INSTITUTE FOR RESEARCH ON EDUCATION POLICY & PRACTICE Informing change & promoting innovation through rigorous & collaborative research

Getting Down to Facts: School Finance and Governance in Summary By Susanna Loeb, Anthony Bryk, and Eric Hanushek September 2007

“Getting Down to Facts” is the largest independent investigation ever of how California governs and funds public education. It was commissioned at the request of a bipartisan group of California leaders, including the Governor’s Advisory Committee on Educational Excellence, the President Pro Tem of the California Senate, the Speaker of the California Assembly, the Superintendent of Public Instruction, and the state Secretary of Education.

The purpose of this unprecedented project is to and fail to help students meet state performance Getting Down to Facts describe California’s school finance and gover - goals, especially students from low-income fam - A research project designed nance systems, identify aspects of those systems ilies. While meaningful reform will likely re - to provide California's policy- that hinder the effective use of resources, and quire added investment, it is also clear that makers and other education estimate costs of achieving a range of student absent reform, directing more money into the stakeholders with comprehensive information about the state’s outcome goals. The project is not designed to current system is unlikely to result in the dra - school finance and governance advance specific policy recommendations, but matic improvements in student achievement systems, and lay the ground - rather aims to provide a common factual needed to reach state goals. Our research indi - work for a serious and substantive ground to promote informed conversation cates that what matters most are the ways in conversation about needed reforms. The project was made among policymakers and the public as they which current and new resources are used. To possible by grants from the consider necessary reforms. this end, the Getting Down to Facts reports Bill & Melinda Gates Foundation, Coordinated by ’s provide a framework for assessing our reform the William and Flora Hewlett Institute for Research on Education Policy and options going forward. Foundation, the James Irvine Foundation, and the Stuart Practice, the project was independently fi - Foundation. nanced by the Bill & Melinda Gates The Problem: California lags Foundation, the William and Flora Hewlett significantly behind other states This summary was prepared by IREPP and approved by the authors. Foundation, the James Irvine Foundation and in student achievement the Stuart Foundation. The resulting 23 reports Despite the development of challenging educa - For the full text of the author’s represent 18 months of research completed in tion standards and sustained attention to report and the other studies in early 2007 by scholars at universities and re - school improvement over the past decade, this project, go to http://irepp.stanford.edu/ search institutions across the nation. California continues to lag behind other states projects/csfg.html. The research carefully documents what many on several different measures of student educators report anecdotally: While good achievement. The problem is serious. For ex - things are happening in many districts, schools ample, on the 2005 National Assessment of

