Running Head: CHARTER SCHOOL SPENDING AND SAVING IN CALIFORNIA. Charter School Spending and Saving in California WORKING DRAFT.

Transcription

1 Running Head: CHARTER SCHOOL SPENDING AND SAVING IN CALIFORNIA Charter School Spending and Saving in California WORKING DRAFT February 2, 2015 Sherrie Reed PhD Candidate, University of California, Davis Director of Research, New Tech Network Heather Rose, PhD Associate Professor, University of California, Davis

2 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 2 Abstract Examining resource allocation practices, including savings, of charter schools is critical to understanding their financial viability and sustainability. Using nine years of finance data from California, we find charter schools spend less on instruction and pupil support services than traditional public schools. The lower spending on instruction and pupil support is offset, not by administrative costs, but by higher spending on operations, consultant services and a greater rate of saving. Spending on consultant services by charters may suggest efficiency in the early years of operation. Further, the greater saving rate of charter schools implies preparation for economic uncertainty, vital for financial viability.

3 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 3 1. Introduction to Charter School Spending and Saving Charter schools represent the primary vehicle for school choice in the U.S. Yet, the long-term viability of charter schools is unknown. While nearly 6,000 charter schools are currently operating across the nation (National Alliance for Public Charter Schools, 2013a), the Center for Education Reform reports that over 1,000 charter schools have closed since 1992 (Consoletti, 2011). The closure of charter schools is natural part of the market in which charter schools operate, where charter schools not meeting consumer demands are driven out of business (Huerta & Zuckerman, 2009). The Center for Education Reform describes five primary reasons for charter school closure including financial deficiencies, mismanagement, poor academic performance, insufficient facilities and the absence of a good working relationship with the charter authorizing agency. The most prominent reason--contributing to 42% of charter school closures--is financial deficiencies (Allen, Consoletti, & Kerwin, 2009; Consoletti, 2011), which may stem from inadequate revenue or ineffective resource allocation. The fiscal challenges of charter schools are a concern for multiple reasons. The charter school movement has long suggested that school choice provides quality schooling opportunities for students who might not otherwise have access (Resmovits, 2014). If quality charter schools are closing merely due to financial problems, it is critical to understand the funding and spending patterns that may contribute to the financial viability of schools. In addition, advocates of charter schools often suggest charter schools are cost effective, delivering higher performance at a lower cost (Resmovits, 2014; Winters, 2012). If this is the case, understanding how charter schools achieve this cost efficiency has important implications for education finance policy more broadly. Finally, the federal government, many state governments and numerous philanthropic organizations invest substantial sums of money to promote the successful implementation of charter schools. This investment of taxpayer and philanthropic money is wasteful if charter schools are unable to remain fiscally solvent and operational. Current research on charter school finance is inconclusive and often conflicting. Much of the research in the field of charter school finance focuses on the adequacy and equity of charter school funding, concluding that charter schools receive less public funding than traditional public schools

4 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 4 (Batdorff, Maloney, May, Speakman, Wolf & Cheng, 2014; Consoletti, 2011; Nelson, Muir & Drown, 2003; Speakman & Hassel, 2005). In contrast, other research indicates that charter schools receive greater funding than traditional public schools (TPS) when philanthropic funding is included (Baker, Libby & Wiley, 2012). However, researchers Gary Miron and Luis Huerta argue that the study of revenues alone is misleading because it does not account for the types of students enrolled and services delivered (in Resmovits, 2014). Examining charter school spending, then, provides a more complete picture of the fiscal practices that lead to efficient and sustainable operations by illuminating how limited funds are apportioned rather than simply looking at the total amount available for spending. The research about charter school resource allocation to-date provides some insight as to how charter school resource allocation differs from that of traditional public schools. On average, charter schools receive less revenue, spend less overall (Arsen & Ni, 2012; Miron & Urschel, 2010, Perez et al., 2006; Zimmer et al., 2003) and save more than traditional public schools (Arsen & Ni, 2012), providing evidence of efficiency and financial viability. However, this lower overall spending manifests in lower spending on instruction and student support services (Arsen & Ni, 2012; Miron & Urschel, 2010, Perez et al., 2006; Zimmer et al., 2003), and higher spending on administration (Arsen & Ni, 2012; Miron & Urschel, 2010, Perez et al., 2006; Zimmer et al., 2003), practices which may undermine consumer demand for charter schools if families believe fewer funds are directed towards the classroom. The existing research of charter spending, however, contains numerous limitations. Todate one national study of charter school spending exists (Miron & Urschel, 2010), but differences in funding mechanisms, reporting requirements and policy contexts across states limit the conclusions that can be drawn. This national study includes only eight of the nearly 1,000 charter schools in California, the state with the largest share of the nation s charter schools, and lacks any data for charter schools in Florida, the state with the second largest share of the nation s charter schools. Two studies which examine charter school resource allocation practices in California (Perez, et al., 2006; Zimmer et al., 2003) use only personnel staffing data and survey data about the perception of fiscal flexibility and

5 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 5 autonomy, possibly missing important findings for spending on equipment, materials, and facilities. Also, all of the existing research uses a single year of data, where shifts in the number of charter start-ups or a charter policy may potentially skew findings. Further, little research accounts for the length of operating history of charter schools, despite evidence that it matters (Zimmer et al., 2003), as charter schools are often plagued with heavy start-up costs in the first few years (Hayes & Keller, 2009). Finally, only limited research examines charter school savings patterns, important to paying off short-term debts and surviving revenue deficits, common in times of economic uncertainty (Wohlstetter, Smith & Farrell, 2013). This study adds to the existing body of research on the resource allocation practices of charter schools by capitalizing on publicly available data for California charter schools. California represents an important segment of national charter school community, as California is often cited as a model for charter school legislation and equitable funding (NAPCS, 2013b). California charter schools account for over 18% of all charter schools and 21% of all charter school students in the U.S. (NAPCS, 2013a). California also leads the nation in the number of charter schools in a single state. California s 1,000 charter schools make up ten percent of public schools in the state and enroll seven percent of the state s K-12 students. 1 Since , California charter schools have reported financial data, providing nine years of data to analyze. We use these multiple years of detailed charter school financial records, combined with relevant information on the duration and location of charter school operation, as well as the characteristics of enrolled students, to address the gaps in the literature and answer each of the following research questions: 1. How do charter schools in California allocate resources (i.e. instruction, administration, pupil support, operations and savings)? 2. How does resource allocation differ between charter schools and traditional public schools? 1 1 Numbers representing California charter schools and student enrollment are calculated based on administrative data from California Department of Education.

6 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 6 3. How does resource allocation differ between newly founded charter schools and charter schools with a long operating history? 4. How do school characteristics (i.e. total enrollment, percent of students eligible for free and reduced lunch and geographic location) explain noted differences in resource allocation between charter schools and traditional public schools? 2. Policy Context Since their advent in Minnesota 1991, charter schools continue to grow, both in numbers of charter schools operating and numbers of students served. In , over 5,500 charter schools operated in 39 states and two territories, accounting for nearly six percent of the nation s public schools and four percent of K-12 public school students (NAPCS, 2013a). Recent legislative action to expand charter school operations in several states indicates the number of charter schools in the U.S. will continue to grow. In 2012, two additional states, Washington and Mississippi, approved charter school laws (NAPCS, 2013). Other states, such as Hawaii, Idaho and Missouri, lifted the caps on the numbers of charter schools allowed to operate within the state (NAPCS, 2013). Further, the federal government provides financial incentives for states to improve charter laws and increase the number of charter schools through Race to the Top grants and the National Charter School Grant Program. Noteworthy philanthropists like Reed Hastings and Walton Family Foundation, and advocacy foundations such as Thomas B. Fordham Institute and National Alliance of Public Charter Schools, also lend their influential support to ensure the growth of charter schools. The growth in the number of charter schools provides educational options for an increasing number of students. In , charter schools in the United States enrolled nearly two million students (NAPCS, 2013a). The National Alliance for Public Charter Schools estimates that another one million students across the nation are on charter school wait lists, providing evidence of the demand for more charter schools (Kern & Gebru, 2014). 2.1 California Charter Schools

7 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 7 Since the passage of the Charter School Act of 1992, California has been a leader in the charter school movement. California was the second state to enact charter school legislation and its present-day charter school legislation is a model for other states. The National Alliance for Public Charter Schools (NAPCS, 2013b) points to California as a leader in the areas of equitable operational funding, equal access to categorical funding, and equitable access to capital funding and facilities. Additionally, the Charter Schools Development Center and the California Charter Schools Association, California s charter school advocacy organizations both headquartered in Sacramento, are important leaders in shaping both state and national policy for charter schools. In order to operate as a public school in California, a charter school must be authorized by a public school district governing board, a county office of education governing board or the State Board of Education. Charter schools are initially authorized on the merits of a charter application outlining operating principles and measurable outcomes for which the school will be accountable. Typically, authorizers grant operating permission for five years, with a renewal application and review process necessary to continue operation for a subsequent five years. During the initial five year term, newly founded charter schools often experience non-recurring costs associated with initial planning and getting the school up and running (Hayes & Keller, 2009). The charter renewal process varies by charter authorizing agency, but the authorizer is responsible for oversight of the charter school for adherence to local, state and federal education regulations, including financial regulations. It is presumed therefore that, charter schools demonstrate financial stability and document success on measurable outcomes in order to earn the renewal for continued operation. Charter schools operating for six to ten years are often in their second term and those operating for more than 10 years have typically passed two renewal processes. 2.2 Charter School Funding California Education Code (1999) defines the sources and means for funding charter schools. Like all public schools in California, general purpose funding for charter schools is based

