Turnover at the Top: Causes and Consequences of Leadership Change in Public Agencies
This essay will define leadership turnover as change and replacement of the occupants of formal, non-elected executive and managerial positions within public agencies.
Since at least the time of the Volker Report (The National Commission on Public Service 1989), the phenomenon of personnel turnover in the federal civil service, particularly at or near the top of agencies, has been regarded as a problem. Comparable issues of turnover in state and local government agencies have also been recognized as a matter of some concern, with articles by scholars and practitioners and speeches by politicians and interest group leaders expressing concern over this trend. Whether we see this as a “quiet crisis” (see Lewis 1991; Light 2008) or simply a natural, manageable problem, the issue deserves careful examination. This essay surveys the scholarly literature that examines the causes and consequences of leadership turnover within public agencies. The focus of the essay is not turnover among elected leaders, or of rank-and-file personnel, although some of the research on those topics may be instructive. The focus is upon turnover in occupants of formal executive and managerial leadership positions. The research in the business literature on turnover among CEOs and boards of directors may also be of interest but is not the essential issue to be studied.
The essay proceeds as follows. The first section very briefly summarizes the research on the causes of turnover. Of necessity, this section relies to a great degree on literature covering rank-and-file employee turnover. The next section discusses research on the implications and consequences of executive turnover. Finally, the concluding section discusses how and why the meaning and significance of turnover is context specific and dependent upon opportunity structures of not only agencies but of the entire labor market.
Causes of Turnover
In public organizations and in many private ones, leaders can be evaluated based upon the performance of the programs that they lead. If their programs appear to be successful (e.g., if a coach’s team wins ball games), leaders are retained and even rewarded. In the political science literature, an argument about the way that voters make decisions about candidates (particularly presidential candidates) has been based upon this tendency. The political science scholarship describes this phenomenon as “retrospective voting” (Fiorina 1981). According to this argument, voters evaluate the incumbent based upon the performance of the government that he/she leads. In most of the research, the incumbent who is examined is the incumbent president, and the major focus of the voters’ evaluation is the state of the economy. There is considerable evidence that these retrospective models do explain quite a bit of the variance in citizen evaluations of the president and the vote share obtained by incumbent presidents and the presidential candidates of the incumbent’s party. Nonetheless, the near-exclusive focus on the state of the economy offers a somewhat incomplete view of the concerns of all voters, since most voters care about a wide variety of other issues (see also Berry and Howell 2007).
Public agency leaders may not be held accountable by the public in general (although there are many executive agencies at the state level that are chosen through election). Nonetheless, the biggest problem in evaluating many agency leaders is the multiplicity of the tasks that they perform and the difficulties in measuring and monitoring their performance. Most of the evaluation of these leaders is done by elected officials and administrative superiors in their organizational hierarchy. In practice, though, much of the research literature, particularly in public administration and human resource management, does not view turnover as primarily a function of unfavorable evaluations of performance, but as a result of unfavorable conditions within the workplace. In other words, much of the research explains turnover as a choice made by dissatisfied employees. This tendency may be attributable to the prevalence of civil service rules that make dismissal difficult. However, agency leadership would not normally be covered by those rules.
One of the notable characteristics of much, although not all, of the literature is the usually unstated premise that turnover is something to be avoided (see, e.g., Lewis 1991; Light 2008). For that reason, the research seems dedicated to the quest of identifying variables that tend to reduce the frequency with which public employees or officers leave office. Much of the research is quantitative, with regression-based analyses that predict turnover or turnover intention among public employees for a given set of agencies within a particular period of time. In some cases, the research does not aim to account for individual turnover or turnover intentions but instead attempts to explain a rate of turnover among personnel within an organization, management team, or program leadership structure. To a great extent, with some notable exceptions, there is not much elaborate and elegant theorizing in this research. For the most part, the literature tests the importance of a number of favorable inputs (i.e., nice stuff to have) that might discourage personnel from leaving their agencies. In general, better pay and benefits (Lee and Whitford 2007), friendlier co-workers, on-site child care (Selden and Moynihan 2000), and better relationships with superiors all tend to discourage turnover (Whitaker and DeHoog 1991). Sometimes a measure of fiscal stress in a community is a strong predictor of turnover (e.g., Tekniepe and Stream 2012). Some behavioral indices of attitudes, such as organizational commitment, organization-person fit, and public service motivation, as well as aggregated indicators of supervisor practices, are also negatively associated with turnover, although somewhat surprisingly opportunities for “voice” are not significantly related to turnover (Moynihan and Pandey 2008; Selden and Sowa 2015). Some research indicates that turnover diminishes as employees are supervised by someone of their own race or gender (Grissom et al. 2012). It does appear that turnover may be greater in rural locations than in urban or suburban locations, perhaps because the rural areas are not as attractive for some households (Grissom and Andersen 2012). However, Watson and Hassett (2003) find that some of the longest serving city managers are located in small, homogenous towns. It also seems to be the case that public university presidents have much shorter tenures than private university presidents (Monks 2012). Agency leaders who came up through the ranks appear to stay longer in their positions than leaders who were recruited from outside the organization (Grissom and Andersen 2012). Ironically, managers with higher education may have shorter tenures than those with less (Mani 2014).