Institute for Research on and classrooms, California’s school finance and Education Progress, California ranked 7th low - Education Policy & Practice governance systems are fundamentally flawed est in eighth grade math among the 50 states Stanford University 520 Galvez Mall, CERAS Building Rm #518 Stanford, CA 94305 These studies need to be taken very seriously by everyone in the education irepp.stanford.edu debate. This is just a starting point for what I hope will be a renewed focus in www.irepp.net “the Legislature on increasing student achievement with needed reform. 650.736.1258 [email protected] —Governor Schwarzenegger, March 14, 2007 ” and the District of Columbia. The replace policies that don’t work and to have the resources, incentives and au - story is at least as bad in other sub - maintain or expand those that do. thority to accomplish their obligations. jects. California performed 3rd lowest However, California does not have in reading, ahead of only Hawaii What do California’s school such a system and, as a result, local and the District of Columbia, and finance and governance responses are not as intended. The 2nd lowest in science, ahead of systems look like today? current system of parallel public re - only Mississippi. Some suggest that School governance in California is porting of school performance under California’s position simply reflects characterized by a hodgepodge of restrictive the federal the large minority populations in rules and regulations that often hinder, (with its associated adequate yearly the state, but the facts belie this. rather than promote, student achievement. progress requirements) versus school California schools do not do well Any informed discussion of school fi - performance under state law (with for any sub-group, including non- nance requires an understanding of goals under its academic performance Hispanic white students. Significant the governance system in which it op - index, or API) sends mixed signals to progress will require fundamental and erates and an assessment of any ob - parents and educators. comprehensive change. stacles to reform imposed by that Perhaps even more important than The low achievement of Cali- system. Researchers identify a num - this confusing lack of alignment, how - fornia’s students will almost certainly ber of key characteristics that should ever, are the substantial constraints on hurt their economic outcomes later in frame consideration of any system: resource use that local personnel face in life and is likely to be detrimental for transparency; simplicity; innovation, trying to respond to the challenges they the state as a whole. There is mounting flexibility and responsiveness; ac - encounter. While other states with evidence that educational quality countability; and stability. Getting strong accountability systems have re - measured by test scores is directly re - Down to Facts’ conclusions about the duced regulations to enable local im - lated to individual earnings, worker California system of governance stem provement initiatives, (Florida and productivity and . In from this framework. Connecticut, for examples), California a global knowledge economy, the eco - Excessive regulation. California has not. Instead of encouraging flexi - nomic growth of regions and nations is places substantial restrictions on bility and innovation at the local level, affected by the skills of workers, which schools’ and districts’ use of resources. many of California’s state policies con - in turn are directly related to student These restrictions impose heavy com - strain local actors – forcing very similar learning outcomes. pliance costs and make it difficult for policies regardless of either local needs local actors to respond to incentives in or capacities. Ultimately, accountability The Research Questions the accountability system. Regulatory and responsibility are not well aligned The hypothesis underlying “Getting requirements in an Education Code because schools and districts are held Down to Facts” is that improving with 500 chapters and more than accountable for student learning but California’s school finance and gover - 1,250 articles, stifle local innovation they are not given responsibility and nance structures will enable schools to (such as extending the school day, pro - authority to allocate resources. be more effective and to address an viding for teacher collaborative time or Low priority given to administrative achievement gap that remains signifi - improving reading instruction). They capacity. Despite the increasing com - cantly wide. In light of that assump - also impose needless obstacles on local plexity of local administrative roles, es - tion, researchers sought to answer school administrators causing them to pecially the modern principalship, the three questions: focus on compliance, and its attendant state places little emphasis on adminis - G What do the school finance and gov - paperwork, rather than on meeting trative capacity. Though the empirical ernance systems look like in teaching and learning goals. At the evidence on the effects that principals California? same time, surveys of superintendents have on student learning is not as exten - G How can dollars be used more effec - and principals reveal that constant sive as it is for teachers, it seems clear tively to meet student outcome change of state-level policy hinders nonetheless that principals play a key goals? planning and frustrates school and dis - role in the effective leadership of in - G To what extent are additional re - trict staff. structional improvement at the school- sources necessary to meet state Governance and accountability sys - site. Principals in California are less goals? tems are often at cross-purposes. In a likely than principals in other states to Answers to these questions can serve well-designed accountability system, have participated in an administrative as the basis for assessing proposals to all players understand their roles and internship, to have access to mentoring

