Kansas Policy Review
Current Issue
Vol. 29, No. 1 Spring 2007

Printable Version (PDF)

Editor’s Comments
Joshua L. Rosenbloom

Baby Boomers and Immigrants on the Range: Population Trends in Kansas
László J. Kulcsár

The State of Innovation in Kansas
Joshua L. Rosenbloom

Sizing up Kansas Public Finance
Glenn W. Fisher, H. Edward Flentje, W. Bartley Hildreth, and John D. Wong

Remarks from Keynote
City-County Consolidation: Reshaping the Local Government Landscape

Suzanne M. Leland

KPR Home Page

KPR Archive

IPSR Home Page

What will it Take to Make Kansas School Funding "Cost Based"?

Bruce D. Baker


As discussed in the article by Preston Green in this issue, on January 3, 2005 , the Kansas Supreme court found the legislature’s current school finance law in violation of Article 6 (the education article) of the Kansas Constitution in two ways. First, the current system fails to guarantee that districts have adequate resources as measured against the legislature’s own study of cost. Second, the ways in which vastly differing amounts of resources are distributed across Kansas school districts are "politically distorted" and "not cost-based." The implication of this latter concern is that for any legislative remedy to be considered constitutionally "suitable," that remedy will have to be shown to be cost-based or at least no longer politically distorted.
In the aftermath of the January 3, 2005 Supreme Court decision, most media, legislative and judicial attention has been paid to the aggregate shortfalls in current funding when compared with the legislature’s own cost study released in 2001. That shortfall was estimated at approximately $853 million. Much less attention has been paid to the question of political distortion and cost basis. This article will explain what is meant by the court’s declaration that the current state school finance system is politically distorted and not cost-based . Further, it will explain how those political distortions came into being.

How did the School District Finance Act Become "Politically Distorted"?

In this modern era of increasingly complex and state legislatively controlled school finance formulas there exist new opportunities for legislators to either improve dramatically or erode the rationality and cost basis of state aid allocations to schools. Throughout the 1970s and 1980s in Kansas and elsewhere, the politics of state aid to local schools centered primarily on the level and mix of taxes to be used for funding schools, and the basic level of funding to be put into the system. The political dynamic has changed with the proliferation of need-based or cost-based aid formulas that include a multitude of weightings or cost adjustments that may benefit some districts and not others and that have differing affects on total costs to the state.

For example, Baker and Duncombe (2004) show how the difference in urban/rural balance between Kansas and Texas has led to a very different balance in these two states’ funding formulas, with Texas having a far greater share of its population and political power in major urban areas (63% compared to Kansas’ 32%). Both state’s aid formulas include similar basic structures district size weights, student poverty, and bilingual education weights. But, in Kansas , the small district size weighting dwarfs the student need weightings, driving over 10% of general funding, with poverty-based weighting driving only 2% and bilingual program weighting driving less than one-half of 1%. In Texas, small district size and sparsity weights drive under 3% of funding while compensatory weightings drive nearly 10% and bilingual programming over 1%. In addition, the state aid formula includes a teacher wage index that tends to drive a greater share of aid to urban areas, and in total, drives over 10% of aid. 1 Baker and Green (2005) show how legislators in Kansas and Alabama have created cost/need adjustments that specifically disadvantage minority students and do so to a greater extent than in any other states. 2

Some distortions were introduced into the School District Finance act (SDF) from its inception in 1992, while others emerged over the next decade as legislators became more familiar with the new system. The original distortion of SDF occurred in the calculation of the small district adjustment. That calculation was tied to the historical spending of those districts which had (a) been provided since 1965 with more state support per pupil, and (b) allowed to grow their budgets annually at a much faster rate.

Table 1 summarizes the data underlying the low enrollment weight calculation. In addition to the effects of state aid and budget growth limits, it also turns out that the very smallest districts that were spending 2.14 times what larger districts spent possessed 3.25 times the taxable assessed value per pupil of larger districts. Similarly, districts with 200 to 400 students, which spent 1.58 times the amount spent in larger districts, possessed 1.76 times the taxable property wealth per pupil.