There are some troubling aspects of this literature. In addition to the generally low quality of theorizing, there is only occasionally a distinction made between voluntary and involuntary turnover. It also is not clear whether turnover reflects a departure from the agency or simply a move to another unit within an agency. Also, there tends to be a focus on variables that pertain to agency or supervisor characteristics. The environment outside the organization in most instances does not get as much attention. It is also the case that institutional arrangements regarding particular positions (e.g., employment contract length, vesting rules within pension systems, post-termination restrictions on employment) are often unexamined, perhaps because the large-N survey-based studies do not include items regarding those issues on their questionnaires.
Consequences of Turnover
Just as much of the literature seems implicitly aimed at reducing turnover, much of the research on the consequences of turnover presumes that the results are bad. Turnover can reduce, at least temporarily, the workforce available to do agency tasks. It can involve a loss of human capital, agency expertise, and institutional memory (Duffield et al. 2011). It may affect the time horizons and the capacity of the agency to make credible commitments to fulfill long-term commitments, such as municipal bond issues and service contracts of long duration (Clinger et al. 2008). However, the research also finds that the results of turnover may be contingent upon what kind of turnover is involved and how the agency was performing prior to turnover. Ironically, involuntary turnover may be associated with improved organizational performance, while voluntary turnover appears linked to performance declines (Boyne et al. 2011). Furthermore, turnover in general seems to lead to improved performance in agencies that had been performing poorly, while turnover seems to be followed by performance declines when an agency had been performing well. In some cases, moderate turnover is associated with performance improvements, but very high turnover is associated with declining performance (Meier and Hicklin 2008). Whether these relationships are actually causal or just a statistical artifact, a “regression toward the mean” is not clear. But it does seem likely that a badly performing organization is not likely to improve unless some changes are made.
Recommendations Regarding Understanding Turnover
Understanding turnover in personnel at the top, middle, and bottom of an organization should be enhanced by the accountability movement that is manifested at all levels of government now in the form of organizational report cards, Program Assessment Rating Tool (PART) evaluations, performance measurement, etc. Perhaps most obviously, in the realm of education, school leaders will be scrutinized on the basis of how their students perform. That does not mean that dismissals will take place immediately after a poor showing on standardized test. It does mean that elected officials, citizens, and the media will begin to ask some questions. Resistance to this scrutiny could lead to “voluntary” turnover, even if involuntary dismissals do not take place.
One aspect of the research that is mostly lacking is a recognition that agency leaders are often participants in a national or at least regional market. Some scholars, such as Grissom et al. (2016), have emphasized this aspect of the employment of agency leaders, but many other researchers ignore it. When viewed in this way, turnover is a phenomenon that should increase as job opportunities for advancement exist in other agencies, often in other states and sometimes in other levels of government. Such opportunities may also exist with nonprofit organizations, think tanks, and for-profit firms. Research that only examines workplace variables or the behavior of immediate supervisors will miss the significance of the broader labor market. When examining these market effects, researchers should take care to take into account the agency-specific or position-specific human capital that agency leaders bring to a job (or take away from a job) should they leave their present positions. Opportunities outside current agencies are going to be much more numerous and lucrative for certain individuals than for others. Of course, some of the human capital that agency leaders have is not just technical. Some may have more general administrative capability or personal political connections than other agency leaders. Whether “who you know or what you know” is more important is a promising question for research to answer.
More research should also be focused on formal policies that agencies use to discourage turnover or at least shape behavior after leaders leave office. For example, many agencies have policies regarding post-termination employment. Specifically, regulatory agencies may forbid their employees form going to work for the firms that they formerly regulated. This presumably promotes a more ethical stance by the regulators and prevents conflict of interest. Perhaps it was assumed that with more restricted employment opportunities available after working for the agency, staffers and commissioners would remain with the agency longer. Ironically, research by Law and Long (2012) indicates that these post-employment restrictions by state utility commissions lead to the recruitment of employees with weaker credentials who subsequently serve shorter tenures with the government. The implications of restrictive employment rules of this kind deserve to be explored.