2 | School Finance and Governance in California | September 2007 or coaching in their work or to a princi - The System is simply a barnacle of the past. Almost no one, pal’s network while on the job, and to save a few highly paid experts, understands it. have participated regularly with teach - ers in professional development. “ Moreover, California has more students —Mike Kirst, Stanford University, former member California State Board of Education per school administrator and fewer dis - ” trict administrators per school adminis - leaving districts with little ability to local preferences. trator than the rest of the country. The raise additional funds for school oper - Inequitable by any measure. regulatory environment discussed above ations. Perhaps as a result of the state Differences in spending across imposes a heavy compliance burden on playing such a central role in finance, California districts are substantial school administrators. With these re - the state has also taken control of and not systematically tied to costs, sponsibilities and limited numbers of other aspects of school policy, requir - needs or demands. Despite a court-or - school administrators, it is not surpris - ing districts to spend their revenue in dered school finance equalization ing that principals in California report specific ways. In fact, California pre - plan, there remains a wide variation that they spend substantially less time scribes at the state level more of how in spending across California school overseeing instruction at their schools dollars should be spent than do other districts. The difference in total ex - than do principals in other states. There states. There is some research evidence penditures, excluding capital outlays, is no reason to believe, however, that that districts use state-prescribed aid in a district at the 25th percentile of simply expanding staff by itself will lead less effectively for the purpose of im - spending and a district at the 75th to significant improvements in student proving student outcomes than general percentile of student-weighted spend - outcomes. purpose aid. These findings suggest ing is more than $1,000 per student. that, in California, districts can allo - The system could still be considered California’s school finance system is cate resources more effectively when equitable if spending patterns effec - unnecessarily complex and is not rationally given flexibility than when the alloca - tively accounted for differences in aligned to support the accountability and tion is determined solely by the state. local needs. In fact, however, district performance standards imposed on local A complex and irrational finance poverty level, racial and ethnic educators. system. The number of dollars avail - makeup, urban status and district The school finance system determines able to each school district is largely grade span explain only a small por - not only the dollars that flow to dis - an historical artifact of spending in tion of the variation in spending. tricts, schools and classrooms but also the 1970s combined with confusing Unstable revenue from year-to-year. in many cases how these dollars may be categorical grant programs. As a re - The source of school funding is unsta - used. As with governance, there is no sult, similar districts can receive sub - ble both in terms of revenue fluctua - consensus on the one best way to fund stantially different revenues per pupil, tions and the lateness of the budgeting a public school system. In assessing and differences in student needs across process. Stock price volatility and the California’s state system of finance, the districts are not systematically ac - state’s relatively progressive personal in - Getting Down to Facts studies consid - counted for in determining revenue come tax have created years of boom ered a variety of indicators: equity, suf - levels. In addition, the finance system and bust for California schools. The im - ficiency of dollars, clarity and is extraordinarily complex and im - portance of stability is evidenced in the simplicity, administrative requirements, poses substantial and costly compli - principal and superintendents surveys. the extent to which the system facili - ance burdens on school districts. More then three-quarters of superin - tates or hinders the effective use of re - Pre-dating the implementation of tendents responded that knowing the sources for meeting goals, and the modern accountability systems, the budget earlier would be a great deal of stability of funding sources. current finance structure has never help or essential to improving outcomes Highly centralized. In California, been updated to align with the states for students. district spending levels are set, with accountability system, nor redesigned only minor exception, at the state level to help local officials meet student California spending is below the national and a higher proportion of funds come performance goals. By contrast, a average. from state revenues than in most other number of states including Texas have Even factoring recent substantial in - states. This state control is, at least in implemented reforms in which the creases, California lags behind a major - large part, a result of Proposition 13 dollars going to districts are much ity of other states in education which limits the local property tax; more closely tied to needs, costs, and spending. For example, California gen -

September 2007 | School Finance and Governance in California | 3 Figure 1 Staffing Ratios in California and Other States Common Core • teacher preparation requirements affects of Data 2003-04 school district demographic data the pool of available teachers. Whether or how changing these requirements will 50 translate to improved student outcomes 45 California 40 New York is less clear. In light of this, and the clear 35 Texas effect of certification requirements on 30 Florida the pool of teachers, it is an area worthy 25 All Other States of experimentation. 20 The story is similar with professional 15 development for principals. California 10 principals report being less engaged in 5 evaluating and supporting teachers, in 0 working with teachers to improve their Student/Teacher Student/School School Administrator/ Administrator in 10s District Administrator practices, in helping to develop curricu - lum plans, in fostering teacher profes - sional development, and in using data to monitor and improve instruction than do erates approximately the same revenues ters’ degrees demonstrate no measurable principals in other states. Given the im - per pupil as Texas and Florida, approx - benefits for students. There seems to be portance of school-based leadership to imately $5,500 less than New York, little reason to keep these requirements student outcomes, and the sparse evi - and approximately $630 less than the or to peg salary enhancements to them, dence on how to enhance it effectively, average of the remaining states. as is currently the case in most districts. this area is ripe for innovation coupled Moreover, California’s costs are higher In contrast, field-based experiments with careful assessments of effects. than those in most other places, due pri - have demonstrated that targeted profes - Due-process rules and the inability to marily to the higher wages of college- sional development, aligned to stan - dismiss ineffective teachers. The one fac - educated labor. After adjusting for dards, and implemented well can affect tor that emerged most consistently across salary differences across states, Texas improvements in teaching and learning. studies as inhibiting local leadership was spends approximately 12 percent Dmore Along these lines, recent policies in the difficulty in dismissing ineffective than California; Florida, approximately California have aimed to make profes - teachers. Both the principals and super - 18 percent more; New York, approxi - sional development more relevant to the intendents surveyed ranked this factor as mately 75 percent more, and the rest of work of teaching. However, it is difficult the most important change that could the country, approximately 30 percent to tell whether these new policies, such help them improve student outcomes. more, on average. The lower spending as those supporting school-based Increasing principals’ authority in this in California is reflected also in high coaches or mentors, are delivering the area does not necessitate removing due- student-to-staff ratios, including fewer desired outcomes. This is one of many process rights for teachers. One poten - teachers and administrators per pupil. instances in which a promising initiative tially productive alternative, for (See Figure 1.) has been scaled very rapidly without example, would be to create a fair and concern for whether schools have the accurate system of evaluation for teach - California does not have a coherent system capacity to sustain such development ers that can be used to optimize profes - for supporting the entry, development, and with quality and with no plan to learn sional development and job assignment retention of quality teachers and administrators. from the program so that the overall as well as to provide the basis for an ef - Weak state policies on teacher and prin - productivity of state policy might be en - fective due-process system. In addition,

