f. Division of Research Graduate School of Business Administration The University of Michigan September 1982 CHANGE ASSUMPTIONS OF THE CIVIL SERVICE REFORM ACT Working Paper No. 313 Rodney L. Lowman Donald F. Parker The University of Michigan FOR DISCUSSION PURPOSES ONLY None of this material is to be reproduced without the express of the Division of Research. quoted or permission Paper presented at the Academy of Management Annual Convention, New York August, 1982

I

Change Assumptions of the Civil Service Reform Act* The Civil Service Reform Act (CSRA) was passed in 1978 largely as the result of intense efforts by the Carter administration. It was based on the major premise that, although the federal workforce was "basically honest and competent," it had become "...a bureaucratic maze which neglects merit, tolerates poor performance, permits abuse of legitimate employee rights, and mires every personnel action in red tape, delay, and confusion" (Carter, 1976). This paper will examine the factual and fictional aspects of these and other assumptions made by the CSRA. These assumptions can be placed into several major categories, including: (1) assumptions, apparently unexamined, about the baseline conditions of the federal sector; (2) beliefs about how change most effectively takes place in organizations and social systems; and (3) the curious assumption of a rational and (relatively) apolitical environment in which to create change. Each of these areas is examined in this paper and, where appropriate, data are presented from our large sample of federal workers. Briefly, a team of researchers from the Institute for Social Research was one of three university-based groups contracted to evaluate the organizational effects of the CSRA. Each research team was to study five organizations, to be followed over a period of five years. Four different government agencies, which varied considerably in mission, technology, and size, were represented in The University of Michigan research sites. While we can obviously make no claim that these samples were representative of the entire federal workforce, there was no reason to believe that they were atypical. Moreover, the *Paper presented at the Academy of Management Convention, New York, August 1982.

2 similarity of responses across these geographically and technologically divergent organizations lends credence to the idea that our samples were reasonably characteristic of federal employees. Assumptions about Baseline Conditions Although Carter's CSRA was said to have been based on seven months of work by a task force of more than 100 career civil servants who had carefully explored the strengths and weaknesses of the existing Civil Service (Thayer, 1978), it is not evident that many of the basic objectives of the CSRA were empirically well-grounded. In this section of the paper, we examine seven fundamental goals of the CSRA using our baseline data to check the validity of the assumed conditions explicit and implicit in the act. The data base used in these analyses consists of responses to a questionnaire completed by more than 2,700 federal employees at all organizational levels collected after the passage of the CSRA but before implementation of most of its provisions. The data reported here are based on scales developed from extensive factor analyses, and on single items (Organizational Assessments of the Effects of Civil Service Reform, 1980). A. Difficulty in Hiring Competent Personnel President Carter's statement to Congress on the need for the CSRA stated that the federal bureaucracy entangles every personnel action in "...red tape, delay, and confusion" (Carter, 1978, p. 444).* Table 1 displays the responses of supervisors in the agencies to six questions relevant to Carter's assumption. INSERT TABLE 1 ABOUT HERE *All subsequent quotations are from Carter, 1978.

3 From these data, it is clear that most supervisors in our samples agreed in 1980 that it took too long to process the paperwork required to hire new employees. Although the mean values for the remaining items are nearer to neutral, most respondents tended toward agreement that the personnel system is problematic. A site-by-site analysis of these data, however, showed differences among the various agencies, with some demonstrating slight disagreement with these statements. Thus, the data seem to lend limited, but by no means unequivocal, support to the first assumption. Carter's rather categorical statement that this was a "serious defect" of the system is not well supported by our samples. It would be closer to the truth to say that supervisors (at least in our samples) perceived the so-called bureaucratization associated with hiring people as a real, but rather moderate, irritant. B. Employee Motivation By implication, the President and his advisors suggested that many federal workers were either incompetent or unmotivated. He said, for example, that "the public suspects that there are too many government workers, that are underworked, overpaid, and insulated from the consequences of incompetence" (p. 444). Table 2 displays the mean responses to two scales and five single items INSERT TABLE 2 ABOUT HERE concerning employees' perceptions of their work and organizations. Contrary to Mr. Carter's assumption, these data suggest that most employees report relatively high motivation and a rather positive attitude toward performing well on their jobs. They also believe that they have the requisite skills to perform their jobs well. Such data must be cautiously interpreted since the measures of motivation are based on self-report, but they have been well

