Principal leadership changes and their consequences for school performance in South Africa Gabrielle Wills Stellenbosch University, Department of Economics, South Africa A R T I C L E I N F O Article history: Received 15 April 2016 Received in revised form 11 August 2016 Accepted 25 August 2016 Keywords: Educational administration Principal turnover Teacher turnover Retirement Fixed effects South Africa A B S T R A C T School leadership changes are gaining momentum in South Africa as larger proportions of incumbent principals near retirement age. While this presents opportunities to replace weaker school leaders with better ones, these changes may also destabilize school environments and impede on learning. This study explores this issue using payroll data on public school principals in South Africa linked to national data on schools and matriculation examination outcomes. School fixed effects and propensity score matching with difference-in-difference estimations confirm that principal changes are indeed detrimental to school performance with larger disruptive effects observed in poorer schools. ã 2016 Elsevier Ltd. All rights reserved. 1. Introduction Internationally, the age profile of school leaders is rising across high and middle-income countries (Pont et al., 2008). With confirmatory evidence on the importance of quality school leadership for student outcomes (Branch et al., 2012; Coelli and Green, 2012; Chiang et al., 2016; Grissom et al., 2015), rising principal retirements present an opportunity for improving the leadership trajectories of schools if better quality replacements are hired. This is especially the case where school leaders have long average tenure and are seldom dismissed for poor performance. However, anticipated principal exits also pose significant chal- lenges. In addition to the cost of recruiting principal replacements, there may be unintended negative consequences of leadership changes (or otherwise referred to as principal turnover) as these events disrupt the school environment and impact on learning. Using a unique administrative dataset linking South African data on schools to payroll information on the population of public school principals, this study considers how school leadership changes impact on student performance in the short to medium term across secondary schools. The country is facing a substantial and rising number of school principal retirements (Wills, 2015). This occurs in a context of growing concern about corruption, nepotism and union interference in the appointment and promotion of school personnel (RSA DBE, 2016). Moreover, current policies and their implementation in the area of school leadership development, selection, hiring and performance management remains weak while no policies exist to manage school leadership successions. In estimating the relationship between school leadership changes and student performance, a key challenge is disentangling the impacts of a leadership change from various factors that may influence both a principal’s decision to move out of a school and student learning. A school fixed effects strategy including various time-varying school and student compositional characteristics is initially used to control for unobserved school factors that may bias estimates. The results suggest that leadership changes are very detrimental to school performance in the short to medium term, particularly if the disruption is initiated by a principal exit from public education. The validity of the school fixed effects results is supported using an alternative approach which combines propen- sity score matching with difference-in-difference estimation (PSM-DiD) 1 following Heckman et al. (1997). I obtain a suitable counterfactual group of schools not experiencing a leadership change by constraining the potential control group of schools to E-mail address: gabriellewills@gmail.com (G. Wills). 1 Abbreviations/Acronyms used in this article: CIA—Conditional Independence Assumption; DBE—Department of Basic Education; EMIS—Education Management Information System; FET—Further Education and Training (grades 10–12 level); NSC—National Senior Certificate; OLS—Ordinary Least Squares; PSM-DiD—Propen- sity score matching with Difference-in-Difference; REQV- Relative Educational Qualification Value (a system of grading professional and academic qualifications); SGB—School Governing Body. http://dx.doi.org/10.1016/j.ijedudev.2016.08.005 0738-0593/ã 2016 Elsevier Ltd. All rights reserved. International Journal of Educational Development 51 (2016) 108–124 Contents lists available at ScienceDirect International Journal of Educational Development journa l homepage: www.e lsevier.com/locate/ijedudev