Abstract

Several states and the federal government distribute aid for special education programs based primarily on total district enrollment and a fixed aid amount per student, a method called census funding. In this policy brief, we address three questions to help policy makers, educators, and researchers better understand census-funding models and special education finance policies in general. The first question is, what are the key advantages and disadvantages of census-funding models? The second and third questions relate to aspects of policy implementation, in the event a state legislature should choose to adopt the approach. First, we examine what options are available to mitigate concerns about the equity of funding under a census funding model. Second, we examine what other options exist for helping states and districts to contain special education costs while maintaining a high level of quality.

Introduction

The Individuals with Disabilities Education Improvement Act (IDEA) is the federal law in the United States that governs special education and guarantees students with disabilities access to a free and appropriate public education.1 Public schools in the United States must serve all school-aged youth, but only for youth with disabilities must they work with families to develop Individualized Education Programs (IEPs) and provide all special services and supports needed to fulfill those programs.2 The access provided for by IDEA improves education quality for students with disabilities but generally requires that school districts spend more money to provide an education for them than they do for other students.

Since the enactment of IDEA in 1975, the percentage of students served by special education programs has grown significantly. Figure 1 depicts the special education enrollment rate trend between 1991 and 2009, a period during which it rose from about 11 percent to 13.5 percent nationally. The overall trend in special education participation combines trends for thirteen distinct disability categories that do not all follow the same pattern.3 For example, autism is a less common disability but its rate has been climbing at an increasing pace. In contrast, learning disabilities account for nearly half of all students served by IDEA. Learning disability rates peaked in 1999 and subsequently declined, so the overall special education enrollment rate has also fallen slightly since 2004.

Figure 1.

U.S. Special Education Enrollment Trends, 1991–2009. Note: Percentage calculated by dividing all special education students ages 6–22 years by the total enrollment grades 1–12, plus ungraded students.

Figure 1.

U.S. Special Education Enrollment Trends, 1991–2009. Note: Percentage calculated by dividing all special education students ages 6–22 years by the total enrollment grades 1–12, plus ungraded students.

Disability-level trends have important implications for special education finance because the typical costs of educating youth with disabilities tend to be higher for lower incidence disabilities, which recently have experienced growth. Some of the most widely reported data on school spending for youth with disabilities come from a national study called the Special Education Expenditure Project (SEEP). Using those data, Chambers, Shkolnik, and Pérez (2003) found that average school spending in 1999–2000 on students with learning disabilities was about 60 percent higher than typical spending on other students, but for students with autism it was about 190 percent higher.4Figure 1 includes trend lines for terciles of disability categories based on the estimated average cost of educating students with different categories of disabilities from Chambers, Shkolnik, and Pérez (2003).5 Collectively, the trends indicate that the distribution of identified special needs in the United States increasingly features disabilities that are more costly to serve. This shift is the result of falling identification rates for the least costly disabilities and rising identification rates for the most costly disabilities.

Federal and state governments appropriate funds annually to help school districts defray the additional cost of educating students served by IDEA above what they spend for typical students. Using data from California in 2006–07, Lipscomb (2009) found that state and federal aid offset 62 percent of this additional spending on students with disabilities, with school districts paying the remainder with general purpose funds. State aid for special education programs tends to be larger than federal aid across states, making special education finance primarily an issue for local and state policies. School districts are held responsible for meeting student needs and states provide the largest source of external funding.

States use different methods for distributing aid for special education programs to school districts. All special education funding systems require that school districts meet special education needs. The differences lie in what factors are used to determine how much aid a school district receives. Ahearn (2010) grouped states’ primary special education funding formulas in 2008–09 into eight categories: multiple student weights, single student weights, resource-based, percentage reimbursement, block grant, combination, no separate special education funding, and census-based. Weights-based formulas allocate funds by special education enrollment counts, which may or may not vary by disability or type of placement. Resource-based models allocate funds typically by the amount of education resources that districts commit to serving students’ special education needs. Percentage reimbursement models allocate funds based on a predetermined percentage of actual expenditures. Block grants allocate funds based on prior allocations. Although these methods vary, most special education finance models direct more state revenue to districts with greater identified needs as determined by school districts. Figure 2 displays the states using each type of funding formula according to Ahearn (2010), who describes these methods in more detail.6

Figure 2.

Special Education Funding Formulas by State, 2008–09. Source: Ahearn 2010.