4 validated in prior studies of thousands of predominantly private-sector emloyees. The federal employees do not appear to be responding in a substantially different way. Moreover, and again contrary to prior assumption, these workers report themselves to be at least as motivated, and perhaps more so, by intrinsic outcomes as they are by such extrinsic factors as money and security. C. Effectiveness of Federal Managers The creators of the CSRA also had doubts about the competence of many of the managers of the federal government. Table 3 presents the results of five INSERT TABLE 3 ABOUT HERE scales and two individual items often found to be good indicators of managerial competence. Although the values are not extreme, most managers view themselves and are seen by their subordinates as being generally effective, and as maintaining a work atmosphere conducive to employee effectiveness and productivity. There are two possible exceptions to this, however, in the nearly neutral responses to items concerning whether management devotes more attention to quantity than to quality of work, and whether supervisors are seen as taking action to deal with low performers. Both of these areas, measured by single items, were slightly less strongly endorsed than the remaining supervisor scales. Nevertheless, the overall picture is one of a system that is generally perceived by its members to be conducive to productivity and effectiveness. D. Performance-Reward Linkages The CSRA was intended to "provide incentives and opportunities...to improve efficiency and responsiveness..." of the federal workforce (p. 444). It was further assumed that managers had considerable difficulty rewarding the

5 best and most talented people. Table 4 presents scales and items that explore INSERT TABLE 4 ABOUT HERE employees' views about whether the assumed lack of linkage between good performance and positive outcomes occurred in our samples. In this case, the data essentially support the assumptions. In general, employees do not see a strong association between performing well and desirable outcomes such as opportunities for development or promotion. While their response to the question of whether their direct supervisors provide them with rewards is more favorable, even here the average response barely exceeds the midpoint of the scale. In other areas, there is on average clear disagreement with the desired contingency. E. Poor Performance-Negative Outcome Linkages The designers of CSRA believed that most federal workers are relatively immune from negative consequences even if they fail to satisfy minimal standards. As Carter put it, "Managers are weakened in their ability to...fire those who are unwilling to work" and "the sad fact is that it is easier to promote and transfer incompetent employees than to get rid of them" (p. 446). Respondents in our samples were asked whether they thought negative outcomes were associated with poor performance. Results are displayed in Table 5. These data present a mixed picture, and only some of the data support the INSERT TABLE 5 ABOUT HERE assumption. From these results, it is clear that most participants believe that penalties for poor performance are applied inconsistently. Moreover, most of the employees believe that their supervisors will not take formal action against them for unacceptable performance. These data also suggest

6 that many employees are uncertain about what would happen to them, since items relating to being demoted, transferred, or fired showed near-neutral results, with a fairly large variance. There were also marked site differences, implying that blanket generalizations about the entire federal workforce are probably inappropriate. Finally, on the scale asking whether supervisors help poor performers to improve, the responses were near neutral, though in one agency the mean was slightly positive. In general, then, the extreme conditions related to discipline that Carter's CSRA assumed do not appear to be very well supported by our data. Most respondents appeared to find the existing disciplinary system to be rather adequate. F. The Need for Formal Disciplinary Action Implicit in many of the statements supporting a need for CSRA was the idea that disciplinary actions, to be effective, must be formal, and that dismissal was appropriate much more often than had occurred prior to the CSRA legislation. In Table 6, we present the responses of supervisors concerning the desirability of various approaches to dealing with poor performance. These data generally agree with the assumption that the procedures for taking formal disciplinary action are unclear and difficult to accomplish. INSERT TABLE 6 ABOUT HERE Perhaps more important, and contrary to the designers' assumptions, the great majority of our sample believe that they can solve subordinate performance problems without recourse to formal action. Given this preference for informal action, it can be asked why emphasis was placed on the system for formal personnel actions rather than on devoting more attention to training