8 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 8 on the average daily attendance (ADA) rate of enrolled students. 2 Charter schools receive a weighted per- ADA rate that varies by grade levels and needs of enrolled students. Charter schools receive an amount equal to the statewide per pupil funding average for elementary districts for each student in grades kindergarten through five, an amount equal to the average for unified (K-12) districts for students in grades six through eight, and a rate equal to the average for high school districts for students in grades nine through 12. Charter schools also receive Economic Impact Aid, which provides additional per-ada funding for low-income and English language learners. During the time period examined in this study, charter schools in California also received categorical funding. In lieu of nearly 50 categorical programs available to traditional public schools, charter schools receive per pupil funding through the Charter School Block Grant. In addition, charter schools receive a share of state special education funds and are eligible to receive funds from a few large state categorical programs such as class size reduction and transportation. If eligible, charter schools also receive federal categorical funding provided through the No Child Left Behind Act (2001) for educational programs for low-income students or English language learners, professional development for teachers, and the drug-free and safe schools initiatives (CDE, Charter School Categorical Block Grant Programs, nd). Public funding for charter schools is distributed in one of two ways in California. In direct funded charter schools, the state disburses charter school funding directly to charter schools. Direct funded charter schools manage their own fiscal activities independently of their charter authorizing agency (CDE, Charter School Direct Funding Option, nd.); all spending decisions are made in accordance with charter school governing board by-laws and operating procedures. Direct funded charter schools usually have financial personnel on staff or contract with private accounting firms to manage accounts receivable, 2 Charter school funding in California is straightforward compared to the funding of traditional public schools. Until funding for traditional public schools varied throughout the state based on a complex formula of local and state sources. For more information on California s public school funding formula see Weston, Sonstelie and Rose (2009).

9 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 9 payable and personnel payroll. The percent of charter schools that are direct funded has steadily increased from 61% in to 72% in On the other hand, between 28% and 39% percent of charter schools in California operate as dependent charter schools, where funding and operations are closely tied with the authorizing district. In dependent charter schools, also known as locally funded charter schools, funding flows from the state to the local education agency that authorized the charter, most typically a public school district. In the case of locally funded charter schools, the charter authorizing agency often manages part or all of the fiscal activities of the charter school and retains fees for these services as agreed upon between authorizer and charter school operator. Remaining funds are then disbursed from the charter authorizing agency to the charter school. The local funding of dependent charter schools makes it difficult to disentangle charter school fiscal information from that of the charter authorizing agency, a common data limitation noted in prior literature (Miron & Urschel, 2010; Zimmer et al., 2003). 2.3 Charter School Financial Reporting In California, fiscal reporting was first required for charter schools in , though traditional public schools reported fiscal data for decades. Currently, all public schools, including charter and traditional public schools, report financial information to the California Department of Education (CDE) as required annually. Traditional public school districts report financial data using the Standardized Account Code Structure (SACS). Charter schools may be included with the charter authorizing district s financial reporting in SACS or charter schools may report finance data independently of their authorizer using either SACS or the Charter School Alternative Reporting Form (also known as the Alternative Form ), depending on agreements with the authorizing agency. Almost all locally funded charter schools and one-third of the directly funded charter schools report financial data with their authorizing agency in SACS each year. Consequently, the charter schools that report separately from their authorizer, on either the SACS or Alternative Form, are the remaining two-thirds of direct funded charter schools and include almost no locally funded charter schools.

10 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A Data and Methods This study draws on the financial data reported annually by all traditional public school districts and charter schools in California, which is publicly available from CDE. We include nine years of finance data, beginning with , the first year in which charter schools were mandated to report financial data, through , the most recently available data at the start of this analysis. The use of nine years of data addresses the limitation of cross-sectional data and single year studies present in the existing literature and allows me to examine whether or not patterns persist over time as individual charter schools mature and additional charter schools enter the market. 3.1 Sample The sample of charter schools for this study includes those charter schools reporting both schoollevel financial data separate from their authorizing agency, just under half of all charter schools. Table 1 provides detailed information about the number of the charter schools operating in the state, the charter schools excluded, and the percent of charter students served, for each of the nine years included in this analysis. We do not have school-level finance data for the 48% to 57% of charter schools that report their financial data with their authorizer in any given year. For example, in , 487 of the 991 charter schools reported with their authorizing agency. By necessity, these charter schools are excluded from the charter school sample. We also exclude the 3% of charter schools that fail to report any finance data. 3 Lastly, a handful of charters are excluded because they are missing student enrollment data; in , only one school met this criteria, but in , eight schools are excluded for this reason. 4 After exclusions, between 38% and 48% of operating charter schools in any given year are included in the sample. These charter schools are almost exclusively direct funded charter schools and serve nearly half of the charter school students in the state. Though slightly less than half of California charter schools are 3 The reasons that charter schools fail to report finance data is unknown. According to CDE, these schools are expected to report finance data at the school level, but no data appears in either the SACS or Alternative Form data files. It is possible that these charter schools actually report with their charter authorizing agency instead. Review of the list of schools that fail to report in each year demonstrates that the problem is limited to few schools who fail to report in multiple years, of which between 50% and 80% are direct funded. 4 The reason charter schools fail to report enrollment data is also unknown. However, review of the schools who fail to report shows the problem is again limited to a few schools who fail to report in multiple years.

11 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 11 included, the final sample is more robust than other studies of charter school spending. Unlike previous studies, the final sample includes multiple years of finance data. Further, the sample includes substantially more charter schools than other single state studies, with Arsen and Ni (2012) examining just over 250 charter schools in Michigan and our study including over 450 in alone. The comparison group in this paper includes all traditional public schools in California, 5 whose financial data are reported at the school district level. 6 Because of district-level reporting, these data necessarily include the locally funded and direct funded charter schools electing to report with their authorizing district. The inclusion of these charter schools in the comparison sample is unlikely to substantially affect the aggregate values computed for traditional public schools, because the charter schools represent only three to five percent of California s public schools and two and four percent of California s total K-12 enrollment. 5 We exclude schools operated by County Offices of Education (COEs). COES administer professional support to local education agencies within their boundaries and operate special programs for students with special needs or adjudicated youth. While spending on programs serving students are included in COE finance data, so are expenses unique to the administrative and professional development services offered. As such, the finance data does not accurately represent average per pupil spending and is therefore excluded from analysis. This exclusion follows the methods of Rose et al. (2003). 6 Common Administrative Districts serve the district administrative needs of multiple school districts and report financial data in aggregate. In Common Administrative Districts, school districts pool resources and share costs for services such as human resources, federal and state reporting and financial services. For the purpose of analysis, we treat them as a single unified school district. In , there were six Common Administrative Districts.

12 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 12 Table 1 Charter School Finance Data Sample Total Charter Schools Total Charter Students 8 166, , , , , , , , ,856 Exclusions Charters Reporting with District/COE/SBE Charters Not Reporting Finance Data Charters Not Reporting Student Enrollment Sample Charter Schools in Sample Charter Students in Sample 69,387 76,951 93, , , , , , ,470 % of CA Charter Schools % of CA charter students In this table, the number of charter schools includes those charters which are listed in CDE finance data for whom operating status was verified using the Public Schools Database also from CDE. In financial datasets, more charter schools are counted, but often prior to opening or after closing. For example in the financial data files included 512 charter schools with more than 50 not reporting finance data. However, only 460 were actually operating during the school years, bringing the number of charters failing to report data to 14. By , all 991 charter schools included in the financial data were legally operating. 8 In this table, the number of students enrolled in charter schools is calculated from CBEDS student enrollment files downloaded from California Department of Education.

13 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A Categorizing Expenses We use SACS and Alternative Form data to compare spending patterns of charter schools and traditional public school districts in California. The financial data from both formats align considerably well and include detailed accounts, using over 100 object codes. Object codes describe and categorize the source of revenue (e.g., general revenue, property tax income and grants) the items purchased (e.g., instructional materials and computer equipment), services purchased (e.g., legal, professional development and consulting), and staff hired (e.g. teachers and administrators). 9 We adopt methods previously designed by Rose, Sonstelie, Reinhard, and Heng (2003) for the study of resource allocation in traditional public schools in California. In their study, Rose et al. (2003) created eleven expense categories based on the purpose of expenditures: teachers, instructional aides, administrative personnel, pupil service personnel, other personnel, instructional materials, maintenance and operations, other services, capital equipment, food service, and tuition and transfers. In the analysis, we collapse the 11 categories identified by Rose et al. (2003) into four larger groups--instruction, student support, administration, and operations--to align with the existing literature about charter school spending. Instruction includes teacher and instructional aide salaries and benefits, as well as instructional materials. Pupil support services include guidance, psychological, and health services and special education related services. Administration is limited to the salaries and benefits for certified administrators serving at the school and district level. The operations category includes salaries and benefits for clerical and janitorial personnel, the acquisition, rental and/or maintenance of facilities, and the acquisition and maintenance of capital equipment such as computers. We also isolate two additional categories: consulting services and savings. Hypothesizing that charter schools may be purchasing administrative and pupil support services from consultants, separate out the consultant expenses from other services. Finally, we generate annual savings, which is simply total revenue minus total expenses in 9 Both the SACS and Alternative Form utilize the same accounting object codes despite different reporting formats. The consistency in codes in both reporting formats allows for accurate comparisons and valid analysis when categorizing expenses. A detailed crosswalk of the codes from both the SACS and Alternative Form as used in this analysis is available from the author.