cipal A recruitment and professional hanced. Simply, the state has no means though principals cite the difficulty of development. Teachers are central by which to tell whether the program is dismissing teachers as a barrier to in - to improving student outcomes. working and thus no way to know how structional improvement, it is also im - Unfortunately, California’s teacher poli - to adjust the program so that it can best portant to note that theyindicate that cies are not currently coordinated and achieve its goals. they would seek to remove only a small designed to optimize the teacher work - Likewise, state policies for teacher number of teachers—two or less in most force. As an example, generic require - certification and licensure for entry de - schools—if they had the authority to do ments such as units of professional serve careful reevaluation. There is so. These comments reflect a concern development credits or unspecific mas - ample evidence that the nature of that just a few ineffective teachers can

4 | School Finance and Governance in California | September 2007 undermine reform efforts at a school. particularly concerning teachers, that reform, no one policy change that will Along similar lines, teachers are more is collected by agencies other than the forever assure an optimal school sys - likely to engage actively in reform when Department and are difficult to link. tem. Instead, Getting Down to Facts principals have real authority to act, even Most importantly, California’s current points to policy areas that are worth if principals rarely use that authority. system does not follow students and pursuing because the evidence suggests Problems with current salary sched - teachers over time and does not link that changes in these areas could pro - ules. Teacher salary scales also do not them together and to the programs duce benefits for students. Among support a highly effective teacher work - and resources that they experience so these areas are: force. Within most districts teachers that we can evaluate effectiveness. G relaxation of state regulations and with the same years of experience and Several other states have assembled restrictions on categorical funds to education receive the same base pay. such systems to record policies and fol - allow greater local flexibility for re - First, this schedule fails to recognize any low the performance of individual stu - source allocation, including the flexi - differences in the effectiveness of teach - dents from pre-K through college and bility to make more effective use of ers. Second, it fails to recognize differ - even beyond into the labor force. From instructional time and possible ex - ences in the difficulties of some teaching these systems, policymakers and pansion of that time especially in assignments versus others. All other school administrators can track how schools with high concentration of things equal, many teachers prefer students are progressing, how different disadvantaged students; schools with higher scoring and presum - teachers and programs are affecting G simplification and rationalization of ably easier to teach students. Thus this performance, and the effectiveness school finance formulas to promote schools with a high proportion of stu - of different uses of resources. With better strategic planning for the best use dents in poverty are often left with less carefully linked data on students and of resources by local school officials; experienced teachers and teachers with teachers, state education leaders would G efforts to support the recruitment less strong academic preparation – and be able to assess the efficacy of differ - and development of effective teach - salaries do not serve as a counterbalance ent intervention programs for improv - ers and educational leaders through to these forces. Third, salary schedules ing failing schools. Simply knowing new approaches to pre-service that pay teachers the same across fields the actual dropout rate, for example, education, in-service professional also result in much greater difficulty could enhance parental involvement in development, due-process, evalua - staffing some jobs than others. In partic - district governance and provide better tion, tenure pro-cesses, and compen - ular, there tend to be shortages in fields information upon which parents might sation; and with greater outside occupational op - base schooling decisions for their chil - G experimentation with alternative portunities such as science, in fields that dren. California currently lacks the ca - training, induction, development, and require greater training such as second pacity to take any of these steps. evaluation of educational leaders. language learning, and in those that In addition to strengthening data Other policy areas are worth explor - have particularly specialized work re - collection and information manage - ing because of their evident impor - quirements, such as special education. ment, state programs and policies need tance. This would include among to be implemented in ways that allow others: California is incapable of effective system for systematic evaluation. Teachers G enhanced curriculum and instruction learning and continuous improvement, both and principals need access to networks for improving reading comprehension because it lags other states in the development through which they can learn about ef - G improved instruction of English lan - of a longitudinal student and teacher data fective policies and implementation guage learners, and system and because it has not developed challenges. And actors at all levels of G effective approaches for helping con - sufficient analytical capacity. the system need the flexibility to inno - tinuously failing schools. It is almost impossible to think of sys - vate, learn from experience and im - temic performance improvement in prove their practice. Building an information-driven system, focused California without dramatic changes on developing and disseminating knowledge in the state’s approach to information How can dollars be used about effective practices, is the fulcrum for development, use and dissemination. more effectively to meet continuous improvement at all levels. California’s current system is charac - outcome goals for students? The evidence base about how best to act in terized by many unconnected data col - Approaching a more effective system the areas identified in the forgoing sections lections within the Department of of public school governance and finance. is often thin, and the issues are compli - Education as well as important data, There is no silver bullet for school cated. In this context, it is important that