7 supervisors to be more effective in their informal role of helping poor performers to improve. The informal methods may be not only more cost-effective, but also more humane. G. Labor Management Relations President Carter also saw the need to improve labor management relations, with the goal of making the public sector more like the private sector in its labor-management activities. It is unclear whether the designers of the CSRA actually believed that labor management relations were less desirable than they might be. Some have suggested a more politically based motive for these provisions of the CSRA, e.g., strengthening the position of unions in the federal service to avoid their opposition to the act. Table 7 shows the responses of federal workers to six scales related to INSERT TABLE 7 ABOUT HERE labor-management relations. These data suggest that, although both workers and supervisors accept the need for unions and generally agree with their priorities, they give mostly neutral responses to questions concerning the overall quality of labor management relations in the federal agencies, and to items querying whether supervisors try to keep the union informed on matters that affect their interests. In addition, nonsupervisors generally do not trust management. Interestingly, supervisors' mean responses are also slightly negative on the question of trust, although less so than nonsupervisors. Summary The responses of more than 2,800 federal employees tend to support two of the above assumptions, to contradict two, and to present a mixed pattern for the remaining three. These data give some support to Carter's conclusion that the personnel system is time-consuming and somewhat cumbersome. There is also

8 generally strong agreement that clear linkages do not exist between high performance and other positive outcomes for the federal workers we surveyed. The data dispute the President's belief that federal workers are not motivated or dedicated to their jobs and suggest that, contrary to that assumption, government managers are seen by both themselves and their subordinates as being dedicated to high levels of performance and productivity. The data relating to the remaining assumptions are more ambiguous. Assumptions about the Process of Change Although it is still very premature to assess the success of the CSRA, few would argue that the act has had its intended dramatic influence on the federal sector. The results from our second administration of a research survey, which cannot be reported here, lend strong support to the view that very little change has occurred in our study sites as the result of the CSRA. One reason for this may be the assumptions made by the designers of the act about how persons and social systems can most effectively be made to change. A basic concept in the literature of organizational change is that participation in the implementation of a change enhances the likelihood that it will be effectively implemented and supported. A comment about our baseline data is revealing in this regard. Although our study was to have begun in September 1979, delays by the funding agency prevented approval for three months. We thus began collecting data in the spring of 1980, almost two years after the passage of the act. This is noteworthy here only because our worries about not having good baseline data proved to be unfounded. Most of our early contacts at all but the highest levels of the organizations complained that they did not understand why they were selected to be interviewed because they certainly knew nothing about the CSRA. Thus, not only did we appear to have an authentic baseline, but this also suggests that no one at our sites

9 had any part in helping to design the changes they were to implement. Not surprisingly, except for the general (but not universal) desire not to offend the researchers, our respondents initially perceived the CSRA as a nonentity. Cynicism to the effect that "we've been through this one before" was commonplace. We have already observed that the CSRA appears to have been created without benefit of a thorough, empirically based diagnosis. The act assumed that the government had a rather large number of incompetent or poorly motivated individuals, and that therefore the appropriate mechanism for change was to assist in their removal. (The training-and-improvement features of the CSRA have not yet been very extensively developed in our experience.) Changing individuals' behavior, so the intended scenario ran, would ultimately result in changing the grand bureauratic structure itself. The CSRA also assumed that all agencies and organizations should follow a similar process of change. While agency differences were accommodated in many important aspects of the act, no covered agency was permitted not to participate, or to work on personnel problems other than those mandated by the law. Although the role to be played by OPM was in many ways intended to be advisory, it was also controlling, since merit pay plans, performance appraisal systems, and executive selection and development plans had to be approved by OPM specialists. Considerable literature exists which suggests that institutional change is seldom accomplished when primary emphasis is placed on changing individuals, that ordering change is seldom congruent with internalized or long-lasting effects, and that change can seldom be successfully packaged and taken off the shelf like McDonald's hamburgers. Thus, the assumptions made by the CSRA about the process of changing were in many respects naive, perhaps mostly because no one involved with creating this law appears to have thought seriously about the most effective way to create long-lasting effects. Moreover, the problems

10 inherent in such approaches were noted in the literature from the start (e.g., Isbell, 1978; Thayer, 1978). Some agencies did employ participative type approaches, in that representatives from field installations were part of the task forces responsible for creating their specific programs for implementing the reform act. More common, however, was the procedure of developing an agency plan at the headquarters level and distributing it for a (generally cursory) review by the field locations. An interesting aspect of the fact that our studies were based at the organizational level was that we were able to observe at the grass roots level the degree of attachment and involvement felt by individuals, an often crucial requisite for the success of such a change. The modal response to CSRA was that it was something to implement because it was a requirement; most did not actively resist it but, perhaps more deadly, ignored it. Obviously, it would be difficult to involve all employees in the development of every new government program. On the other hand, the failure of the bureaucracy to think seriously about how people change probably accounts in part for the expensive failure of many programs which begin with such good.intentions. While it is not the purpose of this paper to suggest alternative ways of designing change, it does appear clear that simply passing a law is unlikely to have great and long-lasting impact. Assumptions about the Environment The final set of assumptions on which the Reform Act was based were curious ones, involving an apparent belief in an apolitical environment. Whatever else history may record of the Carter administration, ability to work within a political framework is unlikely to be high on any list of strengths. The very nature of the CSRA assumed that the federal government was in need of vast reform. The law, even in the somewhat watered-down version which