14 Percent of total enrollment Per pupil dollars C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 14 the other five categories. Annual savings is essential to understanding charters ability to save for economic uncertainty and long-term financial viability. In total, we include six main expense categories. 3.3 Method of Analysis In order to examine differences in resource allocation between charter schools and traditional public schools, we begin by comparing mean per-pupil spending in each expense category. These descriptive statistics show notable and statistically significant differences between charter schools and traditional public schools in a variety of spending categories. We also compare mean per-pupil spending between charter schools of various lengths of operating history, with few noteworthy differences. Recognizing the numerous possible explanations for the observed mean differences between charter schools and traditional public school districts, we also estimate a series of regressions to better understand the extent to which various factors may explain the difference in resource allocation. For each of our six expense categories (total expenditures, instruction, pupil support, administration, operations, consulting and savings), we estimate a regression in which the dependent variable is per pupil spending in that category. We use the same set of independent variables in each of the models. Table 2 provides descriptive statistics for each of the dependent and independent variables. Table 2 Descriptive Statistics for Regression Variables Charter Traditional Public Schools 25 th p Mean 75 th p 25 th p Mean 75 th p Dependent Variables (weighted by total enrollment) Instruction 3,506 4,140 4,607 5,029 5,438 5,743 Pupil Support ,120 1,214 Administration , ,124 1,243 Consulting 462 1,015 1, Operations 709 1,269 1, Savings Control Variables Middle school students High school students Eligible for free/reduced price meals Total enrollment ,294 6,257 Per pupil revenue (weighted) 7,071 8,171 8,773 8,065 9,044 9,795

15 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 15 The first independent variable is a dichotomous variable indicating whether the entity is a charter school or traditional public school district. The coefficient on this variable indicates the difference in spending between charter schools and traditional public school districts. 10 The second set of variables account for the characteristics of students a school serves. Prior literature suggests that charter schools are more likely to serve elementary students, who are presumably less costly to serve than secondary students (Arsen, Plank & Sykes, 1999; Miron & Urschel, 2010). In fact, the funding mechanism in California provides charter schools with less revenue for elementary students than those in higher grades. Furthermore, differences in grade levels served may also dictate differences in spending needs across categories; for example high schools may need more sophisticated science labs and more guidance counselors than elementary schools. Therefore, the regression models include control variables that account for the percent of students enrolled in elementary, middle and secondary grades, data collected annually by California Department of Education and later made publicly available. 11 Prior research also suggests that charter schools are less likely to serve high needs students (Miron & Urschel, 2010; Perez et al., 2006; Resmovits, 2014; Zimmer et al., 2003), such as those who are eligible for free and reduced price meals, who are presumably more costly to serve. This difference may explain overall differences in total revenue, as schools receive additional funding through federal and state categorical programs for students who are eligible for free and reduced lunch or special education, as well as spending in specific categories such as pupil support staff. To account for these factors, our regression models include the percent of students eligible for free and reduced lunch. These data are also publicly available from California Department of Education. 12 The total enrollment in a school may also influence spending. Prior research points out the difference in per pupil spending may be related to economies of scale, where the cost to educate each 10 With no additional controls in the model, the coefficient on the charter variable simply captures the mean difference in per-pupil spending

16 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 16 student decreases as the number of students increases (Andrews, Duncombe & Yinger, 2002; Duncombe, Miner, & Ruggiero, 1995; Monk & Hussain, 2000; Pandolfo, 2012). As fixed costs are spread across more students, the per-pupil cost of resources actually declines. Guided by this theory, charter schools which are typically smaller in size than traditional public schools are expected to spend more per pupil than traditional public schools in areas such as administration, which have noticeable fixed costs. Therefore, the regressions include a control for a school s total enrollment. This control enters the regression in log form to account for potential nonlinear relationship between size and spending. Although student characteristics and enrollment are important factors in determining spending patterns, ultimately the total revenue per pupil determines the funds available for spending in the various categories, therefore the regressions also account for total revenue per pupil. Another consideration when noting the spending differences between charter schools and traditional public schools is geographic location. Charter schools are often concentrated in urban areas (Ritter, Jensen, Kisida & McGee, 2010), where both salaries and property costs are higher. In California, 214 charter schools, or nearly one-quarter of California s charter schools, are located in Los Angeles (CCSA, California Charter Schools by the Numbers, nd). These Los Angeles charters enroll the largest number of students in any single school district in the country (NAPCS, 2013). Similarly, there are 55 charter schools operating in San Diego, 38 in Oakland, and 36 in San Jose (CCSA, California Charter Schools by the Numbers, nd). Furthermore, previous literature provides evidence of regional wage differences, where salaries are notably higher in urban areas such as Los Angeles, San Diego and San Jose (Rose, Sengupta, Sonstelie & Reinhard, 2008; Rose & Sengupta, 2007). Therefore, if charter schools are disproportionately located in these areas, we may observe higher spending in categories that are wage intensive. We use county level fixed effects to control for regional differences, ensuring that any notable differences between charter schools and traditional public schools are not attributable solely to the concentration of charter schools in certain geographic locations. To account for the fact that multiple districts are located within a county and thus may have correlated error terms, we assume that the

17 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 17 error term in our model has a county-specific component and compute standard errors clustered at the county level. Finally, when we estimate the regressions, we include nine years of finance data. To ensure that time trends are not driving our results, we also include fixed effects for the year. The regression analysis substantiates the findings presented in the descriptive tables and also allows us to show what portion of the charter schools spending differences are attributable to these additional factors. 4. Findings and Discussion After briefly reviewing overall spending and revenue trends, this section will evaluate spending by category. We outline the similarities and differences in the spending patterns of traditional public schools and charter schools. We also explain differences between charter schools operating for varying lengths of time, where findings are notable. Throughout the section, we suggest possible explanations for our findings and consider how our findings align with or diverge from those in previous studies of charter school resource allocation. Summarizing nine years of data, Table 3 displays overall trends in revenue and expenses for charter schools and traditional public schools in California. Columns A and B show the spending in each category as a share of total expenditures for charter schools and traditional public schools. Columns C and D provide the average per-pupil revenue and the average per-pupil spending in each spending category for both charter schools and traditional public schools. Column C shows the differences in spending between charter schools and traditional public schools over this near decade, and column D shows the regression-adjusted differences. Although Table 3 gives the nine-year averages, the discussion below includes figures depicting the annual data and trends overtime. We convert the data for individual years to dollars using the all-urban consumers Consumer Price Index (CPI). Furthermore, Appendix Tables A.1 and A.2 provide spending by category for each of the nine years and Appendix Table A.3 shows the detailed regression results used to determine the regression-adjusted differences.

18 Operations Pupil Support Instruction C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 18 Table 3 Differences in Revenue and Spending between Charter Schools and Traditional Public Schools ( dollars) A B C D E F Mean Percent of Total Current Expenses Mean Per Pupil Spending Mean Difference Regression Coefficient with Controls Charter TPS Charter TPS Charter TPS Charter TPS Total Revenue NA NA 8,171 9, * NA Total Expenditures NA NA 7,840 8,885-1,045* -129* Instruction ,140 5,438-1,298* -618* Teachers ,205 4,639-1,434* -928* Instructional aides * -43 Instructional materials * 353* Pupil Support , * -626* Pupil service personnel * -481* Tuition and transfers Administrative personnel , * -166* Consulting services , * 613* Operations , * 669* Maintenance and operations * 453* Other personnel * 28 Other services * 106* Capital equipment * 83* Annual Savings * 129* *p<.05

19 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A Total Expenditures and Revenue Researchers agree that charter schools spend less than traditional public schools (Arsen & Ni, 2012; Miron & Urschel, 2010; Perez et al., 2006). Our analysis also supports this conclusion. In California, charter schools spend less than traditional public schools in every year since charters began reporting financial information in On average, charter schools across the U.S. spend 19% less than traditional public schools (Miron & Urschel, 2010), while in California the magnitude of the difference is not as great. Across the nine years examined, charter schools spent an average of $7,840 per pupil, about $1,000 per pupil less than traditional public schools. Figure 1 displays the average total per pupil spending for charter schools and traditional public schools across the nine years. The difference is largest in , when charters spend $1,352 less than traditional public schools, and smallest in , when charters spend only $774 less. While the differences vary slightly from year to year, the trend of lower spending is consistent, with charter schools spending an average of 12% less than traditional public schools. The lower spending by charter schools may in part be due to the fact that they receive less revenue, though the entire 12% gap in spending is not explained by lower revenue. In California, charter schools receive an average of 10% less per pupil in state and federal revenue than traditional public school districts, with the gap as small as 6% in and as great as 14.5% in In addition per pupil total spending, Figure 1 depicts the trends for annual per pupil revenue for both charter schools and public school districts from until

20 Per Pupil Dollars C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 20 Figure 1 Revenue and Expenditures for Charter Schools and Traditional Public Schools ( dollars) $10,000 $8,000 $6, Revenue--TPS Expenditures--TPS Revenue--Charter Expenditures--Charter While some might hypothesize that long-term charter schools are more likely to mirror traditional public schools in spending patterns than new charter schools (Arsen & Ni, 2012), the greatest difference in overall spending is between long-term charter schools, rather than newly-founded charters, and traditional public schools. Newly founded charter schools actually spend more per pupil on all expenses combined than long-term charter schools. On average, long-term charter schools spend 16% less than traditional public schools, or about $1,400 per pupil less. In contrast, newly founded charter schools spend only 7%, or $600 per pupil, less than traditional public schools. In other words, newly founded charter schools spend an average of $800 more per pupil than long-term charters. Though this gap is notable, the magnitude of the difference fluctuates markedly over the decade as seen in Figure 2. In , newly founded charter schools spent only $276 per pupil more than long-term charters, but in the difference peaked at over $1,200 per pupil. Despite the fluctuations, the higher spending by newly founded charter schools compared to those operating for a much greater length of time persists in every year in the past decade.