Figure 2.

Special Education Funding Formulas by State, 2008–09. Source: Ahearn 2010.

In this policy brief, we examine the final category of special education finance models used by states, an approach called census-based funding. Under a census-funding model, state aid is based primarily on total district enrollment and a fixed aid amount per student. Census funding differs from all other methods used by states for distributing special education funds because aid is largely independent of the characteristics of special education programs, such as the number of students served. Census funding was seen as a cost containment strategy during a period of growing special education enrollment by several states adopting it in the 1990s. States adopting census funding limited their fiscal exposure to rising disability rates by ceasing to provide aid based on factors that districts arguably can influence (e.g., the number of students identified, the nature of their disabilities, how students are served across educational environments, or the resources committed to helping each student), thus forcing school districts to assume more fiscal responsibility for additional special education services. These states assume that additional costs to meet special education needs are distributed approximately uniformly across districts based on total enrollment.7 Currently, seven states and the federal government have adopted census funding for special education finance (Ahearn 2010).8

Although we focus on the census-based funding model in this brief, we do not view it necessarily as the preferred policy option for all states relative to other funding models in use. Rather, our intent is to address three questions to help policy makers, educators, and researchers gain a fuller understanding of census-funding models and, more generally, about the implications of special education finance policies. The first question is, what are the potential advantages and disadvantages of census-funding models? We focus on key advantages and disadvantages and draw on findings from relevant research studies where applicable. As Baker and Ramsey (2010) point out, it is reasonable to be concerned about funding equity with this approach because program funds are intended to support only a subset of students that may vary in percentage terms and in terms of needs across districts. The second and third questions relate to aspects of policy implementation, should a state legislature choose to adopt the census-funding approach. First, we examine what options policy makers have for mitigating concerns about the equitability of funding under a census model. Second, we examine what options exist for helping states and districts to contain special education costs while maintaining a high level of quality.

We do not address what the adequate levels of state and federal aid for special education programs should be. Instead, we study the features of census-funding models and assume that any finance model can provide adequate funding if a legislature's appropriation is sufficient. Adequate funding levels are likely to vary across states and depend on multiple factors, such as available financial support for regular education programs. Determining funding adequacy depends crucially on being able to measure the costs of educating students with disabilities. Measuring costs is extraordinarily difficult to do because education agencies track spending, which overstates costs unless services are provided efficiently. Cost minimization cannot be assumed if districts know that states will reimburse the marginal cost of special education services they are considering providing to students.

At a time when fiscal pressure on many state and local budgets has put a high premium on cost containment and when more students with the most costly disabilities are being identified, this brief provides timely information about one option for containing special education costs. Census funding can be a viable option for states, and indeed a number of states and the federal government have adopted it. But the fact that most states have not adopted census funding suggests that it may not be a preferred option across the board, and indeed our own endorsement of the approach in this brief is a highly qualified one. The equity and other issues about the model we describe below pose legitimate concerns that should be considered carefully by states undertaking finance reform in the future. In particular, a pure census model makes no attempt to account for factors that drive cost variation other than to assume they are uniformly distributed (Baker and Ramsey 2010). We also argue that any cost containment provided for by census funding comes about through a very blunt approach. Like Hartman (2001), we recommend state policy makers not limit their search for ways to deliver cost effective programs to choices about the funding model alone. We offer several suggestions directed at alleviating each of these concerns in addressing the research questions posed in the brief.

Description of Census-Funding Models

The defining feature of a census formula is that funding does not depend on how many students are identified with disabilities or how these students are served. In fact, revenue per special education student is systematically higher in districts with a lower percentage of special education students and systematically lower in districts with a higher percentage of special education students. Funding equity is therefore commonly understood to depend on the assumption of a uniform distribution of special education needs across districts. More precisely, we believe equity depends on a uniform distribution of average additional special education costs (i.e., amount per total enrollment) regardless of the number and types of special education needs served. Under a census-based model, school districts use state and federal aid to pay for the additional costs of serving students with disabilities (relative to other students) until those funds are exhausted. They then pay fully for any remaining additional costs using general purpose or locally sourced funds.

