Author:Hayes, Michael S.

    Understanding the factors that influence employee turnover is important because personnel decisions have important repercussions for organizations and organizational performance (Kim, 2002; Meier and Hicklin, 2007; Pitts, 2005; Ronfeldt, Loeb, and Wyckoff, 2013; Shaw, Delery, and Jenkins, 2005). For example, in the context of public education, Guarino, Santibanez, and Daley (2006) suggest that the most effective teachers are more likely to leave their school relative to the less effective teachers. However, there is also research that suggests removing ineffective teachers through performance-assessment reforms can improve student outcomes (Adnot et al., 2017). Either way, actual turnover--even healthy turnover--has significant financial costs on organizations (e.g. Meier & Hicklin, 2007; Park & Shaw, 2013). For example, estimates of the cost of teacher turnover range between $12,000 to $52,500 per teacher (Alliance for Excellent Education, 2005; Texas Center for Educational Research, 2000).

    The literature on employee turnover in public management and organizational theory focuses primarily on how employee- and organizational-level characteristics influence employee turnover. However, other determinants of employee turnover, such as factors external to the organization, are understudied (Mobley et al., 1979; Moynihan and Pandey, 2008; Meier and O'Toole, 2009; Llorens and Stazyk, 2011). The current study contributes to this literature by focusing on the interaction of two external factors: environmental shocks to organizations and budgetary constraints on organizations. (1)

    Specifically, the current study examines how the passage of the No Child Left Behind Act of 2001 (NCLB) and state-imposed binding tax and expenditure limitations (TELs) on school districts interact to affect the likelihood of teacher turnover. Previous research suggests that the passage of NCLB created an environmental shock by placing a significant financial burden on school districts by requiring them to invest in "highly qualified teachers" and designing and implementing new student assessments (Dee, Jacob, and Schwartz, 2013; Goertz, 2005; Hayes, 2015; McGuinn, 2005). (2) At the same time, the passage of NCLB had non-financial effects on teachers, such as diminishing teachers' classroom autonomy, reducing teachers' perceived job security, and increasing teachers' stress levels (Daly and Chrispeels, 2005; Figlio and Loeb, 2011; Reback, Rockoff, and Schwartz, 2014). Interestingly, there has been mixed evidence on the impact of the passage of the NCLB on teachers' job attitudes, satisfaction, and mobility (Feng, Figlio, and Sass, 2018; Grissom, Nicholson-Crotty, and Harrington, 2014; Reback, Rockoff, and Schwartz, 2014; Sun, Saultz, and Ye, 2017).

    One possible reason for this mixed evidence on teacher mobility is that teachers were not impacted equally by the fiscal and non-fiscal shocks of the passage of NCLB. For example, some states impose budgetary constraints on school districts. Common state-imposed budgetary constraints on school districts are binding TELs, which restrict districts' abilities to raise additional revenue or increase expenditures. School districts in states with binding TELs were less likely to be able to respond to the environmental shock of NCLB by raising additional own-source revenue (Hayes, 2015). For example, as Figure 1 shows, the gap in local revenue per pupil between school districts in states without binding TELs compared to those school districts in states with binding TELs increased following the passage of the NCLB act. As shown in Figure 1, there was a statistically insignificant gap in per-pupil local revenue during the pre-NCLB years. However, in the first year of the passage of NCLB, this gap in local revenue increased by more than $300 per pupil relative to the 2001-02 academic year and continued to grow over the next several academic years. This suggests that school districts in states without binding TELs were able to increase local revenue more than school districts with binding TELs following the passage of NCLB.