September 2007 | School Finance and Governance in California | 5 Figure 2 Percent of Students Participating in Subsidized School Lunch • First, often the academic goals set by Program and API, K –5 and K–6 Schools, 2004 the state for students are substantially 1,000 higher than current student outcomes. As Figure 2 illustrates, few high 900 poverty schools reach the state’s 800 API goal. In such a situation, there x e

d 800 may be very little information avail - n I e

c able about how to achieve such goals n a 700 m and thus the dollars or resources r o f r needed for success. If we do not know e

P 600 c i how to achieve a given level of student m e

d performance, we cannot estimate the a

c 500 A cost of attaining that goal. Second, districts and schools differ 400 in their capacities to transform re - sources into achievement - say, because 300 0% 20% 40% 60% 80% 100% of differing leadership skills or ability Percent of Students in School Lunch Program to use of information effectively. Thus, it is difficult to ascertain what re - sources are actually necessary for any given outcome. whatever California does be undertaken in G independent evaluation of programs A third difficulty in determining the a way that we can rapidly and systemati - and policies based on state data and amount of funding necessary to achieve cally learn what works, what doesn’t, and other data sources, a given outcome stems from substantial why. Too many times in the past, we have G support for the creation of networks differences in needs across districts. pursued initiatives that appeared promis - of schools or districts, or networks of These differences come from variation ing only to be deeply disappointed by the school leaders to allow for sharing of in the student population served as well ultimate results. For example, the 1997 information on effective practices, as variation in local labor markets for class size reduction program is currently and teachers and administrators. funded at $2 billion per year, yet there is G local capacity building to promote A fourth factor confounding estimates scant evidence that it has been effective in data-driven knowledge generation and of resource needs is that estimates are improving student outcomes. It continues use by school principals and teachers only applicable to the existing gover - in part because we have not developed ap - inform instruction and practice. nance and finance arrangements. As propriate systems to analyze its effective - Plans to expand and improve such, the dollars necessary to reach a ness and to make decisions based upon California’s data systems are underway given outcome goal under current insti - evidence as opposed to hunch. but these need to be deepened and accel - tutional arrangements may be quite dif - Producing dramatic improvement in erated. When better data are combined ferent from what would be needed student outcomes will require innova - with purposeful policy implementation under other conditions. tion and the creation of an information so that the effects of policies can be care - Finally, any estimate of resource infrastructure that will support continu - fully evaluated, our understanding of needs depends on the state of our cur - ous improvement. This would require: policy impacts can improve quickly. rent knowledge about effective school - G better data on student performance, ing. Innovations in curriculum or linked to teachers, schools and dis - To what extent are additional instruction, for example, may reduce tricts to facilitate better policy and resources necessary to meet the cost of achieving a given goal and, program choices, our goals? in some cases, investments in research G policies and programs implemented Determining how much and under what and development may be a better use of in a manner that allows for rigorous circumstances schools need additional funds for improving outcomes than ad - assessment of effects on students, resources is a complex task. ditional dollars for current instruction. G infrastructure at the state-level for in - Estimating, with any degree of cer - Not withstanding these analytic dif - formation collection, evaluation, and tainty, the resources districts need to ficulties in estimating the resource knowledge dissemination, meet state goals not an easy task. needs associated with achieving a