Health maintenance organizations (HMOs) use an approach to cost containment similar to the one theorized to help census-funding models contain special education costs (Dhuey and Lipscomb 2011). HMO providers are typically paid a fixed amount to treat plan members, regardless of how many they actually see or what treatments are required. The HMO model reduces the moral hazard in a fee-for-service system that can lead to overspending on health care. But it has also been shown to incentivize providers to seek healthier patients (Newhouse 1996). Census-funding models for special education are similar because they attempt to slow cost growth by redirecting the incentives of school officials toward reducing caseloads and services.

Most census-funding states adopted the finance model during the 1990s. Pennsylvania and Massachusetts were the initial adopters in 1992 and 1993, respectively. Alabama, California, Idaho, and Montana followed suit during the rest of the decade. New Jersey is the most recent adopter, having switched to a census-based model in 2008. In addition to these seven states, the U.S. government distributes its special education aid through a census formula. Under the federal formula, 85 percent of funds are allocated based on total enrollment and 15 percent are allocated based on the degree of poverty. As of 2008–09, eight other states were considering some type of change to their special education finance methods (Ahearn 2010).

What are the Key Advantages and Disadvantages of Census-Funding Models?

In this section, we describe key advantages and disadvantages of census funding. We begin with the advantages, which are summarized in table 1, and then discuss disadvantages in turn.

Table 1.
Advantages of Census-Funding Models for Special Education Finance
1. Allocations are determined by a transparent process. 
2. Legal protections are maintained for youth with disabilities. 
3. The funding model is adaptable and can include weights for external characteristics. 
4. Financial incentives to overidentify disabilities are reduced. 
5. Districts are incentivized to use cost effective placement options for students. 
1. Allocations are determined by a transparent process. 
2. Legal protections are maintained for youth with disabilities. 
3. The funding model is adaptable and can include weights for external characteristics. 
4. Financial incentives to overidentify disabilities are reduced. 
5. Districts are incentivized to use cost effective placement options for students. 

Advantages

Census-funding models can offer states and districts welcome relief from the complexities of modern-day education funding formulas. Census-based allocations are determined in a highly transparent manner. Because allocations are based on average daily membership, districts enrolling the same number of students from one year to the next obtain equivalent allocations whether they have more or fewer students with disabilities in any particular year. Transparent formulas can help school districts reliably gauge future funding levels and thus better plan for high-quality special education programs over multiple years.

Census-funding models maintain legal protections that are afforded to students with disabilities under IDEA. These protections include the right to a free and appropriate public education and the right to due process. School districts in states that change to a census-funding model must continue to recognize these legal requirements regardless of whether educating students with disabilities can be done within the limitations of state and federal aid or whether locally sourced revenues must also be used.

Census-funding models are inherently adaptable. The archetypal model is based exclusively on average daily membership but modified versions could include weights for characteristics thought to drive variation in additional costs of meeting special education needs. The best characteristics to use are those that school districts cannot influence to overrepresent special education needs. For instance, the formula used by the federal government allocates larger amounts to districts and states with more families with children living in poverty. External weighting factors can be added and adjusted over time to meet changing needs of states, districts, and special education students. Weighted-student formulas have been proposed for regular education funding as well, based on the premise that external characteristics affect the resources districts need to meet state academic standards (Sonstelie 2007).

Census funding reduces fiscal incentives to overidentify disabilities. Researchers have found that school districts respond to incentives in special education funding formulas when determining how to classify and serve students. For instance, Dempsey and Fuchs (1993) found that school districts in Tennessee changed how often they used different placement settings for educating special education students in response to changing fiscal incentives in the funding model. Other studies, such as Kane and Johnson (1993) in Vermont and Cullen (2003) in Texas, have found disability rates are sensitive to changing funding structures.

Recent literature has examined the effects of adopting census funding specifically, finding that state-level disability rates tended to fall following census funding reforms (Greene and Forster 2002; Mahitivanichcha and Parrish 2005; Dhuey and Lipscomb 2011). Kwak (2010) also reached a similar conclusion when examining district-level disability rates in California around the time the state adopted census funding in 1997. In our 2011 analysis of the entire United States, we found that state disability rates fell by approximately 10 percent following census reforms, primarily in subjectively diagnosed categories, such as specific learning disabilities and intellectual disabilities, and in early and late grades (Dhuey and Lipscomb 2011). Reductions in the disability rate appeared immediately in less severe categories and gradually in more severe categories. We also found that census-funding policies were associated with subsequent increases in the share of funding that was locally sourced.