In testimony presented at trial in Montoy v. Kansas in the fall of 2003, Baker showed that removing transportation costs and controlling for differences in taxable wealth (and income) and administrative expenditures leads to a reduction in maximum weight from 2.14 to 1.75 for the district with 100 pupils. 3 It turns out however, that the most significant distortion in the "low enrollment" weight results from the fact that legislators drew straight lines rather than a more sharply dropping curve through their points. By alternative estimates district with 600 to 1,000 students that presently receives a weight of 1.41 would have received only 1.19 had a more appropriate curve been fit to the points (Baker, 2003). The effect of this "error" is that many medium sized Kansas districts, often adjacent to much larger districts received a substantial boost in funding (for example, Piper USD, carved out of the city limits of Kansas City has historically received about $800 in revenue per pupil from this adjustment). Arguably, it was partly the exaggerated size of the low enrollment weighting that led to sufficient support among key rural legislators to pass SDF. 4

In addition to errors made in establishing the low enrollment weighting, transportation weighting was added as an independent component of the new formula, despite the fact that general fund expenditures per pupil used in calculating the low enrollment weight included expenditures on transportation.

While not necessarily plain common sense, it is relatively well accepted in the field of school finance that economies of scale do exist across districts and schools of varied size, typically leveling off for districts with greater than 2,000 students. In addition, low density, large land area districts face higher per pupil transportation costs. 5 Yet, it became obvious to some Kansas legislators that these adjustments were overstated, so obvious that one of the first modifications to the new formula was a "large district" weight called correlation weight. Correlation weight was specifically intended to offset the first few percentage points of low enrollment adjustment, and eventually to level large districts up to no less than 6.32% above basic aid. This meant that by 2003, when base aid per pupil had climbed from $3,600 in 1992 to $3,863, the actual minimum funding available to a large district would be 6.32% above $3,863, or $4,107.

Among the state’s larger districts, different political divisions emerged around weighting factors. Fast-growing suburban districts were able to successfully argue that they faced substantially greater costs to their annual operations from opening new school facilities. Representatives of these districts negotiated a 25% adjustment per pupil for each pupil attending a new facility in its first two years of operation, if and only if that district had (a) already been able to garner the local support to construct new facilities, and (b) maximized all other local options for raising tax revenues. 6

The policy specifically favored Johnson County districts in the Kansas City metropolitan area. Legislators did little to hide their targeting preference when in 1996-97 they increased the weight from 25% to 33% for two districts only, Blue Valley (USD 229) and Olathe . Targeted by name or not, the weight was designed to serve specific suburban districts, and succeeded in doing so almost exclusively. Baker (2003) notes: In 1995, Blue Valley and Olathe alone received 64% of all new facilities aid, and with their neighbor Shawnee Mission (discussed in the previous example), the three received 82% of all new facilities aid. 7 In more recent years, a related provision was added, called ancillary new facilities adjustment, which would allow districts already maximizing all other local taxing options and receiving new facilities aid to request additional general fund tax authority.

Table 2 shows the collective effects of formula distortions on Kansas City area districts in 2004–05. Table 2 specifically addresses the General Fund Budget authority (legal maximum) of districts. Table 2 uses weighted pupil count data from school year 2004–05. The general fund weighting ratio represents the number of weighted pupils allocated to each district as a cumulative function of district size adjustment, transportation adjustment, new facilities and ancillary new facilities, at risk pupils, bilingual and vocational program contact hours. A 1.29 ratio implies that the district would receive 1.29 x $3,863 = $4,983.27 per pupil in cost adjusted general funds, while a ratio of 1.23 would generate $4,751.49. If we assume state defined general fund authority to reflect differences in district need, then the implication of FY 2005 funding for Kansas public schools was that Piper school district had, by far, the greatest needs among districts in the Kansas City metro area. Further, Blue Valley ranked second and DeSoto third, each with greater implied need than Kansas City . In 2003, Olathe , Blue Valley and DeSoto had received as much as 10% more need adjusted general fund budget authority as Kansas City. 8