11 ultimately passed Congress, was intended to produce a major upheaval in a weather-beaten and somewhat corrupt system, an approach hardly designed to win over converts from the federal agencies. The law also mandated an external evaluation of the act, resulting in part in the five-year organizational studies reported earlier. While five years may seem a relatively short time period in a longitudinal research project, the history of a nation, or the life of a sequoia tree, it is something akin to light years in the political context. The anticorruption votes that helped to make Carter President and facilitated the passage of the CSRA in the first place were soon to be forgotten in the Reagan landslide. Ironically, Reagan was elected in part on a platform of controlling waste and fraud in government, views quite consistent with the spirit and intent of the CSRA. Yet Carter's legacy could hardly be Reagan's mandate, and OPM was to put CSRA, and particularly its careful evaluation, on the same level as research on promulgating pacificism or expanding the Department of Education. Changes in presidential administration are only one of the political realities with which the CSRA has had to contend. Court rulings (e.g., the Weingarten decision and Wells v. Harris), GAO orders to withhold merit payouts, the PATCO firings, and a reconceptualization of the proper role of the Office of Personnel Management all had to be accommodated and served to dilute —permanently or temporarily —the implementation of CSRA. The assumption of an apolitical environment which could be "held off" for at least five years to determine whether the Reform Act would work was untenable. It is easy to say that this set of assumptions embodied the prior assumption of Carter's reelection, but if the intent was to make long-lasting change, the CSRA would have had to be more "administration-proof" (like deficit financing and Social Security). Thus, by the third year of our study, when many aspects of the CSRA were just beginning to be put into place, the research teams

12 suddenly had to consider whether the study would continue to be funded at all.- The uncertainty of the survival of many aspects of the Reform Act, (e.g., delegated authority) in the intended form created similar concerns on the part of many of our research respondents. Summary and Conclusions This paper has reviewed three major types of assumptions made by the Civil Service Reform Act of 1978. Empirically, the beliefs about prior conditions of the federal workforce were partly supported by a study of over 2,700 federal workers. In most cases, the conditions on which the CSRA were based were not nearly so extreme as the creators of the act had assumed. In some instances, the assumptions were not valid. Equally important for its eventual success, the Reform Act appeared to be based on rather little understanding of how people and organizations change. The creators of the act were also naively optimistic about the act's ability to withstand political influences long enough to be both successfully implemented and evaluated. The ultimate success of the CSRA may well be far less than it might otherwise have been if greater attention had been directed from the outset to issues such as those presented in this paper.

13 Footnotes 1. Subsequent to the writing of this paper, the intended five year assessment being carried out at The University of Michigan was terminated by OPI at the end of its third year.

14 References Carter, J. Federal civil service reform. In Weekly Compilation of Presidential Documents, 1978, 14 (9), 444-448. Isbell, F. Carter's CSRA: 35% ifs, buts and maybes. The Civil Liberties Review, 1978, 5, 6-15. Organizational assessment of the effects of civil service reform. Ann Arbor: Institute for Social Research, 1980. Thayer, F. C. The president's management "reforms": Theory X triumphant. Public Administration Review, 1978, 38, 309-314.

Table 1 Effectiveness of the Personnel Systems* N Mean Standard Deviation 1.75 ITEMS It is difficult for me to get competent people to fill job openings. 518 4.4 It takes too long to process the paperwork 521 5.5 1.Z needed to fill vacancies here. The Personnel Office helps me find good 520 3.8 1. people when I need them. Most new employees who come to work for me 520 4.3 1. have the skills to do their jobs. I don't have enough authority to hire 520 4.3 1. competent people when I need them. *All items on this table were completed by supervisory personnel only. aResponses range from (1) Strongly Disagree - - -> (4) Neither Agree nor Disagree - - -> (7) Strongly Agree. 47 67 64 85