Although the research literature has established a linkage between adopting census funding and post-reform reductions in disability identification, it would be wrong to conclude that census funding models eliminate fiscal incentives to identify disabilities. On the contrary, they may incentivize less identification and less costly placements. Descriptive statistics also dispel the myth that census funding is associated with “low” disability rates in absolute terms. Massachusetts and Pennsylvania are examples of census-funding states that have large proportions of special education students. The funding structure alone does not hold the key to understanding variation in disability rates across states.

Finally, census-based funding formulas may incentivize districts to make better use of cost-effective placement options. A recent survey of district representatives in New Jersey noted that there was pressure after adopting census funding to bring students who had been served outside the district back to regular district schools. Most of the representatives indicated this saved money and allowed districts to better understand how well they were serving students (Augenblick, Palaich, and Associates 2011). Census funding gives districts greater control over service decisions by freeing them from state compensation rates that may have guided placements previously. Control over placement decisions may help districts to use the regular education environment to a larger degree for students with disabilities, which is consistent with expectations set forth by IDEA. Parrish (2001) posits that some students may be better served outside of the special education environment because special education placements tend to segregate and label students, which may not be in their best interest.

Disadvantages

All education-funding models have flaws, including census-based models. In table 2, we list key disadvantages of census-funding models. The primary disadvantage is that the model relies on nontestable assumptions about the uniformity of per-student additional costs of meeting special education needs. Several studies have used data from California, New Jersey, and Pennsylvania—all census states—to examine whether disability rates are evenly distributed across districts (Parrish et al. 1998, 2003b; Baker and Ramsey 2010). The findings from these studies firmly reject a uniform distribution. For example, Baker and Ramsey found a statistically significant relationship between higher poverty rates and higher disability rates in New Jersey and Pennsylvania.

Table 2.
Disadvantages of Census-Funding Models for Special Education Finance
1. The underlying model assumptions are nontestable. 
2. Funding allocations may be inequitable. 
3. Funding allocations can become inadequate over time. 
4. Disability identification and more costly, higher quality services can be deincentivized. 
5. Legal protections for youth with disabilities cannot be maintained if youth are not identified. 
1. The underlying model assumptions are nontestable. 
2. Funding allocations may be inequitable. 
3. Funding allocations can become inadequate over time. 
4. Disability identification and more costly, higher quality services can be deincentivized. 
5. Legal protections for youth with disabilities cannot be maintained if youth are not identified. 

High disability rate clusters can originate for a number of reasons. For instance, certain districts may have developed good reputations for serving special education students with specific types of disabilities. Augenblick, Palaich, and Associates (2011) describe a New Jersey school district that is located adjacent to a military base. The military tries to station personnel with high-needs children at this base because the neighboring district is known to have a good program for serving these children. Another possibility is that districts have higher disability rates because they were able to secure greater funding prior to census-funding reforms by designating more students as eligible for special education programs. It is impossible to untangle these various possibilities fully in analyzing geographic variation in identified disability rates.

Even if disability rates are uniformly distributed, it does not follow that the additional costs of meeting student needs would necessarily be uniformly distributed, too. Different types of districts, such as elementary, secondary, K–12, vocational, or special services districts, may have different costs for providing services (e.g., higher costs related to diagnosing disabilities for elementary districts). Geography and regional labor market conditions can lead to cost variation as well. For instance, the costs of serving students in rural communities may be higher because of transportation issues, or urban communities could face higher costs because of higher expected wages. Here, too, the underlying model assumptions are either nontestable or plausibly violated.

A second disadvantage, related to the first, is that funding allocations may be inequitable. Baker and Ramsey (2010) argue that most people would find it illogical to distribute school aid for children living in poverty or for English language learners based on total enrollment. Special education funding allocations similarly may be fundamentally inequitable for districts with higher and lower disability rates. Policy makers attempting to rectify this through formula weights may not solve the problem if the formula adjustments are wrong. As Parrish et al. (2003b) note, adjustments need to be based on data that are accurate and sufficiently reliable.

The funding allocation may become inadequate over time unless policy makers maintain at least the same real rate of funding. Funding levels could be threatened if support for the funds diminishes once taxpayers can no longer link money to specific students (Parrish 2001).

The flip side of redirecting fiscal incentives away from over identifying disabilities is that census funding can also deincentivize identification and the provision of higher quality services that may be more costly to districts. This is desirable to the extent that foregone services are unnecessary, but it becomes problematic if it negatively affects the quality of education that students receive.

