Representative Bureaucracy, Organizational Effectiveness and the South African Public Service Sergio Fernandez Indiana University School of Public and Environmental Affairs & Sam Koma University of Pretoria School of Public Management and Administration
Theory of Representative Bureaucracy Focus on the relationship between demographic representation and • bureaucratic decisions, outputs and outcomes that benefit represented group (Kingsley 1944; Mosher 1986; Krislov 1974) • Empirical studies show increasing bureaucratic representation can produce substantive benefits for represented groups and improve program/organizational effectiveness (Meier 1993; Meier, Wrinkle and Polinard 1999; Meier and Nicholson-Crotty 2006; Fernandez, Smith and Malatesta 2013; Wilkins and Keiser 2006: Bradbury and Kellough 2008) • Theoretical argument: – Members of a social group undergo similar socialization experiences and come to share values and attitudes that set them apart from others – Values and attitudes influence behavior – Bureaucrats with discretionary authority act in ways that advance their interests and those of their social group
Theory of Representative Bureaucracy Preconditions for producing substantive benefits • – Bureaucratic discretion during implementation – Policy saliency for represented group – Ability to impact represented group • Sources of direct substantive benefits (Lim 2006; Hindera 1993) – Direct • Advocacy or partiality • Better communication between bureaucrats and citizens • Empathetic understanding by bureaucrats of the needs and interests of citizens – Indirect • Efforts by bureaucrats to monitor and mitigate bias expressed toward their group by other bureaucrats • Positive changes in behavior of citizens (e.g., greater program uptake and compliance with the law)
The South African Context A public service broadly representative of the population is a constitutional • principle and legislative goal • To be achieved through affirmative action and related policies • Government pronounced an instrumental role for representative bureaucracy: – Greater representation of Blacks, women and those with disabilities will increase the public service’s legitimacy, competency, responsiveness and effectiveness (DPSA 1995, 1997; Ndletyana 2008) Today, in general, public service closely mirrors the South African • population in terms of racial and gender representation • Critics argue, however, that transformation has adversely affected managerial capacity, staffing and morale
Research Questions Does representation of historically disadvantaged groups – Africans, Coloureds, Indians and women – impact the effectiveness of national departments? What are the causal mechanisms by which greater representation of historically disadvantaged groups improves the effectiveness of these organizations?
Data and Methods • Mixed methods approach – econometric analysis and semi-structured interviews • Quantitative data: department annual reports, Department of the Presidency (MPAT scores), Auditor-General (audit scores) • Interview data: high ranking officials in national departments, with knowledge and experience related to transformation and affirmative action • Samples: – 188 national departments (2006-2014, unbalanced panel) – 15 interviewees in 7 national departments Statistical modeling (in Stata 12) • – Random Effects Model (unbalanced panel, grouped by department, with dummy variables for years) – Cluster robust standard errors
Data and Methods • Dependent variable – Percent of goals in “ Programme Performance “ sub-section of department ’ s annual report achieved in year • Goals must be linked to departmental outcomes, strategic priorities of the government and goals in performance contracts of senior managers • “Competing values” measure of organizational effectiveness – most goals relate to program outputs and outcomes, some focus on resource acquisition and system maintenance activities • Independent variables – Percent African, percent Coloured, percent Indian, percent White and percent female; race and gender of department head • Control variables – Race and gender of minister, percent employees trained, median occupational level, total expenditures, total employees, turnover rate, vacancy rate, total goals set, clean audit result, reporting rate, percent goals uncertain, percent Blacks and women trained, percent Blacks and women promoted, dummy variables for policy type (cluster) and year