In short, Table 2 shows that the Kansas Legislature by adopting cost and need adjustments based solely on majority vote in the political process without serious consideration of actual costs and needs has created a system of school finance that directly (rather than indirectly via local control) favors districts with lower poverty rates and far fewer minority students. Indeed it may be too high a standard to imply that all legislative decisions of this type be based on the best possible empirical evidence. However, a particularly relevant question in the current political and legal context is just how far out or line, or just how "politically distorted" must a school finance system become before judicial intervention is warranted, or required? Further, if judicial intervention is warranted, how and to what extent must the distortions be alleviated?

How Much More and Where Should it Go?

In this section, I evaluate the funding deficits under current (at the time of the January 3 ruling) relative to two alternative empirically-based benchmarks of costs. First, I construct a simulation of core components of the cost study prepared for the legislature by the consulting firm of Augenblick & Myers of Denver, Colorado (referred to as AMSB). 9 Second, I construct an alternative simulation of core components of the AMSB study, using marginal cost adjustments for economies of scale and student needs drawn from more rigorous recent research.

Note that in the following analyses, I focus specifically on school districts’ General Fund Budgets, pulling out certain components of General Fund Budgets for the purpose of making apples to apples comparisons between actual funding and estimated costs. In Kansas school finance terminology, a school district’s general fund budget is the core component of a district’s annual operating budget, and the general fund budget authority for each district is defined by the School District Finance act. In short, the General Fund Budget is what the state aid formula guarantees that each district shall have available for paying teacher, administrator and support staff salaries and benefits, keeping the lights, air conditioning and/or heat on, keeping the halls clean, and paying for instructional and non-instructional materials and supplies. The Supplemental Fund Budget, which may be used to generate additional annual operating revenues, is derived from local option taxes (Local Option Budget, or LOB). Programs for children with disabilities are partially financed by the state through a separate mechanism, but one which interacts with the general fund in that each child with a disability also generates regular education funding. Separate funds are used for generating revenues for capital outlay and bonded indebtedness.

In this same volume, Lori Taylor of Texas A&M has discussed education cost analysis and the specific application of cost analysis to Kansas and estimating the cost of a constitutionally adequate education. Studies of the cost of education attempt to measure both the bottom line of spending required in a state to meet constitutional mandates, as well as the ways in which costs vary across school districts based on a number of district characteristics and student population characteristics outside the control of local school officials. That is, education cost analysis identifies a basic underlying cost, usually either the cost of achieving a given set of outcomes in a scale efficient district with relatively few additional costs and needs (low to average teacher labor market costs and few special needs students), or the cost of achieving a given set of outcomes in a district of average characteristics. Education cost analysis also provides guidance on the additional amounts of funding needed for marginal costs of achieving the same set of educational outcomes in smaller school districts, districts in areas with higher competitive wages, districts with higher concentrations of at risk and limited English proficient students. 10

Some background on the AMSB study is warranted. The study’s origins date back to 1999 and a report issued by the Vision 21st Century Task Force convened by Governor Bill Graves. One of several working groups of citizens and legislators, the task force on public school finance recommended to the legislature:
To date, no one has defined what constitutes a suitable education in Kansas. Therefore, it has been impossible to put a price tag on it. When the current school finance formula was drafted, cost figures including the base state aid of $3,600 per pupil and the various pupil weightings were derived primarily from political deliberation. The Task Force concluded that it is of critical importance that the first step toward public education finance reform in Kansas is to conduct a professional evaluation to determine the cost of a suitable education. (p.1.)

Kansas legislators responded by appropriating funding for the study and collaborating with the Kansas State Board of Education on developing operational definitions of suitable provision , tied to current statutes prescribing curricular programs consisting of the subjects and courses required under the provisions of K.S.A. 72-1101, 72-1103 and 72-1117. Members of the Legislative Coordinating Council then oversaw the study through its implementation and completion.

