Portal de Periódicos da CAPES

External Control and Red Tape: The Mediating Effects of Client and Organizational Feedback

2012; Taylor & Francis; Volume: 15; Issue: 3 Linguagem: Inglês

10.1080/10967494.2012.725291

1559-3169

Gene A. Brewer, Richard M. Walker, Barry Bozeman, Claudia N. Avellaneda, Gene A. Brewer,

Auditing, Earnings Management, Governance

ABSTRACT Bozeman's (1993; 2000) external control model of red tape posits that organizations with higher degrees of external control will have higher levels of red tape. According to the model, this effect is compounded by entropy affecting the communication of rules and their results, limited discretion over rules and procedures, and non-ownership of rules. However, the model predicts that red tape will be mediated by communication from clients and within the organization. Bozeman's model is often cited in the literature, but it has not been subjected to comprehensive empirical verification. This study tests the model using data from a multiple informant survey of 136 upper-tier English local government authorities conducted in 2004 and several secondary sources. Statistical results show that external control does indeed lead to higher levels of red tape. We then test a number of organizational feedback mediators and find that client feedback does little to mediate the effects of red tape; the major factors are trust between politicians and public managers and devolved management. We discuss these findings and propose some changes to the model. ACKNOWLEDGEMENTS This work was supported by the National Research Foundation of Korea Grant funded by the Korean Government (NRF-2011-330-B00194). Earlier versions of this article were presented at the Twelfth Annual Meeting of the International Research Society for Public Management (IRSPM-XII), Brisbane, Australia, March 26–28, 2008; and the 20th Anniversary and 11th Biennial Public Management Research Association (PMRA) Conference, Syracuse, New York, June 2–4, 2011. Notes †p < .10; *p < .05; ***p < .001. Notes: Path c: Model with Red Tape (DV) Regressed on Political Climate (IV), other variables acting as controls. Path a: Model with Mediator (Trust: officers and politicians) regressed on Political Climate (IV), other variables acting as controls. Paths b and c′: Model with Red Tape (DV) regressed on Mediator (Trust: officers and politicians) and Political Climate (IV), other variables acting as controls. †p < .10; *p < .05; ***p < .001. Notes: Indirect effect =.609. Direct effect =.297. Total effect =.907. Proportion of total effect that is mediated =.67. Ratio of indirect to direct effect = 2.04. Notes: (P) = percentile confidence interval; (BC) = bias-corrected confidence interval. Notes: Path c: Model with Red Tape (DV) Regressed on Labour Vote Share (IV), other variables acting as controls. Path a: Model with Mediator (Trust: officers and politicians) regressed on Labour Vote Share (IV), other variables acting as controls. Paths b and c′: Model with Red Tape (DV) regressed on Mediator (Trust: officers and politicians) and Labour Vote Share (IV), other variables acting as controls. †p < .10; *p < .05; ***p < .001. Notes: Indirect effect = −.018. Direct effect = −.003. Total effect = −.021. Proportion of total effect that is mediated =.84. Ratio of indirect to direct effect = 5.59. Notes: (P) = percentile confidence interval; (BC) = bias-corrected confidence interval. Notes: Path c: Model with Red Tape (DV) regressed on Non-employment Rate (IV), other variables acting as controls. Path a: Model with Mediator (Devolved Management) regressed on Non-employment Rate (IV), other variables acting as controls. Path b and c′: Model with Red Tape (DV) regressed on Mediator (Devolved Management) and Non-employment Rate (IV), other variables acting as controls. †p < .10; *p < .05; ***p < .001. Notes: Indirect effect: −.051. Direct effect: −.055. Total effect: −.106. Proportion of total effect that is mediated =.48. Ratio of indirect to direct effect =.93. Notes: (P) = percentile confidence interval; (BC) = bias-corrected confidence interval. Buchanan's measure of red tape was identical to the measure that many other researchers had used as an index of formalization. It is perhaps more accurate to say that his study is the first published work in the public administration literature to focus on formalization measures. Bozeman (2000, 82) later changed the litmus test for red tape from rules that fail to achieve their functional purpose to those that are burdensome and negatively impact organizational performance. In the organizational studies literature, entropy is used to convey weakening of bonds in organizational systems (Leifer Citation1989; Oliver Citation1992; Parks 1983; McKelvey Citation1997), though its origins are in thermodynamics and information theory. The usage here follows Bozeman's (Citation2000, 107) usage "as a metaphor for understanding processes of change in organizations" and particularly in relation to the tendency of social bonds and interactions to dissipate. Thus, we expect that, all else equal, there will tend to be a reduction in the amount of energy devoted to rule implementation as a function of