21 Per Pupil Dollars C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 21 Figure 2 Revenue and Expenditures for Newly Founded and Long-Term Charter School ( dollars) $10,000 $8,000 $6, Revenue--New Charters Expenditures--New Charters Revenue--Long-term Charters Expenditures --Long-Term Charters As expected, the overall spending patterns reflect the overall revenue patterns. Newly founded charter schools receive an average of $764 more per pupil in funding than long-term charter schools, allowing for the greater spending. This difference is possibly attributed to grant funding available for start-up charters. In California, charter schools may apply for up to $575,000 in grant funding for the planning and implementation of a new charter school through the Public Charter Schools Grant Program (CDE, Public Charter Schools Grant Program, nd.). 13 A review of revenue supports this notion, as the revenue differences between newly founded and long-term charter schools is evident only as restricted revenue. Restricted revenue includes funds distributed for specific purposes and where spending of the funds is restricted to items or services that support those specific purposes. Unrestricted revenue is general purpose funding that can be spent on any educational function allowable under state law. Table 4 shows three years of revenue data for newly founded charter schools, mid-length charter schools and long-term charter schools. In no years are there statistically significant differences in the unrestricted revenue of charter schools operating for various 13 The Public Charter Schools Grant Program is funded by the federal government, who makes large grants to state education agencies, such as California Department of Education, for expanding the number of high quality charter schools across the Nation (U.S. Department of Education, nd.). CDE then manages the grant application, award, oversight and fiscal obligations for new charter schools seeking the start-up grant funds.

22 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 22 lengths of time. In contrast, in and , as it is in every year across the decade except , newly founded charter schools receive substantially more restricted revenue than mid-length or long-term charter schools (p<.01). Table 4 Per Pupil Revenue for Charter Schools of Varying Lengths of Operations ( dollars) Restricted Unrestricted Restricted* Unrestricted Restricted* Unrestricted New Charters Mid-length Charters Long-term Charters *All significantly different from each other (p<.01). It is important to note, public funding may not tell the entire story of charter school revenue. Researchers suggest charters are more likely than traditional public schools to raise funds through private sources (Baker et al., 2012; Miron & Urschel, 2010; Speakman & Hassel, 2005; Zimmer et al., 2003). Further, education foundations, such as Walton Family Foundation and Bill and Melinda Gates Foundation, have frequently offered grants for new charter schools over the last decade. Revenue from and goods and services purchased with philanthropic funding are often not included in state mandated charter school finance reporting systems as they may be purchased directly, leading to likely underestimation of total charter revenue and spending (Baker & Ferris, 2011; Miron & Urschel, 2010). In this case, philanthropic funding and expenses charged to this revenue source may fill in the gaps in funding and spending or in fact contribute to fiscal advantage of charter schools (Baker & Ferris, 2011; Gittleson, 2010), some of which may be undetected in analyses of charter school finance data based on state department of education reports. 14 Whether or not charters receive philanthropic funding, the lower average public funding and spending of charter schools compared to traditional public schools provides evidence that charter schools may cost taxpayers less to operate, an argument often presented in the case for charter school expansion. 14 Some philanthropic revenue may be included in California finance data as local revenue. However, in progress research from the Public Policy Institute of California provides evidence that most philanthropic funding for schools is not captured in SACS and Alternative Form data (Weston, personal communication, 2014).

23 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 23 However, a lower public cost does not necessarily equate to more efficient or more effective. Further, a lower public cost may not be sustainable if charters are dependent on philanthropic dollars that may wane over time. It is therefore important to understand how lower overall spending by charter schools is accomplished and consider the potential implications of resource allocation. 4.2 Instructional Expenditures Supporters of charter schools argue that the autonomy granted to charter schools leads to more spending on instruction than occurs in traditional public schools (Hill & Roza, 2008). However, our analysis and the body of research on charter school spending do not support this argument. Generally, charter schools spend less on instruction, including salaries and benefits for teachers and instructional support personnel, as well as instructional materials, than traditional public schools. In fact, while charter schools in California spend 12% less on average than traditional public school districts on all expenses, they spend an average of 23% less on instruction related expenses. In California, charter schools spend an average of $4,140 per pupil, or 53% of total current expenditures, on instruction related expenses, whereas traditional public school districts spend an average of $5,438 per pupil, or 61% of total expenses, towards instructional purposes. (Table A.1 shows expenditures as a percent of current expenditures.) This spending gap is generally consistent over the nine years examined, with the largest difference, $1,384, occurring in and the smallest difference, $1,149, occurring in

24 Per Pupil Dollars C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 24 Figure 3 Spending on Instruction for Charter Schools and Traditional Public Districts ( dollars) $6,000 $4,000 $2, Instruction -TPS Instruction -Charter Within the category of instruction, the gap between charter schools and traditional public schools in instructional spending is primarily evident in spending on salaries and benefits for teachers and instructional aides. In fact, the substantial gap in spending on teacher salaries is masked when considering all instructional related expenses, as charters spend between $150 and $300 more per pupil on instructional materials making up for some of the difference. Charters spend an average of $1,434 per pupil less on teacher salaries and benefits than traditional public schools. Further, charters spend $133 less per pupil than public school districts for instructional aides. In general, the spending gap between new charter schools and mature charter schools on total instruction expenses is not significant or substantial; however, the shifts in spending within the category are notable (see Table A.4). Spending on teacher salaries and benefits is higher for long-term charter schools, while spending on instructional materials is lower, indicating that charter schools may shift spending from materials to personnel as the school matures. Charter schools in their first five years of operation spend about $4,200 per pupil on teachers and nearly $850 per pupil on materials, on average. In contrast, long-term charters spend $200 more per pupil on teachers, but $300 per pupil less on materials. Both of these trends appear to persist in all nine years, however differences observed are only statistically significant in some years.

25 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 25 The lower spending by charter schools on teachers and instructional aides, documented both in this study and prior research (Arsen & Ni, 2012; Miron & Urschel, 2010), debunks the myth that autonomy granted to charters leads to greater spending on instruction. This does not mean, however, that fewer personnel are devoted to instruction in charter schools. Prior research suggests this lower spending in instruction is a result of lower teacher salaries. In their study of staffing patterns in California charter schools, Perez et al. (2006) find charter schools generally have the same teacher-pupil ratio as non-charter schools. Charter schools may spend less on teacher salaries while maintaining teacher-pupil ratios similar to traditional public schools due largely to the fact that, on average, charter school teachers have fewer years of experience and therefore are paid less (Perez et al., 2006; Zimmer et al., 2003). Some researchers hypothesize the differences in lower spending on instruction by charter schools is due to the types of students served in charter schools. Specifically, researchers suggest that charter schools are more likely to serve less costly elementary students. Yet, in California, this trend does not explain the difference in instructional spending between charter schools and traditional public schools. In California, charter schools and traditional public schools actually have fairly comparable distributions of students by grade level (see Table 5). If anything, charter schools actually have proportionally more students in the higher grades compared to traditional public schools. Of the total students enrolled in charter schools in California, nearly 40% are students in grades kindergarten through fifth grade, and another almost 40% are enrolled in high school grades, with the remaining 20-25% of students in grades six through eight. Traditional public schools, on the other hand, serve a slightly larger proportion of elementary students (45%) and smaller proportion of high school students (30-32%), a pattern mirrored in both charter and non-charter schools across the U.S. (NAPCS, 2013a). Table 5 Percent of California Student Enrollment by Grade Level Charter TPS Charter TPS Charter TPS Elementary Grades (K-5) Middle Grades (6-8) Secondary Grades (9-12)

26 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 26 Regression analysis, summarized in Table 3 and Table A.3, provides further evidence that the difference between charter schools and traditional public school spending on instruction is not driven entirely by grade level of the students served. When controlling for grade level of students served, and other factors that may influence spending in both charter schools and traditional public schools, charter schools overall spending on instruction remains lower than traditional public schools, by more than $600. This lower spending by charter schools represents about half of the original gap spending between charters and traditional public schools, meaning that grade level and the other factors explain about half, but not all, of the difference. Furthermore, the relationship between grade level and spending works counter to expectations; as the percent of students in grades six through 12 increases, spending per pupil decreases. In other words, relative to a school with all elementary students, a school comprised entirely of middle or high school students spends about $500 less per pupil on instruction. This may be a result of elementary class size reduction efforts across the state. 4.3 Pupil Support Services Expenditures In addition to spending less on expenses directly related to instruction, charter schools spend less on pupil support services (i.e. guidance, psychological, health services and special education related services) as well. Between and , charter schools spend an average of $455 per student, or 6% of their total expenses, on pupil support while traditional public schools spend more than twice as much, $1,120 per pupil, or 13% of their annual total spending. The resulting gap between charter school and traditional public school spending on pupil support services is $665 per pupil, on average, with charter school spending less. This pattern is consistent over nine years, with the widest gap, $771, evident in and the smallest gap in , at $574. Figure 4 Spending on Pupil Support Services for Charter Schools and Traditional Public Districts ( dollars)

27 Per Puil Dollars C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 27 $1,400 $1,000 $600 $ Pupil Support--TPS Pupil Support--Charter The finding that charter schools spend less on pupil support services than traditional public schools aligns with prior research. In their study examining charter spending across the nation, Miron and Urschel (2010) find that in charters spend less than non-charters on student support services, such as attendance, guidance, psychological and health services and special education related services. Arsen and Ni (2012) find charters in Michigan, charters spend $400 less on student support services (e.g. guidance, speech, health and social work) and instructional support personnel (e.g. librarians and curriculum specialists) than traditional public schools), a notable, but lower gap than what we find in California. Unlike expenses related to instruction however, the lower spending by charters on pupil support services may be directly related to the employment of fewer personnel. In fact, according to Perez et.al. (2006), charter schools in California have about one-third fewer pupil support staff per 100 students than traditional public schools. It logically follows that the lower spending and fewer personnel may then be due to less demand, or fewer students in need of specialized support services. Prior research suggests charters may not serve students needing the range of services provided by traditional public schools (Arsen et al., 1999; Arsen & Ni, 2012; Miron & Urschel, 2010). Exempt from many public school regulations, charter schools are often not obligated to provide services such supplementary instruction and summer programs, which noncharter public schools may be legally bound to deliver (Arsen & Ni, 2012; Arsen et al., 1999; Miron & Urschel, 2010). Further, charters often serve fewer special education students and English language