Finally, the most extreme case of the redirection of the fiscal incentives is for students who lose their right to a free and appropriate public education under IDEA based on not having been identified as having a disability. Without being identified with a disability, a student is no longer eligible for an IEP and supplemental services associated with an IEP.

Policy Implementation

In this section, we discuss aspects of census-based funding policy implementation for states that are considering whether to adopt the approach for special education finance. Specifically, we describe options for easing equity concerns and possible ways to contain costs while maintaining quality. Any decision to adopt census-based funding must weigh the approach against the advantages and limitations of other funding models, which are summarized in Baker, Green, and Ramsey (2012). For instance, other methods, such as percentage reimbursement, may be preferable if tying funding to identified needs is viewed as critical for reasons of fairness, regardless of whatever adjustments to census-based models may be possible and states are willing to accept that needs are defined at the district level.

What Options Exist for Easing Equity Concerns with Funding Allocations?

Some census states have used funding adjustments based on external characteristics to help address equity issues introduced by census funding because they appear to provide the clearest path toward improving the funding model. Weighted-student formulas (i.e., census models with per-student adjustments based on external characteristics) retain transparency in the funding process while recognizing that certain external factors are thought either to drive cost variation or to relate closely to factors that drive cost variation, without altering existing identification and placement incentives. The federal special education funding model is the clearest example of a weighted-student formula because the per-student allocation generated by students from low-income families is higher than for other students. Recognizing such factors—if indeed they explain variation in costs—makes the funding model more equitable by providing additional funds to districts thought to have greater additional costs of meeting special education needs based on factors outside their control. In contrast, others may argue that once adjustments become the rule the funding model itself should be reevaluated (Baker and Ramsey 2010).

We consider several possible adjustment factors in this section, including adjustments for regional wage levels, population density, poverty, and students with exceptionally high-cost disabilities. Each option has merit, and some or all of these may be appropriate for individual states. We recommend states carefully consider their own geographic cost patterns before selecting adjustment factors to gain a better understanding about the best formula option. States should also commit to periodically recalibrating the adjustments and the weights to be responsive to changing student demographics and drivers of costs.

Personnel costs are the largest expenses for school districts, and variation in regional wage levels affects their purchasing power under equal budgets. For example, Winters (2008) found that teacher salaries in Georgia varied across districts because of differences in experience, educational attainment, and local salary schedules. The differences in local salary schedules were related to factors mostly outside of the control of school districts, such as average enrollment per school, wages of comparable workers, property tax wealth, and the composition of the property tax base. To deal with these kinds of variability, researchers have developed comparable wage indices based on the typical salaries of noneducators with similar labor market qualifications as educators (Taylor 2006; Rose and Sengupta 2007). Census-based models for special education finance could incorporate factors that help to equalize district purchasing power, particularly in states where wage levels vary considerably across regions.

A related factor is the enrollment density of districts. Rural districts are less able to leverage economies of scale within special education programs, and thus they potentially face higher per-student costs through providing certain services to fewer students. Some states adjust for density. In particular, California's formula includes a funding “floor” for sparsely populated regions that does not depend on district size. Idaho uses different funding multipliers for smaller districts than for other districts. Transportation costs for rural districts serving special education students may also be significantly more expensive because of large distances between students’ homes and schools. Students with high cost and/or low incidence disabilities often need to be transported to a placement setting outside the regular school district to receive specialized services. Montana encourages districts to participate in local district cooperatives. These cooperatives are given additional allocations to help defray their higher transportation costs.

Poverty has been shown to relate to higher disability rates, particularly among more severe conditions. Evidence of this relationship from the special education literature comes from Lipscomb (2009) and Parrish et al. (2003b) for California and from Baker and Ramsey (2010) for Pennsylvania and New Jersey. For example, severe disability rates in California were 12 percent higher when 60 percent of students were eligible for the free or reduced-price meals program than when 30 percent of students were eligible (Lipscomb 2009). Parrish (2001) describes relationships between special education rates and poverty as somewhat tenuous because they are based on identification rates rather than prevalence rates and may vary by disability category. Parrish, however, also points to parental reports suggesting an inverse relationship between disability prevalence and income and agrees that a vast literature supports a conclusion that lower income levels are typically associated with worse health outcomes, which in turn could lead more students to need special education services. Finally, Baker and Ramsey (2010) show that background factors collected from census data, including measures of poverty and geographical location, can be used to predict approximately one-third of the variation of populations of special needs students within a district. Given these research findings, a poverty-based adjustment could be a sensible option for states considering census-based formulas.