greater time between the rule's origins and subsequent implementations. In other words, rule implementation becomes routine even though the rule environment may change. While we draw from several data sources to study the external control model of red tape, we are unable to operationalize the concept of discretion. Echelons are levels in the organizational hierarchy (for example, see Aiken and Hage Citation1968, 918). Seven key services were surveyed: education, social care, land-use planning, waste management, housing, library and leisure, and benefits. Pandey and Kingsley's (Citation2000, 782) definition of red tape was also useful in designing this study: "impressions on the part of managers that formalization (in the form of burdensome rules and regulations) is detrimental to the organization." Simply put, red tape exists when managers view formalization as burdensome and detrimental to organizational purposes (Pandey and Scott, Citation2002, 565). The groups comprised 12 National Statistics Socio-Economic Classifications: large employers and higher managerial occupations, higher professional occupations, lower managerial and professional occupations, intermediate occupations, small employers and own account workers, lower supervisory and technical occupations, semi-routine occupations, routing occupations, never worked, long-term unemployed, full-time students, and non-classifiable. The models are not troubled by multicollinearity: the highest variance inflation factor recorded in all three models was 1.86—well below the level of 10 at which it can become a concern. We tested the mediation effects one at a time rather than lumping them together in an omnibus test in order to calculate the direct and indirect effects for each mediator. This revealed some effects that were otherwise masked. These tests are performed using the Stata command sgmediation and specifying the option bootstrap.. We recognize the potential of having a third variable acting as a confounder or a suppressor rather than as a mediator. To reduce this confusion, we followed MacKinnon, Krull, and Lockwood's (Citation2000) conceptual differences among the three types of effects. For instance, we followed MacKinnon, Krull, and Lockwood's (Citation2000) interpretations of signs on the indirect and total effects in order to differentiate suppression from mediation effects. Moreover, the potential for confounding effects is undermined by our theoretical justifications, which propose a causal relationship from the independent variables to the mediators—contrary to having a causal relationship from the mediator to the independent variable. From an open systems perspective managers in the organization react to signals from the environment and put in place managerial and organizational behaviors to overcome red tape. See Kline (Citation2005, 128–131) for more details about estimation of indirect and direct effects in mediation tests. Testing mediation effects with cross-sectional data, as was the case in this study, raises concerns about the causal relations implied by mediation models (Cole and Maxwell Citation2003; Maxwell and Cole Citation2007). In fact, "mediation is a causal chain involving at least two causal relations (e.g., X →M and M→Y), and a fundamental requirement for one variable to cause another is that the cause must precede the outcome in time" (Maxwell and Cole Citation2007, 559). The norm has been graduating from cross-sectional to longitudinal designs to make more rigorous inferences about the causal relations. Therefore, future studies should test the meditational hypotheses suggested in this study with longitudinal data covering several years to avoid the potential misleading results of cross-sectional meditational tests (Maxwell and Cole Citation2007). Additional informationNotes on contributorsGene A. Brewer Gene A. Brewer (cmsbrew@uga.edu) is Associate Professor of Public Administration and Policy at the University of Georgia and Visiting Professor of Public Management at Utrecht University in the Netherlands. He is an internationally recognized expert on public management research and practice. Richard M. Walker Richard M. Walker is Professor of Public Management and Policy in the Department of Public and Social Administration at City University of Hong Kong. His most recent books are Strategic Management and Public Service Performance (Macmillan Palgrave) and Public Management and Performance: Research Directions(Cambridge University Press). Barry Bozeman Barry Bozeman is Regents' Professor of Public Policy and Ander Crenshaw Chair at University of Georgia. His research focuses on public management and science policy. Claudia N. Avellaneda Claudia N. Avellaneda is an Assistant Professor at the University of North Carolina-Charlotte. Her research interests include public management, comparative politics, and comparative public policy with a regional focus on Latin America. Gene A. Brewer Gene A. Brewer, Jr., is Assistant Professor of Psychology at Arizona State University. His research focuses on the relationship between human memory and attention. He often consults in other areas on issues related to research methods and statistics.