One component of the study involved convening professional judgment panels to identify the inputs necessary to deliver a suitable education, while another component involved calculating the average expenditures of districts presently meeting required outcome standards. In their final recommendations, AMSB suggested using a base cost figure close to their findings of average expenditures of successful districts—$4,650— but using marginal cost estimates from professional judgment analysis, because successful district analysis provided no guidance on marginal costs.

In Montoy v. Kansas, as discussed by Preston Green in this volume, the court has relied almost exclusively on the AMSB study as the basis for declaring that the total amount of funding available is insufficient. The AMSB study also addresses how that funding should be distributed across districts by costs and needs, and the court has declared that current funding is not distributed by costs and needs. One can only assume that the court believes that the legislature should not just meet the total spending target of AMSB, but should distribute that funding using AMSB district level cost variations as a guide. The court, however, has failed to articulate distributional benchmarks thus far. In their January 3, 2005 decision, the court indicated that more money would be necessary but that there were literally thousands of ways to meet the constitutional mandate. In June, 2005, the court indicated that not enough new funding was added in the spring of 2005, and ordered an additional $142 million but did not specify where that money should be allocated.

The justices reasoned that the $142 million was an appropriate amount because, combined with the first $143 million added in the spring, the total would achieve the first one-third of meeting the costs proposed in the AMSB study ([142 + 143]/853 = .33). This assumes that 100% of each of the first two allotments was allocated toward closing existing gaps between existing district budgets and district needs estimated by AMSB. Surely the legislature would not have accomplished one-third of the goal had all of the new money been allocated only to small, rural, districts or wealthy suburban ones. Yet, the court did not specify as much.

For the comparisons in this section, I remove from current school district General Fund Budgets, transportation and vocational education because these costs were not estimated directly in the AMSB study. Drawing on AMSB recommendations, I assume a basic annual operating cost of $5,420 per pupil in 2004, 11 similar to the figure used by the court in reference to the cost study. Using this base cost figure results in an aggregate shortfall of general fund budgets totaling $851 million when compared against similar budget components in 2004-05. The total cost in 2004–05, across all districts, of General Fund Budgets excluding transportation and vocational education was approximately $2.14 billion and the total cost of funding AMSB recommendations for the same formula components (base aid, size adjustment, at risk and bilingual education adjustments), was approximately $2.99 billion.

To simulate AMSB district level costs, I fit a curve (rather than using straight line segments as recommended by AMSB) to their cost estimates for districts of varied size. AMSB estimate the marginal costs for the smallest districts at between 40% and 50% above base costs, a much smaller margin than in the current law (101.4%). AMSB estimate weights to be applied to school district subsidized lunch counts (free only) ranging from 33% in a small district to 56% in a large district, and weights for bilingual education programs ranging from 61% to 103%. 12 These are much larger than current adjustments.

Studies that rely on professional judgment panels to prescribe resource configurations can lead to irregular findings regarding marginal costs associated with variations in district size and student needs. 13 As an alternative, I use a method recently proposed by Bifulco (2005) to adjust for economies of scale and student needs. 14 This method uses a 3rd order polynomial curve, which represents aggregate findings of costs associated with economies of scale across several states, to adjust costs for differences in district size. 15 Bifulco also uses student need weights derived from education cost function analysis by Duncombe and Yinger (2005). 16

Duncombe and Yinger (2005) show that the marginal costs of achieving comparable outcomes (to the average student) for limited English proficient (LEP/ELL) students is approximately 103% and for children in poverty, as measured by the U.S. Census Bureau, about 149%. When child poverty is alternatively measured by subsidized lunch rates, including those qualifying for free or reduced price lunch, the marginal cost per child in poverty declines to 109%. For alternative cost estimates, I apply Duncombe and Yinger’s LEP/ELL weight of 1.03 to districts’ actual LEP/ELL student populations, rather than bilingual contact hours. In Kansas, LEP/ELL counts are typically 2 to 3 times the reported full-day bilingual contact hour counts. Next, I apply Duncombe and Yinger’s 1.49 weight to the percentage of public school enrolled children between the ages of 5 and 17 from families living in poverty.

