Essentially, partisanship is less important to federal employees than stable workplaces.
The impacts of political environments on employees are strongest on those who want to transfer within government. This is an important distinction as hiring preference makes it easier for federal agencies to hire from within, and for employees, transferring to a new agency involves less risk and is often easier than leaving the federal government. The size of the federal government and locations across the country also remove geographic constraints on changing jobs while staying with the federal government. This finding is consistent with previous studies that view moving within government as driven by different factors than leaving government (Ali, 2017; Ertas, 2015; Kim & Fernandez, 2017; Whitford & Lee, 2014; Wynen et al., 2013).
Limitations and Future Research
The results of this research may understate the significance of the political environment on turnover because it involves all levels of federal employment. It is reasonable to assume that the impacts of politicization are punctuated among certain groups and are diluted as people’s jobs move farther from the political center. A study focusing on senior employees and those working in agency headquarters may show greater influence as suggested by Maranto’s qualitative findings in 2005. Data limitations also do not allow for understanding how individual political preferences and ideologies influence turnover decisions. For instance, an extremely conservative employee working for a liberal agency or appointee may choose to leave based on that mismatch rather than agency relationships with the broader political environment. However, the variables needed to explore these subgroups are not available in the public use FEVS data.
Demographic differences are another avenue to consider as it is possible that some populations are more alienated by political environments than others. For instance, newer employees may be more influenced by the political nature of an agency’s work than those who have worked for the federal government for decades. Members of groups that have a history of being disenfranchised from the political system may also respond to politicization differently. The author considered using interaction terms to explore these factors, but concluded that these questions were worthy of a full future study that differentiates turnover intent based on populations.
There are also questions about how people psychologically process the impact of the political environment. One’s response to politicization is tempered by the belief that it is going to be the same if he or she goes somewhere else. It is also possible that those who work for the government consider the political environment a cost of doing business that is outweighed by the advantages of federal employment. This analysis only assesses people’s past and current political environments without measuring their perceptions of the future environment. Although this presents a challenge where ideological matching is concerned, it is less of an issue where the measure is of volatility or turbulence over time.
The impact of the political environment on careerist experiences and job choices is ripe for continued research. Although this study does capture a component of time in the volatility measures, it does not look at how politicization functions over time and across administrations. As there is a natural cycle in presidential administrations, the findings for the first year of an administration may differ from this study that serves as a snapshot of the seventh year of President Obama’s administration. Research on political environments that incorporates time also allows for studying the influence of varying amounts of contentiousness and conflict between Congress and the White House. As a snapshot, this study only looks at the influence of the political environment at one moment. The divisiveness that currently dominates may yield more dramatic findings. Conversely, the impact of political environments may function differently when one party dominates both Congress and the White House.
This study ignores the economic environment as a driver or inhibitor of turnover intent and future job preferences (see Ali, 2017). Although this research asserts that based on the systems scholarship, the stability or volatility of an agency’s political environment is the key measure, there is also room to study whether there is a difference between positive and negative political attention. The questions raised here also beg for application to data on actual turnover rather than turnover intent, and state and local questions as it is in these governments that careerists often have the most direct contact with political actors.
There is also a need for future research on whether the impact of political environments on turnover and turnover intent varies based on organization types such as those in Wilson’s (1989) and Lowi’s (1979) typologies, and agencies of different ages as proposed by systems theorists. As this study does not assess the management styles of agencies, but relies instead on widely recognized propositions in both public and private sector systems and strategic management literature asserting that conditions influence management, in-depth case studies could examine whether the changes in management in response to political environments are consistent with the propositions from these scholars, and then examine individual responses to specific management changes. In addition, there may be some sections of agencies that are inherently more connected to their political environment than others. For instance, the budget office and legislative affairs offices are more likely to be aware of agency politics than field offices.