28 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 28 learners deferring the higher costs of pupil support services to traditional public schools, a finding also noted in prior research (Perez et al., 2006; Zimmer et al., 2003). 15 As presented in Table 6, California charter schools serve a student body with a smaller proportion of special education students and English language learners, 7.4% and 15% respectively, than the 10.5% of students eligible for special education and 22% of students that are English language learners in traditional public schools. The narrowed range of services necessary to meet the needs of students enrolled in California charter schools may drive the lower spending on students support services. Table 6 Student Demographics for Charter and Traditional Public Schools ( ) Percent of Students Served 16 In U.S. In California TPS Charters Difference Charter - TPS TPS Charters Difference Charter - TPS Free & Reduced Lunch English Language Learner Special Education White African American Hispanic Asian Other However, regression results indicate that even when controlling for the percent of students eligible for free and reduced price meals, a proxy for students with greater academic needs, the gap between charter schools and traditional public school spending on pupil support services decreases by only $39 per pupil (Table 3). In other words, the student characteristics that we add to the model explain almost none of the difference in spending on support staff. Regression results show that a larger percent 15 While prior research suggests that charters may be enrolling fewer high needs students than traditional public schools, data retrieved from the National Alliance for Public Charter Schools Dashboard (and presented in Table 6) shows that charter schools across the nation serve a greater proportion of students eligible for free and reduced price meals and English Language Learners than traditional public schools. 16 National and California data based on school year retrieved from National Alliance of Public Charter Schools Data Dashboard 17 California data based on API Growth data files from available at National data based on school year (most recent available), retrieved from National Alliance of Public Charter Schools Data Dashboard

29 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 29 of secondary students is associate with higher spending on pupil support services, a finding that aligns with the general staffing patterns in secondary schools which hire more guidance and attendance personnel. In contrast, regression results show a greater percent of students eligible for free and reduced price meals is associated with lower per pupil spending on support services, despite likely increases in revenue received by the school. Collectively, the evidence from regression analysis demonstrates that charter schools and traditional public schools spend differently in the area of pupil support services even when serving similar students in the same geographic location and when receiving equal revenue. 4.4 Administrative Personnel Expenditures One might think that lower spending on instruction is offset by greater spending on administration, as pointed out in prior research (Arsen & Ni, 2012; Miron & Urschel, 2010). Yet, this is not the case for charters in California. Both charter schools and traditional public schools in California spend a similar proportion of their operating budget, 13%, on administrative personnel. Though similar in proportion, the lower average revenue of charter schools means that charter schools actually spend less per pupil. On average, charter schools spend $961 per pupil on administrative personnel, whereas traditional public school districts in California spend about $163 more per pupil. The noted average gap in spending on administration is a result of a steadily increasing difference between charter school and traditional public school spending over the decade, as depicted in Figure 5. In , charter schools spent just $22 less per pupil than traditional public schools on administrative costs and by charters spend $209 per pupil less.

30 Per Pupil Dollars C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 30 Figure 5 Spending on Administration for Charter Schools and Traditional Public Districts ( dollars) $1,400 $1,000 $600 $ Administration--TPS Administration--Charter The trend of lower spending on administration by charter schools in California when compared to traditional public schools contradicts findings from other states and data samples. Prior research consistently reports that charter schools spend more than non-charter schools for administration. Miron and Urschel (2010) find charters across the nation (excluding California) spend an average of $1,402 per pupil on administration, compared to only $960 per pupil by traditional public schools. In Michigan, the difference in spending on administration is even greater, with charter schools spending nearly $900 more per pupil than traditional public schools (Arsen & Ni, 2012). Even when taking into account alternative categories of spending on administrative functions, our findings contradict this prior research. In our analysis, spending on administration includes certificated administrative personnel only. In contrast, Arsen and Ni (2012) present a more nuanced comparison of administrative spending by charter and non-charter schools. They divide administration into three subcategories: general administration, business administration and school administration. General administration includes spending associated with a school s governing board, legal fees and the portion of

31 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 31 Education Management Organization (EMO) 18 fees associated with executive leadership. Business administration includes typical business office activities such as payroll, accounting, budgeting and purchasing; these activities may also be part of EMO fees. School administration includes spending on site-based administration such as principals, assistant principals and clerical staff, similar expenses to our administrative category. Arsen and Ni (2012) find that charter schools spend more than traditional public in general administration. We include other personnel and services in the operations category. Charter schools in California spend an average of $78 per pupil less on other personnel, including clerical and financial personnel, and an average of $99 more per pupil on other services, including legal services, than traditional public schools. If administration, other personnel and other service expense categories are combined, the difference in spending between charter schools and traditional public schools for administrative services remains almost the same. Similar to the finding about spending on teachers, one of the reasons charter schools in California spend less on administration than traditional public schools may be lower salaries. Scholars find that administrators in charter schools tend to have less experience than those in traditional public schools (Perez et al., 2006), thus charter schools pay lower administrative salaries than traditional public schools (Zimmer et al., 2003). Another possible explanation for the conflicting findings rests in differing state policy contexts and charter school governance structures. Arsen and Ni (2012), who find substantial differences in spending on general administration, examine charter schools in Michigan where 72% are affiliated with an EMO, whereas in California only 16% of charters are affiliated with an EMO (NAPCS, nd). Similarly, Miron & Urschel (2010) find charter schools with EMO affiliations spend more than traditional public schools and more than independent charter schools on administration. Collectively, this evidence 18 Typically, Educational Management Organizations (EMOs) are for-profit businesses that oversee charter school operations and Charter Management Organizations (CMOs) are non-profit businesses that overs charter school operations. In our discussion, EMO refers to both CMOs and EMOs, as do both the Arsen & Ni (2012) and the Miron & Urschel (2010) studies.

32 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 32 suggests that charter schools associated with EMOs spend more on administration than independent charter schools. In states like California where a large number of charters operate independently of EMOs, charters spend the same or slightly less than public school districts on administration. Another possible explanation for the difference in per pupil spending may be related to economies of scale, where the per pupil cost for administration decreases as the total enrollment increases. However, Arsen & Ni (2012) find evidence that the difference in spending on administration may not be attributed solely to economies of scale. Charter schools in Michigan spend $520 more per pupil than traditional public schools in districts with a single school and similar number of students, presumably with similar economies of scale (Arsen & Ni, 2012). Though our finding that charter schools spend less per pupil on administration than traditional public schools conflicts with prior research generally, our analysis also aligns with Arsen and Ni s finding that differences are not solely attributable to economies of scale. We find some evidence of economies of scale in the regression analysis. An increase in the number of students enrolled in a school is associated with lower per pupil spending on administration. However, this factor in addition to all the other controls in the regression model, explain almost none of the difference between charter school spending and spending in traditional public schools. With no controls, the difference is $163; with all controls the difference is practically the same at $ Consulting Services Expenditures The combined gap in pupil support and administrative spending shows charter schools spending $828 less than traditional public schools on average. Combined with evidence from others research that charter schools employ fewer pupil support and administrative personnel, one conclusion is charter school students and staff demand fewer services, possible evidence of fiscal efficiency. Another conclusion is charter schools do not receive enough money to purchase necessary services. However, the observed gap in spending on pupil support and administration is largely offset by spending on consultant services. On average charter schools spend $1,015 per pupil on consulting services each year, while traditional public districts spend $419 per pupil, a gap of nearly $600.

33 Per Pupil Dollars C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 33 The evidence that charter schools spend more than twice as much on consulting services than traditional public schools persists across the nine years of data examined. As illustrated in Figure 6, charters spend about $400 more than traditional public schools in the first four years of data. By , the gap in spending widens noticeably, with per pupil spending by traditional public schools decreasing slightly and charter school spending on consultants increasing to a peak of $1,116 in , $750 more than traditional public schools. Figure 6 Spending on Consultants for Charter Schools and Traditional Public Districts ( dollars) $1,400 $1,000 $600 $ Consulting --TPS Consulting --Charter Though information about the specific services purchased from consultants is not included in the finance data, it is possible that consultant services include pupil support and administrative services. The approximately $600 per pupil additional spending by charter schools for consulting services nearly makes-up the $828 combined gap, between charter schools and traditional public schools, in spending for pupil support and administration. Charters may in fact be spending more on pupil support and administrative services by purchasing part-time or as-needed services for psychology, health and special education related services, as well as general administrative services such as financial management. For example, several charter schools in California with whom we consulted with in the last decade contracted with school psychologists and speech therapists to provide services for students as legally required on Individual Education Plans (IEPs) at an hourly consultant rate. In addition many charter schools contract

34 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 34 for back office support such as human resources, payroll and accounting services, which may show up as fees paid to consultants rather than salaries for administrative personnel. The higher spending by charter schools on consultants, combined with the possible explanation that charters may be purchasing part-time or as-needed pupil support and administrative services, may suggest that charter schools use consultants because the lack economies of scale. However, when controlling for enrollment, charter schools still spend substantially more than traditional public schools, though the magnitude of the difference decreases. In other words, charter schools that serve the same number of students as traditional public schools spend more on consultant services, refuting the proposition that spending on consultants is driven entirely by differences in enrollment. Spending on consulting services is the one area where new and more mature charter schools differ, as shown in Figure 7. Long-term charter schools spend approximately $350 per pupil less than newly founded charter schools average of $1,182 per student on consultants (see in Table A.4). This pattern persists across most of the decade and suggests new charter schools may have more need for professional consulting services than more experienced schools. One possible explanation for this difference may be that new charter schools hire consultants to facilitate initial start-up activities, such as writing charter applications, setting up human resources and financial management systems, and conducting professional development for newly hired instructional staff. Since restricted revenue decreases as charter schools mature, as do consulting services, it is reasonable to believe that a proportion of the consulting fees occurring during the first five years of operation are funded by start-up grants, which disappear as charters gain experience.