Table 3 compares district general fund budgets in 2004–05 with the two alternative cost estimates by district size and district subsidized lunch quartile. The cost comparisons show that all districts are under-funded by comparison to cost targets. More importantly, however, the table shows that the shortfalls for some districts are much greater than the shortfalls for others. For example, from the lowest to highest poverty quartile among scale efficient districts (measuring poverty by rates of subsidized lunch—free only), 2004–05 general fund budgets climb by about $145 per pupil or 3.3%. In contrast, Augenblick and Myers estimate $1,114 per pupil difference in need, or 18% difference in cost of outcomes between these groups. When applying adjusted weights, the marginal costs of districts in the highest poverty quartile are $1,200 per pupil, or 20.5%.

Among the lowest poverty districts, the current marginal cost difference between smallest and largest districts is about 56%. AMSB places this difference at 29% and alternative estimates place the difference at 38%.

Some clarification is in order. How can it be that the system is funded so inadequately across the board, if, as articulated by the state at trial, most Kansas children are doing well? Recall that General Fund Budgets represent only the underlying guaranteed annual operating funding by the state. Both the lower court and higher court have made clear that the state must guarantee a level of funding that meets the constitutional mandate, and that other funds, like those raised from local option taxes are for the extras.

It can certainly be shown that many of the state’s districts with more advantaged student populations like those in Johnson County are currently spending at sufficient levels to achieve constitutionally adequate student outcomes. While not an apples to apples comparison of similar budget components, Blue Valley (USD 299) reported current operating expenditures per pupil in 2003, of $7,236 per pupil (including special education, transportation, vocational). 17 Blue Valley’s estimated need excluding special education, vocational programs, and transportation was $5,471 (A&M) to $5,640 (A&M with adjusted weights). In contrast, Dodge City ’s estimated need ranges from $9,144 (A&M) to $9,225 (adjusted weights) excluding special education, vocational programs, and transportation. Yet Dodge City ’s actual current operating expenditures including special education, vocational programs, and transportation were $5,615 in 2003.

Are Funding Deficits Related to Outcome Deficits?

In the wake of controversial state high court decisions in states including Kansas, New York, and Arkansas, and the increased role of cost studies in these cases, there has been increased scrutiny on the reliability and validity of cost analysis methods and findings. In particular, it has been argued by some that cost analysis methods are so unreliable that measured spending deficits, like those addressed in Table 3, are entirely unrelated to actual outcome deficits. 18 This brief section looks at the relationship between current funding shortfalls and current outcomes on state assessments.

Figure 1 provides a visual analysis of the relationship between funding gaps, relative to AMSB simulated costs, and the percentages of students scoring proficient or higher on state math and reading assessments over the past 5 years (2000 to 2004). The horizontal axis indicates the ratio of current (2004–05) General Fund Budget to AMSB cost estimates (as simulated herein). Only districts enrolling greater than 2,000 students are included. Sizes of the circles and triangles in the graph represent the relative enrollment of districts, with the largest circle and triangle representing Wichita, which sits just above 70% of estimated cost and just below 50% on both reading and math proficiency. In contrast, districts’ with General Fund Budgets exceeding 80% (such that with LOB’s they may exceed 100%), have proficiency rates at and above 70%. A positive relationship exists, whereby districts with closer to 100% of their cost estimate (in general funding alone) have higher proficiency rates.

Table 4 provides a simple statistical test of the relationships between funding deficits (percent below estimated need) and state assessment proficiency rates. Funding deficits are measured with respect to both AMSB simulated costs and AMSB with alternative weights (inverse of the measure in Figure 1). In the case of reading assessments, funding deficits alone explain more than half of the variation in outcomes, with a 1% increase in deficit associated with a nearly 1% decline in proficiency rate. Funding deficits alone explain one-fifth to nearly one-third of variance in math proficiency rates.