    Notes: This figure is based on the author's calculations and the full analysis is available on request. Figure 1 reports coefficients from an event study analysis that estimates the gap in local revenue per pupil between school districts without binding TELs compared to those with binding TELs for academic years before the passage of the NCLB and academic years during the implementation of the NCLB. 2003 is the 2002-03 academic year and the first year of NCLB. The base year is the 2001-02 academic year, which is the last year before the implementation of NCLB. Data comes from all U.S. public school districts reported in the Local Education Agency Finance Survey (F-33) collected by the National Center for Education Statistics. Standard errors are reported in parentheses and are clustered at the state-level, *** p

    The current study tests the hypothesis that the environmental shock of NCLB increases the likelihood of teacher turnover relatively more in states with binding TELs compared to states without binding TELs. Using a nationally representative teacher-level dataset, the current study uses a difference-in-difference type strategy to examine the change in the likelihood of teacher turnover between teachers in states with and without binding TELs on school districts following the passage of NCLB. The main results suggest the likelihood of turnover increased relatively more for teachers in states with binding TELs relative to all other teachers after the passage of the NCLB. Interestingly, teachers did not necessarily move to another school, but it appears that these teachers were exiting the teaching profession altogether. For example, following NCLB, a teacher in a state with a binding TEL is 8.7 percentage points more likely to leave the teaching profession relative to all other teachers. This higher likelihood of leaving the teaching profession is even higher in states without a school accountability policy prior the passage of NCLB.

    The remainder of this paper is organized into five sections. Section 2 reviews the relevant literature and presents testable hypotheses. Sections 3 and 4 describe the dataset and the empirical methodology. Section 5 presents the main results and section 6 concludes with a discussion of the implications of these results.


    The current study sits at the intersection of three literatures: the determinants of employee turnover, the fiscal and non-fiscal shocks caused by NCLB, and tax and expenditure limitations (TELs). In this section, I use the key findings from these literatures to develop testable hypotheses, which I test empirically later in the paper.

    2.1 Determinates of Public-Sector Employee Turnover

    Numerous predictors of employee turnover have been found in prior research. For thorough reviews from the public management and education policy literatures, see Grissom, Viano, and Selin (2016), Guarino, Santibanez, and Daley (2006), Moynihan and Pandey (2008), and Pitts, Marvel, and Fernandez (2011). Below, I summarize the various employee- and organizational-level predictors of employee turnover found in these prior reviews. Additionally, I utilize these previous studies to identify appropriate control variables for the empirical model.

    The decision to leave an organization is influenced by many employee demographics including age, years of work experience, race, gender, and worker qualifications. First, older and more experienced employees are more likely to remain in the same organization (e.g. Guarino, Santibanez, and Daley (2006); Lambert, Hogan, and Barton, 2001). Second, there are mixed findings on whether turnover is related to employee race/ethnicity (Blau and Kahn, 1981; Choi, 2009; Ingersoll, 2001; Kellough and Osuna, 1995; Kirby, Berends, and Naftel, 1999). Third, recent research suggests female employees are less likely to leave their organization than their male counterparts (Guarino, Santibanez, and Daley, 2006). Lastly, the likelihood of a teacher leaving their school is higher if the teacher has a graduate degree, higher standardized test scores, and less teaching experience (Borman and Dowling, 2008).

    Organizational characteristics also influence the likelihood of employee turnover. Not surprisingly, employee satisfaction with pay, benefits, and career advancement are strong predictors of turnover. Research finds higher salaries and more generous benefits decrease the likelihood of employee turnover (Blau and Kahn, 1981; Borman and Dowling, 2008; Grissom and Anderson, 2012; Grissom and Mitani, 2016; Shaw et al., 1998). Additionally, previous research suggests that organizations with more opportunities for career advancement and promotion have lower employee turnover rates (Lee and Whitford, 2008; Selden and Moynihan, 2000). One possible explanation for this negative relationship between opportunities for career advancement and turnover intention is that access to more opportunities for career advancement promotes higher levels of job satisfaction and organizational commitment (Pitts, Marvel, and Fernandez, 2011). The employee's relationship with his or her supervisor is another important factor. More effective public managers tend to lower the likelihood of employee turnover. For example, public managers can decrease rates of employee turnover through better communication, providing more role clarity, and creating a sense of trust (Kim, 2002; Kim, 2005; Pitts, Marvel, and Fernandez, 2011). Grissom (2011) also finds that schools with more experienced principals have relatively lower rates of teacher turnover.

    Schools with more economically disadvantaged students have relatively higher turnover rates (Ingersoll, 2001; Hanushek, Kain, and Rivkin, 2004). Similarly, schools located in an urban area have higher teacher turnover rates relative to schools located in suburban areas (Ingersoll...

To continue reading