35 Per Pupil Dollars C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 35 Figure 7 Spending on Consultants by Newly Founded and Long-Term Charter School ( dollars) $1,400 $1,000 $600 $ Consultants -- New Charters Consultants -- Long-term Charters The difference in spending between new and long-term charter schools appears to be fading in more recent years, as the average per pupil spending of long-term charter schools on consultant services is increasing. The emerging pattern may provide evidence contrary to hypothesis about start-up expenses suggested above. Rather, it may suggest that charter schools are finding consultants a useful way to meet temporary or part-time needs on an ongoing basis. Further examination of the spending patterns as more recent finance data becomes available, as well as additional research about the specific services purchased and decision-making process for the purchase of consultant services, is necessary to completely understand charter school spending on consultant services. 4.6 Operations Expenditures In addition to higher spending on consultants, charter schools also spend more than traditional public schools on operations. On average, charter schools in California spend $1,269 per pupil annually, or 16% of their total spending, on operations expenses, including capital expenses such as the acquisition and improvement of land, buildings and equipment, and non-capital expenses such as rent, maintenance, housekeeping, communications, food service, legal service and operations personnel. In contrast,

36 Per Pupil Dollars C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 36 traditional public schools spend only $783 per pupil, or 9% of their general fund expenditures, on operations. In sum, charter schools spend about $480 more per pupil than traditional public schools. The difference in spending between charter schools and traditional public schools is most notable in maintenance and operations and capital equipment expense categories, listed in Table 3. Charter schools spend an average of $71 more per pupil on the acquisition and upkeep of capital equipment and nearly $400 more per pupil on the maintenance and operation of the facilities than traditional public schools. As previously noted, the average lower spending of $78 per pupil by charter schools for other personnel is offset by the average higher spending of $99 per pupil by charter schools for other services. The pattern of higher spending on operations by charter schools when compared to traditional public school districts is one that is widening overtime. As displayed in Figure 8, spending by charter schools on operations has risen over the decade with a low of $1,115 per pupil in , a peak of $1,355 in , and a rate of $1,258 per pupil in Spending by traditional public school districts has also risen over the decade. While traditional public schools had a similar upward trend in the beginning of the decade and a peak in , and a steady rate of spending since. Figure 8 Spending on Operations for Charter Schools and Traditional Public Districts ( dollars) $1,400 $1,000 $600 $ Operations--TPS Operations--Charter Again regression analysis results support the conclusions drawn above. By controlling for student enrollment, we are assured that the noted difference is not the function of lack of economies of

37 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 37 scale, but rather that charter schools spend more on operations than public school districts with a similar size student body. Further, including county fixed effects accounts for spending differences in operations driven by the geographic location of schools, with charter schools concentrated in high price urban areas. With the complete set of control variables described in the methods section, the regressions show that the difference between charter school and traditional public school spending on operations is slightly higher in magnitude than the mean per pupil spending comparison (see Table 3). The study of spending on facilities, operations and maintenance is complex as it includes both capital expenses and non-capital expenses. Researchers do not agree on resource allocation by charter schools for operations and maintenance. Like the findings in our analysis, Arsen and Ni (2012) find charters spend a larger share of their current operating budget on facilities and operations than traditional public schools. In contrast, Miron and Urschel (2010) claim charter schools spend $420 per pupil less on facilities and operations than their public school counterparts. The disparate findings about charter school spending on facilities may relate to differences in access to capital funding extant in the research on charter school revenues. Research documents the vast inequity in capital funding for charter schools (Krop & Zimmer, 2005; Nelson et al., 2004; Zimmer et al., 2003). Speakman and Hassel (2005) identify only seven states and NAPCS (2013a) notes just five states that do provide equal access to capital funding for charter schools. Miron and Urschel (2010) use national datasets, which include states with a variety of funding mechanisms for capital costs, possibly resulting in findings that conflict with those presented here and by Arsen & Ni (2012). In a state that does not reportedly give charter schools access to capital facilities funds, Michigan charter schools spend more of their operating budget on facilities costs (Arsen & Ni, 2012). Arsen and Ni (2012) argue that charters cannot raise funds for capital construction and are reduced to renting facilities. Rent is usually paid by general (non-capital) funds, rather than capital funds, which may explain the higher spending by charter schools on operations and maintenance than non-charter schools. The observed difference in spending on operations relates only to those expenses occurring from general fund. In California, traditional public schools often use capital accounts for revenue and expenses

38 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 38 related to acquisition, construction, and maintenance of facilities. However, charter schools in California almost exclusively use the general fund for capital expenses. Only eighteen charter schools use a capital expense fund in the nine years of financial data. 19 Given this context, it is likely that charter schools spend less overall for facilities, but more from the general fund. Further research, examining the capital funds of traditional public schools in California, may be necessary to fully understand the differences in spending between charters and traditional public school districts for facilities. In California, a state that provides access to multiple programs for facilities funding for charter schools, charters spend nearly $400 more than public school districts on facilities and operations. Under SB 740, charters serving students in economically disadvantaged areas may receive reimbursement for up to 75% of lease costs annually (California School Finance Authority, nd; CDE, nd; NACPS, 2013a). With Proposition 39, charter schools have the right to rent or purchase facilities from an authorizing district at the proportionate share of the district s costs (CCSA, nd; NACPS, 2013a). Further, the Charter School Revolving Loan Program provides loans of up to $250,000 at below market rates for charter school start-up costs (California School Finance Authority, nd; NACPS, 2013a). In addition, charter schools may access bond financing through the state, county board of education or county treasury (NACPS, 2013a). The Charter School Facilities Program provides grant funding for the construction or renovation of charter school facilities with a statewide bond approval (California School Finance Authority, nd; NACPS, 2013a). Another possible reason for the discrepant findings may include the lack of consideration of the length of charter operations. Indeed, newly founded charter schools, in California, spend more than longterm charters on operations, as outlined in Table A.4. In general charters spend $380 more per pupil each year on operations expenses than public school districts. However, new charter schools spend almost twice as much as traditional public schools, with new charter schools paying an average of $,1472 per pupil in operating expenses. By the time charter schools are in their third term, they are spending an 19 Tables displaying the types of funds and numbers of charter schools reporting financial activity in each fund for every year between and are available from the author.

39 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 39 average of $1,003 per pupil for operations, a decrease of nearly $500. Although the difference isn t statistically significant pooling nine years of data, in many individual years, the difference is statistically significant. The lower average spending by mature charter schools manifests mostly in maintenance and operations. As charter schools mature, their spending on maintenance and operating expenses such as rent and facilities maintenance decreases by nearly $300 per pupil, to only $200 more per student than public school districts. While spending in other subcategories also decreases as charter schools mature, the difference is non-significant and smaller magnitude. The notable lower spending by long-term charter schools, as compared to newly founded charter schools, provides further evidence that new charter schools experience substantial start-up expenses in their first few years of operation that cease or diminish overtime. 4.7 Savings Nine years of financial data clearly indicate charter schools in California spend less for instruction, pupil support and administration than public school districts. Evidence also indicates the lower spending by charter schools for these functions is largely offset by higher spending on consultants and operations. However, charter schools do not entirely supplant lower spending on instruction and pupil services with expenses for consultants and facilities, supporting a hypothesis that charter schools may save more than traditional public schools. It is true that charter schools spend a smaller proportion of their revenue in total than traditional public schools, while saving the remaining revenue. On average charter schools spend 96% of their revenue and save 4%, whereas public school districts spend 98% of their revenue and save 2%. Between and , the annual savings of charter schools averaged $330 per pupil, almost twice as much as the $159 per pupil average for traditional public schools. However, as Figure 9 illustrates, the savings patterns of both charter schools and traditional public schools are inconsistent in the last decade. From through , the rate of annual savings for both traditional public schools and charter schools increases steadily with traditional public school districts

40 Per Pupil Dollars C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 40 saving slightly more per pupil than charters. As economic times turn bad and school revenue falls later in the decade, traditional public schools savings becomes irregular with deficit spending in , little to no savings in two other years, and approximately $200 per pupil savings in two alternating years. In contrast, charter schools do not spend more than the revenue received in any year. Further, despite times of economic uncertainty in the last five years, charter schools saved more than public school districts. While the trend in average savings by charter schools is marked with alternating rise and dips, fluctuating between about $300 and $500 per pupil annually, the pattern is more consistent than that for traditional public schools. Figure 9 Annual Savings for Charter Schools and Traditional Public Districts ( dollars) $600 $400 $200 $0 -$ Annual Savings--TPS Annual Savings--Charter Though prior research on charter school savings patterns is limited, Arsen and Ni (2012) find charter schools in Michigan save more and build greater reserves than public schools. Noting the absence of increased school funding in , they describe how traditional public schools drew down savings to cover increasing employee benefits costs, while charter schools improved reserves by $362 per pupil. Our research aligns with Arsen & Ni s (2012) finding and regression analysis supports the conclusions drawn from descriptive statistics. Charter schools save more, though only slightly more, than traditional public schools on average. Controlling for characteristics of enrolled students and geographic