To some extent, these analyses convey the obvious. Districts with more children in poverty and more limited English speaking children have fewer children scoring proficiently on state tests. AMSB cost estimates among large districts differentiate primarily on the basis of those student needs, providing higher estimates of cost to districts with more children in poverty and more limited English speaking children. Among large Kansas districts, those with more children in poverty and more limited English speaking children receive ostensibly the same need adjusted general funding as districts with far lesser student needs.

Measuring Progress in the Reducing Political Distortions

In this section, I measure the progress made by the legislature in the spring and summer legislative sessions that resulted in a number of incremental changes to SDF. I begin with a review of the changes made to General Fund Budget allocations by the end of the summer 2005 special legislative session on school finance. Next, I focus specifically on the question of whether and to what extent the legislature has, through the incremental changes made thus far, reduced distortions and increased the relationship of actual aid distribution to cost estimates. For this analysis, I evaluate the correlation across districts between formula aid allocations for General Fund Budgets and the two alternative cost estimates—AMSB simulated costs and AMSB costs with alternative weights. Because correlation analysis addresses only the rank order of districts and not whether higher poverty districts have enough additional support, I also calculate regression coefficients for both cost estimates and for General Fund Budgets before and after incremental adjustments.

By the close of the spring 2005 regular session, the legislature had adopted House Bill 2247, which, among other things raised base aid per pupil to $4,222. This increase was accomplished in part by folding the previous correlation weight into the base, such that the increase was actually from $4,107 to $4,222 and not from $3,863 to $4,222. Correlation weight no longer existed. At risk weighting was increased from 10% of $3,863 per child to 14.5% of $4,222 and bilingual program weighting increased from 20% x $3,863 per 1 full time equivalent student to 39.5%. Of the $143 million package, about $115 went to funding adjustments to General Fund Budgets that would move those budgets closer to cost targets. Additional policy changes included increased funding to offset special education costs. Other provisions, included the hotly contested cost of living adjustment are addressed in a separate article in this volume.

By the close of the summer special session on school finance, additional incremental changes had been made to some of the core components of the formula. Base aid per pupil was increased from $4,222 to $4,257. Correlation weight was reintroduced to chip away further at the margin of difference between small and large districts. Bilingual weighting was left at its HB 2247 level, but at risk weighting was increased to 19.3%. These changes added approximately $73 million directly toward the goal of closing gaps between current general fund budgets and AMSB targets. This suggests that as much as $75 million of the $148 million summer legislation funded areas not directly related to reducing distortions and improving adequacy. 19

Table 5 shows the correlations between General Fund Budgets in 2004–05, estimated General Fund Budgets under HB 2247, estimated General Fund Budgets following the special session, and the two cost estimates. As in previous analyses, the comparable components exclude transportation, vocational education, and special education. Across all districts, correlations between general funds and cost estimates were high to begin with, because smaller districts are estimated to have higher costs and smaller districts receive significant cost adjustment. Correlations edged upward slightly with the passage of HB 2247 and subsequent modifications. Across scale efficient districts, correlations were much lower to begin with, but also show incremental improvement with the passage of HB 2247 and subsequent modifications. Perhaps most importantly, among the state’s largest districts, substantial improvements were achieved in alignment between aid allocations and estimated costs, with correlations between general funds and AMSB base with adjusted weights increasing from .342 to .721.

Table 6 addresses the changes in the magnitudes of the relationship between student needs and allocated funding. In 2004–05, across districts enrolling over 2,000 pupils, a 1% increase in poverty was associated with under $17 additional per pupil revenue, where the maximum Census poverty rate is approximately 22% (about one-third the maximum subsidized lunch rate). Using AMSB need weights, a district with 1% more children in poverty would have, on average $105 more in per pupil revenue. Incremental progress was made toward this goal from spring 2005 to summer 2005, bringing the slope to nearly $35 per 1% difference in poverty rate. The case is similar for actual funding and marginal costs associated with limited English proficient children. With the doubling of the bilingual program weight, the slope of the relationship between limited English proficient student shares and funding doubled. Nonetheless, the need adjustment still falls well short of estimated marginal costs.