6 | School Finance and Governance in California | September 2007 S S

Figure 3 District API and Spending per Pupil 2004 –2005 Data from Imazeki (2007) given outcome goal for students, it is • useful to consider the results of various approaches that are designed to inves - 1,000 x e d

tigate how increases in school re - n I 900 e c

sources might affect student outcomes. n a

m 800

Our investigations included the per - r o f r

spectives of each of the commonly em - e

P 700 c ployed methods that have been i m e developed to study school finances in d 600 a c other states plus extensions of those. i A 500 The data available on spending and 5,000 7,000 9,000 11,000 13,000 15,000 achievement in California schools is Dollars per Pupil not sufficient for assessing the effects of dollars on student outcomes. The rela - tionship between dollars and student achievement in California is so uncer - tain that it cannot be used to gauge the three consecutive years. The study mates are not far from the estimates of potential effect of resources on student showed clearly that schools with simi - the effects of resources found in rigorous outcomes. Figure 3 illustrates this lar resources have very different stu - studies of resources such as class size re - point. It plots district API scores as a dent outcomes. If additional dollars duction in other states. function of per pupil spending in 2004- were inserted in the current system Sonstelie (2007) provides an illustra - 2005, and finds essentially no relation - there is no reason to expect substantial tive estimate of the dollars needed for ship between the two. There are a increases in student outcomes related each school in the state to achieve an number of possible reasons for this lack to state goals. 800 API, but this requires extrapolat - of relationship. One is that the data are Professional judgment models provide ing well beyond extant data and his too poor to find a relationship even if suggestive, but not conclusive, insights survey results because currently few one does exist. For example, there are as to the cost of improvement. Another high poverty schools produce API likely to be differences in learning quite different way to probe the re - scores above 800. When he limits the across schools and districts that the sources needs of schools is to ask profes - school budgets to those those situa - simple API measure does not capture. sional educators directly. Because they tions presented to his professional Alternatively, because of the inefficien - are in schools and understand school de - judgment panel, his respondents sug - cies in the governance system, there cision-making processes, they often have gested an estimated a 40 percent in - may, in fact, be little relationship be - informed ideas about what factors most crease from the total expenditure of tween dollars and student outcomes. If help or hinder student achievement. the same districts would still leave 50 this were the case, more money in the When asked how they would allocate percent of elementary schools with current system without significant re - resources to improve student outcomes, APIs less than 800. Five percent of ele - forms would be unlikely to result in superintendents, principals and teachers mentary schools would have predicted students meeting challenging state stan - are generally optimistic that additional scores of less than 736. The estimates dards. Finally, differences across dollars if allocated in specific ways can from a parallel and more traditional districts in factors such as the concen - improve student outcomes. But even professional judgment study con - tration of poverty might mask any ef - these professionals note that the rela - ducted as part of the project are simi - fects of resources but existing data do tionship is not strong. For example, on lar to those of Sonstelie. In other not permit separating this out. average across respondents, an elemen - words, even California teachers and An examination of “Beating the tary school of all poor students with a administrators, who might be expected Odds Schools” supports the conclu - school-level budget of $4000 per-pupil to be quite optimistic about the role of sion that dollars alone do not explain (on top of additional set resources allo - resources, estimate that adding re - learning differences across districts. cated by the district) is estimated to pro - sources alone within the current struc - One study investigated schools that duce an API of approximately 698. An ture of schools has only a small were beating the odds insofar as they increase in the school’s budget by positive effect on student outcomes. significantly outperformed expected $1,000 per pupil increases this predic - For schools in high poverty communi - student performance results for at least tion only by 13 API points. These esti - ties to reach California’s high student