Currently, little attention has been paid to the question of whether turnover has different implications in different task environments or in positions with different professional specialties. Monks (2012) does find that university presidents with degrees in social science, business, medicine, or law are more likely to leave office for “nonpresidential” positions than are presidents with other backgrounds. It is not clear if this is because of greater job opportunities for those with these specializations or if these kinds of presidents are not as eager to pursue a presidential career. It is also not clear whether turnover in a more technical as opposed to a managerial position has different kinds of repercussions for agency performance. One might presume that turnover of leaders with highly specialized, highly scarce, and highly valued skills would have more detrimental consequences for the organization than would turnover among individuals whose skill set was not as unique or in such high demand.
Finally, it would be interesting to learn whether turnover in positions for which there was substantial succession planning was as significant as it might be in organizations that had done nothing formally to respond to the departure of agency leaders. Unfortunately, there is a substantial number of articles written on succession planning, but most are very anecdotal and prescriptive (see, e.g., Pynes 2004; Rothwell and Poduch 2004; Gothard and Austin 2013). Very little empirical analysis of the consequences of succession planning for leadership turnover exists.
Leadership turnover in public agencies has various causes and consequences. Some of the causes of leadership turnover have their origins in the attributes of individual leaders. Other causes exist at the level of the organization or the work unit within the organization where leaders are located. Finally, some turnover is caused by factors at the societal level, particularly involving market opportunities or opportunities for advancement in other agencies. Unfortunately, much of the research has focused upon factors at only one level rather than looking at the issue more holistically. To the extent that public administrators perceive turnover to be a problem, they must consider all the factors that are pertinent. Some factors may not be subject to control. They are embedded in a political and economic context that public administrators can only modestly influence.
Turnover among agency leaders may not be something that, from a normative standpoint, should be reduced. Whether or not turnover is a matter of concern should depend upon the consequences that it will have for public service. In general, the research does not find that leadership turnover necessarily results in lower performance in organizations. Often turnover in poor-performing organizations is followed by better subsequent performance. Turnover in high-performing organizations, not surprisingly, is often followed by lower performance. Leadership turnover may in some cases lead to decisions that reflect the different preferences of the new occupants of those offices. Consequences of turnover may also reflect different levels of competence between the new and old leaders. In short, turnover’s impacts will depend upon who has been replaced and who has been chosen as the replacements.
No matter what the quality of the new occupants of leadership positions, it is likely that extremely high levels of turnover will cause certain kinds of difficulties for public agencies. Uncertainty in organizational functioning may be created, and long-term commitments with employees, contractors, and political superiors may result. Such uncertainty may be more problematic for certain functions and for certain agencies than for others. For example, high turnover might be unacceptable for agencies that are concerned with long-term commitments, such as pension fund management, capital project programming and finance, and comprehensive planning. Turnover might be much less of a problem in functions where the environment is more fluid, and responsiveness to market or political conditions is valued. For example, leadership turnover among individuals involved in economic development, human services administration, and emergency services may be an advantage. In these areas, leaders must respond to changes in policy or circumstances that are outside the control of their agencies. This responsiveness may be easier to achieve with new leaders than with those accustomed to the status quo. In short, the consequences of leadership turnover, both empirically and normatively, vary according to the circumstances in which the turnover occurs.
- Clinger JC, Feiock RC, McCabe BC, Hyung-Jun P (2008) Turnover, transaction costs, and time horizons: an examination of municipal debt financing. Am Rev Public Adm 38(2): 167–179Google Scholar
- Gothard S, Austin MJ (2013) Leadership succession planning: implications for nonprofit human service organizations. Adm Soc Work 37(3):272–285Google Scholar
- Mani BG (2014) Determinants of a city manager’s tenure in office: the person, job, municipality, and election system. Sage Open, 4(1):1–16Google Scholar
- Rothwell WJ, Poduch S (2004) Introducing technical (not managerial) succession planning. Public Pers Manage 33(4):405–419Google Scholar
- Selden SC, Moynihan DP (2000) A model of voluntary turnover in state government. Rev Public Pers Adm 20(2):63–74Google Scholar
- The National Commission on the Public Service (1989) Leadership for America: rebuilding the public service. GPO, Washington, DCGoogle Scholar