Issues for the Coming Legislative Session

The near future will see the release of findings from two additional studies of the cost of a suitable education in Kansas , both under the supervision of the Kansas Legislative Division of Post Audit. As discussed by Lori Taylor in this volume, one of the two studies will attempt to prescribe research-based prototypes for different size Kansas districts serving different student populations. The other, will involve education cost function estimation of the cost of achieving outcome levels mandated by the Kansas State Board of Education. The outcome-based, education cost function analysis has been subcontracted to William Duncombe and John Yinger of Syracuse University, who have published extensively on this methodology.

Thus far, the Kansas Supreme Court has relied on inflation adjusted basic costs drawn from AMSB recommendations, which paralleled their successful schools cost estimate. Taylor , in this volume shows that the successful schools cost estimate produced by AMSB for Kansas is 4th lowest of 39 studies across states since 1995, adjusted for inflation and regional variation in competitive wages. Further, as shown herein, marginal costs associated with size, poverty, and limited English proficiency were lower in the AMSB study than in other recent studies and reports. As such, it is unlikely that Kansas legislators will see substantial cost savings produced by the new studies, or that the studies will indicate lesser need to distributed additional funds more aggressively to high poverty, large districts.

If the court expects new legislation in the spring 2006 legislative session to mitigate past political distortions in addition to leveling up the system, the court may need to be more explicit regarding not only the level but proposed distribution of any new appropriations across districts. The court will have two new studies to consider. However, such direction from the court will likely be met with much resistance by legislators, because disproportionate allocation of new money to the state’s highest need districts is clearly in conflict with the present and recent balance of voting preferences of legislators, even while under recent court oversight. 20 The Supreme Court has indicated that voting preferences alone are insufficient rationale for funding differences across districts and children. The Supreme Court has set a relatively high standard that the legislature should provide a cost basis for their policies. While a precise, empirical cost basis may be an elusive standard for each and every component of a state’s school finance formula, a reasonable standard and fair remedy for all Kansas children likely lies somewhere between precisely cost-based and purely political.


1. Baker, B.D., Duncombe, W.D. (2004) Balancing District Needs with Student Needs: The Role of Economies of Scale Adjustments and Pupil Need Weights in School Finance Formulas. Journal of Education Finance 29 (3) 195-222.

2. Baker, B.D., Green, P.C. (2005) Tricks of the Trade: State Legislative Actions in School Finance Policy that Perpetuate Racial Disparities in the Post-Brown Era. American Journal of Education 111 (3) 372-413. These legislatively created racial disparities are the centerpiece of an ongoing federal equal protection challenge brought against the state of Kansas in the federal district court of Kansas ( Robinson v. Kansas ).

3. Baker, B.D. (2003) Wide of a Reasonable Mark: Evaluating the Suitability of the Kansas School District Finance and Quality Performance Act. Testimony prepared on behalf of plaintiff districts in the case of Montoy v. Kansas . http://www.ku.edu/~bdbaker/Montoy.doc

4. Berger (1998) explains how State Senate Majority Leader Sheila Frahm, from rural Western Kansas, was a "driving force" behind the passage, during the veto session, of the house plan which included the low enrollment weight.

5. Andrews, Duncombe & Yinger (2002) Revisiting Economies of Scale in American Education: Are we any closer to consensus? Economics of Education Review.

6. Note that the add-on of 25% was to be applied to districts general fund, annual operating budgets and was not intended to support the construction of new facilities, which is handled through a separate fund for paying down bonded indebtedness or related capital expenses, which are handled through a separate local mill levy specifically for capital outlay.

7. p. 77.

8. Baker, 2003.

9. Augenblick, J., Myers, J., Silverstein, J, Barkis, A. (2002) Estimating the Cost of a Suitable Education in Kansas in 2000–2001 Using Two Different Analytic Approaches. Report prepared for the Legislative Coordinating Council. Kansas Legislature.

10. Of 39 cost studies reviewed by Taylor , Baker & Vedlitz, none raised the issue of additional operating costs associated with higher percentages of children attending school in new facilities.

