اثر پرداخت اداری و مالیات بر املاک محلی بر نمرات موفقیت تحصیلی دانش آموزان : مدارک و شواهد از مدارس عمومی نیوجرسی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|5296||2013||16 صفحه PDF||سفارش دهید||12615 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Economics of Education Review, Volume 34, June 2013, Pages 1–16
We theorized that student test score performance will be positively related to the percentage of school district revenues raised from local taxes and with salary levels of school district administrators. Using both fixed and random effects panel analyses, we examine data for 217 Kindergarten-to-Grade 12 school districts in New Jersey for the years 2002–2009. Our results support the inference that increases in the percent of school funds raised locally have a positive influence on student test scores. However, the results for our hypothesis involving administrative costs were mixed. Administrative salaries and administrative spending were found to be positively related to test score performance in the one-way time fixed effects model, but not in the two-way models. Finally, classroom spending and the student–faculty ratio were found to be positive and significant in some of the tests, although not robust to alternative specifications.
The international financial crisis of 2008 and the resulting downturn in the economy have put public education funding under immense pressure in the Unites States. As a result, an active debate rages over cost benefit analysis of each dollar spent on public education and its impact on the quality of education. So while on the one hand, it is interesting to look for the best source of funds for our schools, it is equally interesting to ask what these funds should be best spent on. This empirical study investigates the relative influence of funding sources for public education as well as the role of the administrative function on student achievement scores. Over the last decade student achievement scores have often been used as a benchmark for the quality of education provided. This reliance on test scores as a measuring stick for school quality coincides with the enactment of the No Child Left Behind Act of 2001, which triggered increased testing at all levels of elementary and secondary education. The American Recovery and Reinvestment Act of 2009 included $4.35 billion in competitive grants for educational reform under President Obama's ‘Race to the Top’ initiative. This initiative provided for further increases in the use of standardized tests, and at the same time links teacher and principal salaries to student performance. In light of the increased frequency and utilization of standardized tests, additional research on both the optimal mix of funding sources and cost benefit analysis of expenditures on administration is warranted to guide policy-makers in their efforts to reform public education in the U.S.A. This study examines two aspects of the spending-test score relation using a sample of 217 Kindergarten-to-Grade 12 (K-12) school districts in the State of New Jersey. Our first research question focuses on the value of the administrative function on student achievement scores. This question is motivated by the widespread belief that administrators, either through malfeasance, ignorance, or indifference to the objectives of the institutions they serve, divert resources from value-added, education-enhancing activities, to non-value-added administrative activities. The prevalence of this view is clearly exhibited in actions by several state legislatures in the United States to cap spending on administration in public schools. For example, the Chicago School Reform Act passed by the Illinois Legislature in 1988 imposed constraints on non-instructional costs in the Chicago public school system (Hess, 1995). Similarly, Texas adopted legislation in 1993 to limit administrative expenditures to 11 percent of what is spend on instruction for large school districts, and to 35 percent for small school districts (Dee, 2005 and Lewis, 1994). During 2004, the State of New Jersey adopted legislation that anchors school districts per-pupil administrative costs to the lower of: (1) the prior year's administrative costs adjusted by the cost of living or 2.5 percent, whichever is less, or (2) per-pupil administrative costs for the district's region (State of New Jersey, 2004). At the Federal level, the General Accounting Office issued a report that examined the proportion of Title I funds spent on administrative activities relative to instructional and other activities (GAO, 2003).2 More recently, on February 7, 2011, Governor Christopher Christie revised the Administrative Code of the Department of Education (NJDOE, 2011, chap. 23A), placing a limit on the salaries of New Jersey's School District Superintendents. The “maximum salary” cap was imposed on a sliding scale, ranging from $125,000 for the smallest school districts (less than 750 students) to $175,000 for the largest (more than 6500 students).3 According to the New Jersey Association of School Administrators, the salary cap triggered an exodus of Superintendents to neighboring states (Calefati & Terruso, 2011). Angry school districts claim the salary cap places them at a competitive disadvantage in attracting top administrative talent, and have challenged the law in court or otherwise defied it (Davis, 2011). Thus, there is a perception that administrators will arbitrarily increase discretionary and unproductive spending at the expense of more productive spending that contribute to the quality of education. This perception is supported by Brewer (1996) who theorized that the “administrative blob” had a significant negative relationship to student achievement. We argue that, in the current environment, where so much attention is being paid to student achievements and the administrative cost burden posed on the school system, an exhibition of strong expense preference behavior by administrators implies persistent short-term administrative myopia. We theorize that the employment of first-rate superintendents and principals (and other professionals) at what may be regarded as above-market labor rates can lead to organizational and strategic changes that can significantly improve the school district's performance. Our second research question is focused on the influence of funding from local property taxes on test scores. Nationally, the bulk of education funding comes from three sources: federal (8.3 percent), state (45.6 percent) and local (37.1 percent) (U.S. Department of Education, 2005, p. 2). Although the total education funding has increased over time, it is not clear if the increase in funds can be linked to better quality of education.4 This paper investigates the question of whether all dollars contribute the same or can different sources of funding lead to better outcomes in public schools. Hoxby, 1997 and Hoxby, 1999 has theorized that the higher the proportion of the total spending that was raised locally (as opposed to being provided by the state government, the federal government, or from private funding sources), the greater the pressure upon local school districts to provide high-quality efficient schools and to increase student performance. This theory is supported by Fiva and Rønning (2008), who argue that in the education sector where output is not well defined and monitoring is difficult, property taxation can establish a direct relationship between the tax level and benefits received. Jimenez and Paqueo (1996) and James, King, and Suryadi (1996) report findings (from the Philippines and Indonesia, respectively) which suggest that schools that are more financially accountable to the communities they serve are more responsive to their clients—students and their parents. Empirical support is also provided by Hoxby (1997), who compares