41 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 41 location of the school, charter schools save $129 per pupil more each year than traditional public schools. With the addition of annual savings to accumulated reserves, overtime charter schools may build higher per pupil reserves than traditional public schools. In general, researchers in the field of school choice acknowledge the need to study the savings patterns and fiscal reserves of charter schools (Miron & Urschel, 2010, Wohlstetter, Smith & Farrell, 2013). Evidence of annual savings by charter schools indicates financial viability, where charter schools spend less than the revenue received. Further, as annual savings accumulates year over year, charter schools build reserves which will help with long-term sustainability. Charter schools with reserves are likely more able to sustain operations with revenue from the state fluctuates or unforeseen expenses arise. Conclusion Several important conclusions emerge from this research, adding to the existing literature about charter school finance and illuminating resource allocation practices that may lead to the financial efficiency and viability of charter schools. California charter schools receive less funding per pupil than traditional public school districts. With this lower revenue, charter schools in California spend less on instruction, pupil support services and administration, but more on operations, consulting services and annual savings than traditional public schools. Many of these findings are aligned with research about charter schools in other states. Yet, this research reveals some new insights about consultant services and savings patterns in charter schools, which may be important to efficient and sustainable operations. The largest gaps in spending between charter schools and traditional public schools in California occur in instruction and pupil support services, where charter schools spend a combined average of nearly $1,250 per pupil less after controlling for differences in student populations and geographic location. Existing research provides evidence that lower spending on instruction is likely the function of employing less experienced teachers (Perez et al., 2006; Zimmer et al., 2003) and the lower spending on pupil support may be the result of employing fewer personnel (Perez et al., 2006). Further, in California, unlike other states included in prior studies, charter schools spend less on administration than traditional public

42 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 42 schools. Numerous potential explanations for why charters in California spend less on administration exist, including fewer personnel and fewer or lower EMO fees. The lower spending on instruction, pupil support and administration by charter schools may be evidence of fiscal efficiency. Unlike suggestions from other researchers, regression results provide evidence that lower spending on instruction and pupil support is not completely associated with the characteristics of students enrolled. When charter schools serve similar students as traditional public school districts, they spend still less on instruction, pupil support and administration. Charters may in fact be able to meet student needs with fewer and less experienced personnel than traditional public schools. Alternatively, charter schools may be replacing or supplementing pupil support and administrative services, typically provided by employees in traditional public school districts, with consultants. In fact, for charters in California the deficit in spending on pupil support and administrative services is nearly offset by higher spending on consultants. The use of consultants for temporary or parttime needs within a school may be a fiscally efficient practice which allows charters to save on higher full-time personnel costs and potentially permits greater rates of annual savings. This finding provides a unique contribution to the existing evidence about charter school resource allocation practices. In California, charter schools save more each year than traditional public schools. Though the magnitude of the annual difference is relatively small, overtime the annual savings of charter schools may lead to greater reserves. These reserves are necessary for weathering uncertain economic times and may ensure the financial viability and sustainability of charter schools. The evidence in this study begins to answer questions about the savings patterns and reserves of charter schools. Yet, more research is needed understand how and for what purpose individual charter schools build reserves overtime. The resource allocation practices revealed in this study, combined with findings from existing research, shed some light on how charter schools may avoid financial deficiency and ensure their sustainability. For example, charter schools may exercise their fiscal flexibility by hiring less experienced teachers and using lower cost part-time consultants rather than employing full-time personnel for pupil support and administrative services. The savings resulting from this spending pattern may in turn

43 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 43 improve charter schools ability to save for economic uncertainty. If this sort of fiscal flexibility and efficiency leads to the sustainable operation of quality charter schools, the information is important for policy makers and charter school operators who make decisions about revenue and resource allocation. On the other hand, the resource allocation patterns observed in this study may contribute to concerns about the sustainability of charter schools overtime. The fiscal flexibility and efficiency described in the previous paragraph may or may not lead to equal or better academic outcomes for students in charter schools as compared to traditional public schools; that question is outside the scope of this project. However, it debunks the supposition that charter schools granted autonomy and freed from regulation will spend more on instruction. Further, lower spending on instruction and pupil support may in fact undermine consumer demand for charter schools if families of potential students believe fewer resources are targeted toward instruction and student the supports within the school, and rather directed to outside consultants and operations. Finally, if the lower spending on instruction, pupil support and administration results in lower quality instructional experiences and poorer academic performance of students, charter schools may not be able to sustain operations. As noted in the introduction, the primary reason for the closure of charter schools is financial deficiency. However, within the last several years, a resounding call to close low performing charter schools from across the charter school sector is rising. While education research has yet to clearly establish a connection between spending and academic performance, it is plausible that revenue and resource allocation practices may contribute to charter school closure for both reasons, a concern for families of enrolled students, policy makers, and philanthropists, alike. Charter schools represent a small fraction of public schools in the United States. However, the promise of reform that charter schools hold represents a large proportion of the rhetoric of policy makers, the investment of philanthropists, and the hopes and dreams of families. This study shows that charter schools in California exercise fiscal flexibility by hiring consultants and saving more, both of which have the potential to create financial viability and sustainability. The implications however are unclear. Certainly efforts to achieve equitable funding for students in all public school options is imperative.

44 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 44 However, the investment of funds by government agencies and philanthropists, while potentially relieving revenue gaps that lead to financial viability, may not guarantee academic outcomes. Finally, families have choices, but resource allocation patterns likely do not provide useful information in selecting schools. Researchers must continue to unpack the decision making about revenue and spending practices of charter schools and strive to connect resource allocation practices with student outcomes to more fully understand the long-term sustainability of charter schools.

45 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 45 References Allen, J., Consoletti, A. & Kerwin, K. (2009). The Accountability Report: Charter Schools. Washington, D.C. Center for Education Reform. Andrews, M., Duncombe, W., & Yinger, J. (2002). Revisiting economies of size in American education: are we any closer to a consensus? Economics of Education Review, 21(3), Angrist, J.D., Dynarski, S.M., Kane, T.J., Pathak, P.A., & Walters, C.R. (2010). Who benefits from KIPP? Working paper 15740, National Bureau of Economic Research. Retrieved from Arsen, D. & Ni, Y. (2012). Is administration leaner in charter schools? Resource allocation in charter and traditional public schools. Education Policy Analysis Archives, 20(31). Arsen, D., Plank, D., & Sykes, G. (1999). School choice policies in Michigan: The rules matter. Retrieved from Michigan State University, Education Policy Center. Baker, B., & Ferris, R. (2011). Adding up the spending. Fiscal disparities and philanthropy among New York City charter schools. Retrieved from University of Colorado, National Education Policy Center website: Baker, B., Libby, K., & Wiley. K. (2012). Spending by the major charter management organizations: Comparing charter school and local public district financial resources in New York, Ohio, and Texas. Retrieved from University of Colorado, National Education Policy Center website: Batdorff, M., Maloney, L., May, J., Speakman, S., Wolf, P., & Cheng, A. (2014). Charter school funding: Inequity expands. Fayetteville, AR: School Choice Demonstration Project, Department of Education Reform, University of Arkansas. Retrieved December 1, 2014 from University of Arkansas website: Bifulco, R., & Ladd, H.F. (2006). The impacts of charter schools on student achievement: Evidence from North Carolina. Education Finance and Policy, 1(1), Betts, J., & Tang, Y. (2011). The effect of charter schools on student achievement. Retrieved from University of Washington, Center for Reinventing Public Education website: Buckley, J., & Schneider, M. (2009). Chapter 1 Introduction. In Charter Schools: Hope or Hype (pp. 1-29). Retrieved from California Charter Schools Association (CCSA). (nd). Proposition 39 Overview. Retrieved from CCSA website: California Charter Schools Association (CCSA). (nd). California Charter Schools by the Numbers. Retrieved November 22, 2014 from CCSA website: California Department of Education (CDE). (nd). Charter School Categorical Block Grant

46 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 46 Programs. Retrieved from CDE website: California Department of Education (CDE). (nd). Charter School Direct Funding Option. Retrieved from CDE website: California Department of Education (CDE). (nd). Charter School Facility Grant Program. Retrieved from CDE website: California Department of Education (CDE). (nd.) Public Charter School Grant Program. Retrieved from CDE website: California School Finance Authority. (nd). Charter School Facilities Program. Retrieved from California Treasurer website: Card, D., & Krueger, A. (1996). School resources and student outcomes: An overview of the literature and new evidence from North and South Carolina. Journal of Economic Perspectives, 10(4), Center for Research on Education Outcomes. (2009). Multiple choice: Charter school performance in 16 states. Retrieved from Consoletti, A. (2011). The State of charter schools, What we know and what we do not -about performance and Accountability. Washington D.C.: The Center for Education Reform. Retrieved from Center for Education Reform website: Duncombe, W., Miner, J., & Ruggiero, J. (1995). Potential cost savings from school district consolidation: A case study of New York. Economics of Education Review, 14, Frankenberg, E., Siegel-Hawley, G., & Wang, J. (2010). Choice without equity: Charter school segregation and the need for civil rights standards. Retrieved from University of California, Los Angeles, The Civil Rights Project at UCLA website: 12-education/integration-and-diversity/choice-without-equity-2009-report/frankenberg-choiceswithout-equity-2010.pdf Garcia, D. (2008). The impact of school choice on racial segregation in charter schools. Educational Policy, 22, Gill Phillips, S. (2010). Charter school spending: Is there a relationship between spending and student achievement in charter schools. Retrieved from ProQuest. (UMI Number: ). Gittleson, K. (2010). Charter School Philanthropy Gotham Schools. Retrieved October 27, 2010, from Greenwald, R., Hedges, L., & Laine, D. (1996). Interpreting research on school resources and student achievement: A rejoinder to Hanushek. Review of Educational Research, 66(3), p Hanushek, E. (1996). School resources and student performance. In G. Burtless & M. Armacost (Eds.), Does Money Matter? The effect of school resources on student achievement and adult success (pp ). Washington, DC: The Brookings Institution.