11. The inflated value of AMSB recommended $4,650 based in 2001. This value is inflated by the Bureau of Labor Statistics Employment Cost Index. Note that AMSB never found the actual basic costs to be $4,650. Rather, their professional judgment analysis led them to a basic cost of $5,811 and their successful districts analysis identified an average expenditure per pupil across 85 districts at $4,547.

12. Baker and Duncombe (2004) and Duncombe and Yinger (2005) explain the importance of using the correct marginal cost weight with the correct need population count. Where poverty, for example, is counted by subsidized lunch rates, which often run two to three times higher than U.S. Census Poverty rates for the relevant population, the marginal cost of a 1% change in poverty will likely be lower. When used in school finance policy, policymakers should not choose to multiply this weight times the lower U.S. Census poverty rate. Rather, an alternative weight based on Census poverty counts would be estimated, and would likely be higher, producing comparable total adjusted funding. Similar issues apply to differences in bilingual program contact hour counts versus limited English student population counts, the latter in this case being preferable.

13. See Baker, B.D. (2005) The Emerging Shape of Educational Adequacy: From theoretical assumptions to empirical evidence. Journal of Education Finance 30 (3) 259-287.

14. Bifulco, R. (2005) District Level Black-White Funding Disparities in the United States : 1987 to 2002. Journal of Education Finance.

15. For economies of scale:
Cost Index = 10.7028 – 3.41946 (enrollment) + 0.40075(enrollment 2 ) - 0.01556(enrollment 3 ), which indicates that costs drop at a decreasing rate as one moves from the smallest districts to midsize districts, and are roughly equal for districts with greater than 5,000 students. See Also, Baker, B.D. (2005) The Emerging Shape of Educational Adequacy: From theoretical assumptions to empirical evidence. Journal of Education Finance 30 (3)259-287.

16. Duncombe, W.D, Yinger, J. (2005) How Much More does a Disadvantaged Student Cost? Economics of Education Review.

17. U.S. Census Bureau, Fiscal Survey of Local Governments (F-33) Public Elementary and Secondary Education Finances. http://www.census.gov/govs/www/school.html . Current Expenditures include: all expenditures except those associated with repaying debts, capital outlays (e.g., purchases of land, school construction and repair, and equipment), and programs outside the scope of preschool to grade 12, such as adult education, community colleges, and community services. Expenditures for items lasting more than one year (e.g., school buses and computers) are not included in current expenditures. Definition from: http://www.nces.ed.gov/edfin

18. Permission not granted to cite the source of these allegations.

19. At least by the narrow definition applied here, which focuses on raising and adjusting general fund operating budgets to move them closer to AMSB cost estimates. To other areas to which significant funds were added include special education, which the court has recognized as underfunded, and capital outlay matching aid, which the court has recognized as inequitable.

20. A recent example to this effect is that even during the summer 2005 special session, legislators saw fit to provide slightly more aid for property tax relief than the total increase for at risk programming. In spring 2005 session, legislators proposed a cost of living adjustment for districts with high housing unit values that would have generated an additional $265 per pupil for Blue Valley and comparable levels in Olathe and Shawnee Mission, while the additions to at risk and bilingual weighting would lead to an increase of only $202 per pupil in Kansas City .

About the author

Bruce D. Baker is Associate Professor of Teaching and Leadership at the University of Kansas . He is lead author of Financing Education Systems under contract with Merrill/Prentice-Hall, and author of over 30 articles since 1998 in journals including the Economics of Education Review, Journal of Education Finance, Educational Evaluation and Policy Analysis, and American Journal of Education. He has consulted for the Texas , Missouri , and Wyoming legislatures on the design of school finance policy and has served as an expert witness on school finance cases in Kansas and Nebraska.

Please send comments about the Institute’s WWW pages to pri@ku.edu
Copyright© 2006 PRI, University of Kansas, All rights reserved
Last updated: [an error occurred while processing this directive], [an error occurred while processing this directive]