September 2007 | School Finance and Governance in California | 7 achievement goals likely requires new Instead California would benefit of whatever resources are available?” approaches and a system that supports from a policy environment that recog - California is so far from achieving its continuous improvement. nizes the complexity of the task and student outcome goals that marginal the limited state of our knowledge. It policy changes are unlikely to produce Any transition to a new system will initially would focus on reforms that improve the desired outcomes. Instead such require additional resources. the ability of decision makers at all lev - progress requires a new approach to Removing and replacing the dysfunc - els to make good decisions for students reform, an approach that allows state, tional elements of the current finance and to improve outcomes. Such a sys - district and school decision-makers to and governance system requires sub - tem would improve the alignment be - improve their practice and thereby to stantial changes in programs, person - tween the accountability system and enhance the opportunities afforded nel development, and management the decision-making responsibilities, California’s students. structures. Getting from here to there increasing flexibility at the local level. The message of the entire collection cannot be done effortlessly. The neces - It would improve information collec - of studies is that fundamental changes sity for strong political leadership by tion, both at the state level where data will be needed if California is to pro - the governor and the legislature is ob - should follow students over time and vide a high quality school system. vious. But, it will undoubtedly also re - link them with the resources they re - Some changes are easier than others. quire dedicated funds to pay for the ceive and at the local level where net - Some changes are more appealing than introductory phases of new systems works of teachers and administrators others in that they entail less funda - and operations and for the withdrawal could learn from each other’s experi - mental challenges. But picking a small from other programs. Even if the new ences. It would refine policies to at - subset and ignoring the others most system requires no more resources tract and retain high quality teachers likely will have few benefits. than currently being spent, the transi - and administrators, and place a prior - California’s economy is dependent tion would require temporary transi - ity on learning from the effects of the upon the strength of its workers. If tional spending. This funding is policies it implements. It would sim - California students are going to partic - logically distinct from longer run deci - plify its school finance formulas so ipate fully in its future development, sions about the level resources em - that similar districts would be treated they will need quality schools that are ployed to operate California’s schools. similarly and differences across dis - competitive with those in other states tricts would be treated reasonably and and other nations. Without better Conclusions consistently. It would also target re - schools, the future vitality of If our set of studies has one overarch - sources to improve the outcomes of California will depend on its ability to ing conclusion, it is simply this — students in poverty, most of whom are attract workers from elsewhere. A fail - California’s school finance and gover - unable to reach state goals in the cur - ure to act now, abrogates our public nance system are fundamentally rent system. And for all school dis - responsibilities. Ultimately, we fail our flawed. Consequently, California stu - tricts, it would make the state children, our families and the future of dents perform far lower on tests of budgeting process more predictable, our state. achievement than do students in other removing the peaks and valleys in an - states. Within the state, schools with nual appropriations, and establishing Susanna Loeb is an Associate high proportions of students in poverty distributional decisions earlier in the Professor of Education and Director are consistently failing to meet the stan - spring so that school and district lead - of the Institute for Research on dards the state sets out for them. ers could be more strategic in deter - Education Policy & Practice at No one program or intervention will mining how best to use their resources Stanford University. fix the system. California has tried for the next academic year. Anthony Bryk is the Spencer over and over the approach of intro - Finally, we cannot emphasize Foundation Professor of Organiza- ducing separate programs and dis - enough that asking the question, “how tional Studies in Education and jointed new policies. Although each much money will it cost to achieve Business at Stanford University. may have been well intended, the ag - State goals for students?” is meaning - Eric Hanushek is the Paul and Jean gregate mass is now a large part of the less without also asking “how can we Hanna Senior Fellow at the Hoover problem that needs attention. develop a system that makes better use Institution of Stanford University.

i The common nomenclature for the approaches and their associated GDTF studies are: cost function (Imazeki 2007); successful schools or beating the odds schools (Perez et al. 2007); professional judgment approach (Chambers, Levin, and DeLancey 2007; Sonstelie 2007).

8 | School Finance and Governance in California | September 2007