Table 2 Employee Work Orientation and Motivation SCALES N Mean Intrinsic importance - (of doing well on your job and knowing how well you are doing.) - 2 items Extrinsic reward importance - (How important are the amount of pay you receive and your job security?) - 2 items Organization involvement (What happens to this organization is really important to me.) - 2 items ITEMS I have all the skills I need to do my job. 2751 2740 2762 2683 5.7 5.5 5.8 5.2 Standard Deviation 1.36 1.39 1.01 1.51 Doing my job especially well gives me a good feeling. I get upset when I fall behind in my work. I put in extra time on my job when I fall behind schedule. I generally produce high quality work on my job. 2776 2770 2769 2773 6.3.82 5.1 1.53 4.7 6.0 1.83 0.83 aResponses range from (1) Strongly Disagree - - -> (4) Neither Agree nor Disagree - - -> (7) Strongly Agree.

Table 3 Managerial Effectiveness and Work Environment SCALES Clarity of expectations about what goals one is expected to achieve (3 items) Work group openness and cooperation (4 items) Supervisory emphasis on productivity (2 items) Effectiveness of supervisor as manager (2 items) Interpersonal competence of supervisors (3 items) ITEMS Management devotes more attention to quantity of work than to how well I do it. My supervisor criticizes people who perform poorly. N 2728 2273 2657 2655 2612 2749 2781 Meana 5.3 4.5 5.0 4.7 4.7 4.0 4.2 Standard Deviation 1.09 1.12 1.27 1.53 1.42 1.89 1.83 aResponses range from (1) Strongly Disagree - - -> (4) Neither Agree nor Disagree - - -> (7) Strongly Agree.

Table 4 Perceived Linkage between Performance and Rewards Standard SCALES N Meana Deviation Development contingency - (e.g., If you per- 2775 2.8 1.30 form well, how likely is it that you will have an opportunity to develop your skills and abilities?) (4 items) Pay equity - (e.g., Considering what others 2750 3.7 1.58 in your area and agency receive, how fair is your pay?) (3 items) ITEMS Promotion decisions here are based on merit. 2773 3.3 1.76 My supervisor rewards me for good 2787 4.2 1.97 performance. If you perform your job especially well, 2808 2.4 1.66 how likely is it that you will get a cash award, an above-average pay increase, or a quality step increase? aResponses range from (1) Strongly Disagree - - -> (4) Neither Agree nor Disagree - - -> (7) Strongly Agree.

Table 5 Outcomes Associated with Poor Performance SCALES N Mean Standard Deviation 0.96 1.51 Perceived equity of the discipline system 2715 4.5 4.1 Superisor assistance to avoid poor performance 2630 ITEMS Some poor performers here get penalized while others are left alone. 2780 5.3 1.37 If your performance is unacceptable, how likely is it that your supervisor will start a formal personnel action against you? I will be demoted or transferred if I perform poorly. If I perform poorly, I will lose my job. 2762 2802 2802 3.2 4.2 3.8 1.60 1.82 1.84 aResponses range from (1) Strongly Disagree - - -> (4) Neither Agree nor Disagree - - -> (7) Strongly Agree.

Table 6 Ease, Clarity, and Usefulness of Formal Disciplinary Measures Standard N Meana Deviation SCALES Difficulty of taking formal adverse personnel action* (3 items) Clarity of adverse personnel action procedures (2 items) ITEMS I can usually solve subordinate performance problems without having to initiate a (formal) personnel action.* 510 511 514 4.8 4.7 1.34 1.23 5.4 1.04 *Supervisory responses only. aResponses range from (1) Strongly Disagree - - -> (4) Neither Agree nor Disagree - - -> (7) Strongly Agree.

Table 7 Labor-Management Relations Standard SCALES N Mean Deviation Constructive union-management relations exist 2676 4.0 1.02 in this organization*** (3 items) Overall acceptance of the necessity and 2671 4.7 1.17 priorities of the union*** Supervisor acceptance of necessity and 651 4.6 1.32 priorities of the union* Management involvement of union (e.g., "I try 669 4.0 1.36 to keep union leaders informed about changes...")* - 2 items Trust of management** - 5 items 2279 3.0 1.10 Trust of management* - 4 items 507 3.8 1.30 * Supervisors only ** Nonsupervisors only *** All Respondents aResponses range from (1) Strongly Disagree - - -> (4) Neither Agree nor Disagree - - -> (7) Strongly Agree.