educational outcomes for the State of New Hampshire, which relies heavily on local funding and experienced no increase in state aid, to outcomes in six states that increased the proportion of state aid over the 1980–1990 period. According to U.S. Census data for fiscal year 2006–2007, 55 percent of school district revenue in New Jersey was raised locally, placing New Jersey in the top twelve percent of the fifty states in local contributions (Public Elementary-Secondary Education Finance Data—Table 5). Due to the high percentage of school district expenditures funded locally, as well as the variation in local taxes between districts, New Jersey provides an excellent backdrop to test our theory.5 To examine the influence of school district funding sources and the administrative function on student achievement scores, we estimate one-way fixed effects (time-wise), two-way fixed effects (time and school district), and two-way random effects models. Our production function includes a set of variables known or posited to influence student performance, grouped into: socio-economic factors, school district factors outside the control of school district administrators, and school district factors subject to administrative control. The dependent variables are a weighted test score based on the results of three standardized tests taken in the K-12 school districts at the Grade 4 level (NJ ASK4), the Grade 7 level (GEPA, now known as NJ ASK7), and the Grade 12 level (HSPA). Our data consists of 217 K-12 school districts in the state of New Jersey for the years 2002–2009.6 Our principal findings are that, even when the time-invariant socio-economic factors are included in the one-way model with time fixed effects, the proportion of operating expenditures provided by local property taxes is consistently related to test score performance. This finding is robust to when school district fixed effects are introduced, and also extends to our random effects panel model. Results on administrative salaries and total administrative expenditures are mixed. In our one-way time fixed effects model, administrative salaries are positively related to test score performance (both with and without weights), and total administrative spending per pupil is negatively related, as predicted. However, these finding are not robust to our alternative model specifications. Indeed, using dynamic panel estimation with a two-step GMM model, administrative spending per pupil and the share of school funding from local sources are the only statistically significant variables (both with positive coefficients) in explaining test scores. The rest of the paper is organized as follows. Section 2 reviews the literature on the relation between educational expenditures (including the comparative effects of administrative verses instructional spending) and school district performance as well as the effects of funding sources on observed performance. In Section 3, we present our conceptual framework and statistical models. In Section 4 we discuss our data source and provide descriptive statistics on the variables. Our detailed results are presented in Section 5, followed by a summary and conclusions in Section 6.
نتیجه گیری انگلیسی
We set out in this study to examine whether that the proportion of school district expenditures funded by local taxes is positively associated with student test scores. We also wanted to examine whether policies designed to curb administrative expenses could, essentially, be at cross-purposes with the intent to improve educational outcomes as measured by test scores. In general, across all the different panel analysis models we estimated, it is evident that administrative salaries did not have a consistent influence on test outcomes. In the two-way time and school fixed effects model, total administrative salaries and administrative salaries per pupil were found to be not statistically significant. However, using a dynamic panel estimator with GMM, the percentage change in administrative spending per pupil was found to positively (although rather weakly) associated with changes in test scores. An unexpected finding in this study was that the student–faculty ratio was positively related to school district test performance in all three panel analyses. Results of the panel analysis also found that instructional expenditures per-pupil were positively related to student achievement scores. On the other issue of importance in this study, we found much less ambiguous results. The portion of school district operating expenditures obtained from local property taxes was found to be positively related to school district test performance using the one-way time fixed effects panel analysis, the two-way fixed effects, and the random effects panel models. The same positive result was also found in the dynamic panel estimation using a two-way GMM model. This robustness indicates that the results are quite consistent and real, and it also supports the findings reported elsewhere (including Hoxby, 1997). The result suggests that dependence of funding on local sources does exert a disciplinary effect on school district administrators, just as has been theorized in the literature. To our knowledge, this paper is the first one to directly examine at the school district level within a specific state the relationship between student performance on standardized tests and local source of school funding. In light of the increased frequency and utilization of standardized tests, this study has an important implication for policy-makers to consider in their efforts to reform public education. Our findings with respect to the relationship between the proportion of district expenditures obtained from local sources and school district performance suggest that attempts to displace local funding in favor of state or federal funding may be detrimental to improvements in educational achievements, unless some means can be found to make administrators more responsive to local taxpayers. Because local taxpayers are directly affected by school district spending decisions when they pay the highest portion of the expenditures, there is apparently a level of accountability that may not be present if the funding is directly from the state or federal governments. So in public policy deliberations on educational reform, some consideration should be given to mechanisms to retain this local level of interest in the effective and efficient operation of the school districts. Kenyon's (2007) observations on this subject merits attention in light of our findings. First, she recommends against a policy of reforming school finance while ignoring taxpayer concerns. Specifically, states need to consider both the issue of disparities in per-pupil educational expenditure, as well as disparities in the tax burden from an excessive reliance on property taxes to fund public education. Instead, she recommends that state aid be directed at alleviating situations where a taxpayer's property tax burden is high relative to household income. Second, she recommends that federal and state school aid be targeted to specific low-performing schools and students in those districts, citing research that finds that state aid granted to large school districts often wind up being used to subside the higher salaries of teachers and administrators in the more affluent schools within the district. Thirdly, she recommends against states adopting a policy of providing a particular percentage of funding for all school districts because that simply enables rich school districts to divert resources to other spending unrelated to education. Our findings here provide support for the initiation of polices that preserve the incentive for local taxpayers to monitor the performance of their school districts. As such, they are consistent with Kenyon's recommendations.