Conclusion
Contentious political environments for public agencies are increasing in our divisive political climate. Previous norms that sought to buffer careerists from direct political attacks are compromised. Budgets and staffing levels are pawns in relentless partisan battles in the current climate. This study uses data from the Obama administration when there were stark differences in ideologies and beliefs about the role of government. These differences have amplified under the Trump administration. Although the results show the relevance of incorporating political environments in our understanding of what it means to work for the public sector and career change decisions, these results likely underestimate the impacts compared with today, given the direct attacks on federal workers and the longest government shutdown in U.S. history. Without a pendulum swing back to more tempered treatment of agencies and employees by elected officials, politicization of agencies paired with the retirement of baby boomers and an increasingly mobile workforce will force agencies to address challenges in recruitment and retention.
This research seeks to expand the turnover scholarship to incorporate the impacts of the political environment. Taking a systems perspective on exploring the political environment also requires a holistic approach to the political world instead of focusing on one component such as appointee–careerist relations or congressional controls. In this approach, characteristics such as the stability of a system affect organizational form and structure (Burns & Stalker, 1961/1994; Hannan & Freeman, 1977), which, in turn, affect turnover intent (Cameron et al., 1987; Harrison et al., 1988; Harrison, 2004). When looking at an agency as part of an open system, one must explore the relationships that define a federal agency’s environment. This includes exploring scholarship on agency design and the administrative presidency, ideology, bureaucratic responsiveness, and congressional relationships.
Based on these lines of scholarship, measures of the volatility of agency’s political environments through budget and staffing resources were developed to test whether they influence careerist job change decisions. The findings show that the political environment influences all turnover decisions with the greatest impacts on the choice to move within government. Finally, this research supports the concept of measuring the political environment as a function of stability or volatility as well as agency age, agency position in the president’s cabinet, and political appointees. What was not relevant was ideological matching.
For all significant measures of the political environment, agencies that are more politicized experienced higher odds of people wanting to change jobs. The exception was that agencies with greater budget volatility had lower turnover intent. It is important to consider multiple forms of politicization as some are time invariant or structural, whereas others are more likely to change from one administration to the next. For instance, agency age is a set variable except in cases of wide-scale reorganizations such as the formation of the Department of Homeland Security. Similarly, agency placement on the president’s cabinet is rarely changed with only a handful of changes over U.S. history. However, the use of political appointees versus SES member varies from administration to administration as well as the volatility of budgets and staffing. Understanding how the volatility of an agency’s environment influences agency and employee behavior is important in a time when there are wide differences in policy priorities. As we saw in the transition from the Obama administration to the Trump administration, agencies at the center of the political storm change. For instance, where the Department of Justice was not previously considered highly political, it is now receiving new levels of political attention to policies, procedures, and employee behaviors.
Although budget volatility worked in the opposite direction as staffing volatility, it suggests that some forms of volatility may serve as motivators for employees. Political salience as displayed by changing budgets may attract people to stay in public service. Nonetheless, those same changing budgets drive a segment of the workforce to choose to leave as budget volatility increased the odds of people wanting to leave government for a new job, but decreased the odds for wanting to transfer and other futures.
The single measure of political environments that was not significant was the ideological leaning of agencies. This finding suggests that it is not partisan politics that influences careerists, but rather general political attention. This may be explained by Maranto and Hult’s (2004) assertion that careerist compliance with policy changes is the result of those changes falling within a bureaucratic “zone of indifference.” It is also consistent with Golden’s (1992) finding that careerists were more concerned with extreme ideology in appointees rather than partisan preferences. Essentially, as long as agencies are not asked to move too far right or left, careerists accept that partisan differences are natural to government work.
Given that the impacts of political environments on turnover intent are often indirect as they filter through organization structures and behaviors, it is not surprising that the impacts of political environments are smaller than traditional intraorganizational and demographic factors. Although these are still the primary drivers of career change decisions, there is room for considering how the political environment alters the professional lives of public sector employees. As we continue in politically contentious times, where there is a direct assault on the ethics and competency of public employees, and agency activities that were previously insulated from politicization move to the front of the political discourse, it becomes critical for public managers to recognize the implications for recruitment and retention.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