Data and Methods • Semi-structured interviews – From 45 minutes to 2.5 hours – Interviews were conducted March 27-April 2, 2014 and March 16-27, 2015 – Field notes, but no audio recording – Steps taken to maintain confidentiality – Responses regarding underlying causal mechanisms coded using categories from Department of Public Service and Administration (1997, 1998) and Lim (2006)
Variables and Measures Variable Definition percent goals achieved Percent of goals established by department that was achieved during year percent african Percent of department ’s workforce who report being African percent coloured Percent of department ’s workforce who report being Coloured percent indian Percent of department ’s workforce who report being Indian/Asian percent female Percent of department ’s workforce who report being female department head black Coded 1 if department head is either African, Coloured or Indian department head female Coded 1 if department head is female minister black Coded 1 if department minister is either African, Coloured or Indian minister female Coded 1 if department minister is female vacancy rate Percent of positions unfilled in department during year turnover rate Percent of employees who left department during year total expenditures (log) Log transformation of department ’s total budget expenditures during year total employees (log) Log transformation of department ’s total employees during year total goals Number of goals established by department during year, as reported in department ’s annual report percent employees trained Percent of department ’s total employees sent to training during year median skill level Median occupational level for department employees during year (6 = top management, 5 = senior management, 4 = professional and qualified, 3 = skilled, 2 = semiskilled, 1 = unskilled) reporting rate Percent of years department filed an annual report (from 2006 to 2014, only years department was in existence considered) audit result Coded 1 if department earned a clean audit result (unqualified with no findings) percent goals uncertain Percent of goals established by department that could not be coded as achieved or not achieved percent african trained Percent of department ’ s African employees sent to training during year percent coloured trained Percent of department ’ s Coloured employees sent to training during year percent indian trained Percent of department ’ s Indian employees sent to training during year percent female trained Percent of department ’ s female employees sent to training during year percent african promoted Percent of department ’ s African employees promoted during year percent coloured promoted Percent of department ’ s Coloured employees promoted during year percent indian promoted Percent of department ’ s Indian employees promoted during year percent female promoted Percent of department ’ s female employees promoted during year �
Descriptive Statistics 1 0.9 0.8 0.7 0.69 0.7 0.65 0.63 0.61 0.6 0.6 0.6 0.6 0.59 Percent of Goals 0.5 Achieved 0.4 0.3 0.2 0.1 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 Year
Descriptive Statistics 1 0.9 0.8 0.8 0.8 0.78 0.78 0.7 0.74 0.74 0.71 0.69 0.64 0.6 Percent of Percent African Department 0.5 Percent Coloured Workforce Percent Indian 0.4 Percent White 0.3 0.22 0.18 0.17 0.2 0.15 0.14 0.12 0.12 0.1 0.10 0.09 0.09 0.1 0.08 0.07 0.1 0.07 0.07 0.06 0.06 0.04 0.04 0.04 0.03 0.04 0.03 0.03 0.03 0.03 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 Year
Descriptive Statistics 1 0.9 0.8 0.7 0.6 0.54 0.53 0.53 0.53 0.53 0.53 0.53 0.52 0.52 Percent of Department 0.5 Percent Female Workforce 0.48 Percent Male 0.47 0.48 0.47 0.47 0.47 0.47 0.47 0.46 0.4 0.3 0.2 0.1 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 Year
Regression Results Variable Model 1 Model 2 Model 3 Model 4 Coef. Coef. Coef. Coef. percent african 0.93*** percent coloured 1.51*** 0.51 percent indian 2.77*** 1.74** percent white -0.87*** percent female -0.46** -0.44** percent african male 1.26*** percent coloured male 1.66** 1.35* percent indian male 2.98*** 2.74** percent white male -0.35 percent african female 0.63*** percent coloured female 1.55* -0.05 percent indian female 2.41* 1.63 percent white female -1.12** black department head 0.00 0.00 0.01 0.01 female department head 0.01 0.00 0.01 -0.01 black minister -0.06 -0.07 -0.06 -0.05 female minister 0.06** 0.06** 0.06** 0.06** constant 0.47** 1.38*** 0.21 1.16*** N 188 188 188 188 R 2 , Within/Between/Overall 0.36/0.71/0.58 0.36/0.68/0.55 0.36/0.73/0.57 0.33/0.69/0.56 *** p < 0.01 ** p < 0.05 * p < 0.10 Note: Models 1-4 include all control variables and dummy variables for years (2006-2014, 2006 is excluded year). �
Recommend
More recommend