47 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 47 Hanushek, E. (1997). Assessing the effects of school resources on student performance: An update. Educational Evaluation and Policy Analysis, 19(2) Hayes, C. & Keller, E. (2009). A Cost Estimation Tool for Charter Schools. Retrieved November 22, 2014 from the National Resource Center on Charter School Finance & Governance website: Hill, P., & Roza, M. (2008, April 29). The end of school finance as we know it: A brief history, and new direction. Education Week, 36. Hoxby, C., & Murarka, S., (2009). The New York City charter schools evaluation project: How New York City s charter schools affect achievement. Working paper 14852, National Bureau of Economic Research. Hoxby, C.M., & Rockoff, J.E. (2005). The impact of charter schools on student achievement. Retrieved from RAND website: Kern, N. & Gebru, W. (2014). Waiting Lists to Attend Charter Schools Top 1 Million Names. Policy report retrieve November 22, 2014 from National Alliance for Public Charter Schools website: Report.pdf Krop, C., & Zimmer, R. (2005). Charter school type matters when examining funding and facilities: Evidence from California. Education Policy Analysis Archives, 13(50). Lake, R., Dusseault, B., Bowen, M., Demeritt, A., & Hill, P. (2010). The national study of Charter Management Organization (CMO) effectiveness: Report on interim findings. Retrieved from University of Washington, Center for Reinventing Public Education website: Miron, G. & Applegate, B. (2009). Review of Multiple choice: Charter school performance in 16 states. Retrieved from University of Colorado, National Education Policy Center website: Miron, G., Evergreen, S., & Urschel, J. (2008). The impact of school choice reforms on student achievement. Retrieved from University of Colorado, National Education Policy Center website: Miron, G., & Urschel, J. (2010) Equal or fair? A study of revenues and expenditures in American charter schools. Retrieved from University of Colorado, National Education Policy Center website: Monk, D. H, & Hussain, S. (2000). Structural influences on the internal allocation of school district resources: Evidence from New York State. Educational Evaluation and Policy Analysis, 22(1), National Alliance for Public Charter Schools. (nd). CMO and EMO public charter schools: A growing phenomenon in the charter school sector. Retrieved from National Alliance

48 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 48 from Public Charter Schools website: content/uploads/2014/01/napcs-cmo-emo-dashboard- DETAILS_ T pdf National Alliance of Public Charter Schools. (2013a). The public charter schools dashboard. Retrieved from National Alliance of Public Charter Schools website: National Alliance for Public Charter Schools. (2013b). Measuring up to the model: A tool for comparing state charter school laws. Retrieved from National Alliance for Public Charter Schools website: National Alliance for Public Charter Schools. (2013c). A growing movement: America s largest charter school communities. Retrieved November 22, 2014 from National Alliance for Public Charter Schools website: Nelson, F., Muir, E., & Drown, R. (2000). Venturesome capital: State charter school finance systems. Washington, DC: American Federation of Teachers Educational Foundation. Nelson, F., Muir, E., & Drown, R. (2003). Paying for the Vision: Charter School Revenue and Expenditures. Washington, DC: Office of Educational Research and Improvement, U.S. Department of Education. Nelson, F., Rubenstein, M., & Mahoney, L. (2004). Financing Autonomy: Charter School Finance Case Studies. Washington, DC: American Federation of Teachers. Osberg, E. (2006). Charter school funding. In P. Hill (Ed.), Charter schools against the odds: An assessment of the Koret Task Force on K-12. (pp ). Standford, CA: Hoover Institution. Pandolfo, J. (2012). The effect of economies of scale on California school districts expenditures (Unpublished doctoral dissertation). University of California, Davis, CA. Perez, M., Anand, P., Speroni, C., Parrish, T., Esra, P., Socias, M., & Gubbins, P. (2006). Charter schools in California: A review of their autonomy and resource allocation practices. Retrieved from American Institutes for Research website: Ravitch, D. (2010). The death and life of the great American school system. New York: Basic Books. Resmovits, J. (2014, April 30). Charter schools get less money that public schools. Is that a problem? Huffington Post. Retrieved from: Ritter, G., Jensen, N., Kisida, B., & McGee, J. (2010). A closer look at charter schools and segregation, flawed comparisons lead to overstated conclusions. Education Next, 10(3), Rose, H. & Sengupta, R. (2007). Teacher compensation and local labor market conditions in California: Implications for school funding. Retrieved from Public Policy Institute

49 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 49 of California website: Rose, H., Sengupta, R., Sonstelie, J., & Reinhard, R. (2008). Funding formulas for California schools: Simulations and supporting data. Retrieved from Public Policy Institute of California website: Rose, H., Sonstelie, J., Reinhard, R., & Heng, S. (2003). High expectations, modest means: The challenge facing California s public schools. Retrieved from Public Policy Institute of California website: Rowan, B., Bosser, S., & Dwyer, D. (1983). Research on effective schools: A cautionary note. Educational Researcher, 12(24), Sohoni, D., & Saportio, S. (2009). Mapping school segregation: Using GIS to explore racial segregation between schools and their corresponding attendance areas. American Journal of Education, 115, Speakman, S., & Hassel, B. (2005). Charter school funding: Inequity s next frontier. Retrieved from Thomas B. Fordham Institute website: U.S. Department of Education. (2004). Innovations in Education: Successful Charter Schools. Retrieved from U.S. Department of Education website: U.S. Department of Education (nd.). Programs. Charter Schools Program State Educational Agencies (SEA) Grant. Retrieved November 27, 2014 from U.S. Department of Education website: Weston, M., Sonstelie, J., & Rose, H. (2009). Winters, M. (2010, March 1). Charter Schools Aren t Just Better, They Cost Less! New York News. Retrieved December 1, 2014, from: Daily Wohlstetter, P., Smith, J., & Farrell, C. (2013). Choices and Challenges: Charter school Performance in Perspective. Cambridge, MA: Harvard Education Press. Zimmer, R., Buddin, R., Chau, D., Daley, G., Gill, B. Guarino, C., Brewer, D. (2003). Charter school operations and performance: Evidence from California. Santa Monica, CA: RAND Corporation.

50 Operations Pupil Support Instruction C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 50 Appendix A Table A.1 Charter School and Traditional Public School Expenditures as Percent of Total Current Expenses ( dollars) Decade Average TPS Charter TPS Charter TPS Charter TPS Charter Instruction Teachers Instructional aides Instructional materials Pupil Support Pupil service personnel Tuition and transfers Administrative personnel Consulting services Operations Maintenance and operations Other personnel Other services Capital equipment Annual Savings

51 Operations Pupil Support Instruction C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 51 Table A.2 Mean Per Pupil Expenditures of Charter Schools and Traditional Public Schools ( dollars) Decade Average TPS Charter TPS Charter TPS Charter TPS Charter Total Revenue 8594* * * * 8171 Total Expenditures 8647* * * * 7840 Instruction 5412* * * * 4140 Teachers 4614* * * * 3205 Instructional aides 349* * * * 240 Instructional materials 449* ** * * 695 Pupil Support 1077* * * * 455 Pupil service personnel 873* * * * 248 Tuition and transfers * * 207 Administrative personnel * * * 961 Consulting services 439* * * * 1015 Operations 641* * * 1269 Maintenance and operations 318* * * * 695 Other personnel 240* * * * 176 Other services 30* 91 42* 61 30* * 133 Capital equipment 53* * * 94 54* 125 Annual Savings -54* * * * 330 * Denotes where charter schools differ from traditional public schools in the same category and year (p<.05) Note: Mean per pupil amounts are weighted by enrollment.

52 C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 52 Table A.3 Regression Results: Comparison of Charter School and Traditional Public School Spending, Controlling for Student Characteristics and Geographic Location, Using Panel Data ( dollars) Charter without controls Charter with controls Instruction Pupil Support Admin Consulting Operations Savings -1, (46)** (52)** (17)** (21)** (27)** (29)** (89)** (162)** (33)** (50)** (121)** (34)** % Middle School Students % High School Students Enrollment (log) % FRPM Per Pupil Revenue (255) (294) (104) (96) (179) (91) (85)** (86)** (37)** (45) (80) (47) (18)** (44) (8) (4)** (18) (6)** (104) (119)** (42)** (57)* (77)** (43)** (0.0)** (0.1)** (0.0)** (0.0)** (0.0)** (0.0)** , Constant (359)** (837) (141) (112) (266) (129)** R N 10,850 10,850 10,850 10,850 10,850 10,850 County FE YES YES YES YES YES YES Year FE YES YES YES YES YES YES * p<0.05; ** p<0.01 Note: The coefficient on charter schools in the top row is the result of a regression that contains only the dichotomous variable for charter. Below the line are the coefficients for the regressions with all control variables. We include the top row to show spending differences change with the addition of our control variables. In almost every category, the additional controls cause very little change in the charter coefficient, with the instruction category as the only exception. Adding the controls explains about half the observable difference in instructional spending between charter schools and traditional schools. Even with the controls, charter schools still spending significantly less on instruction.

53 Operations Pupil Support Instruction C H A R T E R S C H O O L S P E N D I N G A N D S A V I N G I N C A L I F O R N I A 53 Table A.4 Per Pupil Expenditures of Charter Schools with Various Length of Operating Histories ( dollars) Decade Average New Long New Long New Long New Long Total Revenue * Total Expenditures * Instruction Teachers * * Instructional aides Instructional materials * Pupil Support Pupil service personnel Tuition and transfers Administrative personnel Consulting services * * Operations * * Maintenance and operations * * Other personnel Other services * * Capital equipment Annual Savings *p<.05