Finally, many existing census-based formulas provide additional funds for serving very high-cost or low-incidence disabilities. For instance, Massachusetts reimburses 75 percent of the costs districts incur over a prespecified threshold in serving individual students. Idaho reimburses districts that identify and serve high numbers of students with emotional disabilities at a higher rate. California reimburses 100 percent of the “excess cost” of some nonpublic school placements. Adjusting for very high-cost/low-incidence disabilities is different from the previous factors because school districts can influence how to serve these students. Nevertheless, the adjustment can be viewed as legitimate within the context of census-funding models because students with very high-cost/low-incidence disabilities are by definition extreme cases. Particularly for small districts, the services required for these students can take up a large proportion of the special education budget and, in theory, affect program quality for other students unless additional funding is made available.

How Can States and Districts Contain Special Education Costs While Maintaining Quality?

Census funding offers a blunt approach to cost containment by forcing districts to pay the full marginal cost of services after exhausting state and federal aid. For some school districts, the responsibility to deliver costly but necessary programs at high-quality levels is difficult to meet financially, and the risk is that fiscal realities will lead to a gradual lowering of quality standards. If services could be provided less expensively, then necessary programs could continue to be provided without sacrificing quality.

For this reason, we recommend that policy makers consider ways to promote cost-effective program delivery in tandem with proposals to contain costs through reforms to the funding mechanism. We offer several suggestions, but this is an area where additional research is needed to identify evidence-based practices for special education that are also affordable to districts. For instance, encouraging inter-district service coordination may help individual districts to specialize and better utilize economies of scale. Private service providers may also be cost effective in some instances and preferable to training and staffing district personnel when it comes to highly specialized services. Another possibility could be for states to push for more inclusive placements. There is some evidence, albeit from preschool, that the cost of providing special education services is lower in inclusive models than in traditional forms of program delivery (Odom, Parrish, and Hikido 2001). Developing and focusing on early interventions can also contain costs by potentially helping students early so that they can learn techniques to cope with the educational process in a more efficient manner throughout their school career. Finally, states should consider alternative educational intervention systems. Some students are identified as having a disability because it is the best available way to provide them with supplemental services. Research into alternative education interventions for students who need additional supports but who could be served outside of the special education system would improve our understanding of cost effectiveness and efficiency in special education.

Conclusions

In response to tightening budgets and growing special education costs, states have looked to the funding mechanism for ways to deliver high quality services more cost effectively. This brief discussed one kind of special education finance model, census funding. Like all education finance models, census funding has advantages and disadvantages. We recommend states weigh these against those of other funding model alternatives, and that any future state proposals to enact census-funding reforms recognize and address the model's main deficiencies explicitly. We describe possible solutions to the main problems, such as use of a weighted-student formula to correct for external factors outside the district's control that affect variation in special education needs. We also recommend states and districts consider ways to increase cost-effective program delivery in addition to considering alternative funding structures for special education. This will help students with disabilities in the United States receive an appropriate public education despite budgetary realities.

Notes

1. 

The law continues to be known as IDEA even though Congress added the word “improvement” to the law's name when it was last reauthorized in 2004.

2. 

An IEP is a legally binding document that describes the services districts will provide to students, how often and for how long services will be provided, measurable goals for the student, and how progress will be assessed.

3. 

The federal disability categories are autism, deaf-blindness, developmental delay, emotional disturbance, hearing impairments, intellectual disabilities, multiple disabilities, orthopedic impairments, other health impairments, specific learning disabilities, speech and language impairments, traumatic brain injury, and visual impairments.

4. 

Across disabilities, average school spending on students with disabilities was 90 percent higher than typical spending on other students according to Chambers, Shkolnik, and Pérez (2003). There is considerable variation in these percentages by state. An analysis of ten states in the SEEP study revealed that spending on special education students in 2001–2002 ranged from 56 percent to 189 percent higher than spending on regular education students (Parrish et al. 2003c). Lipscomb (2009) found that the equivalent statistic for California in 2006–07 was 140 percent higher spending.

5. 

The first tercile includes specific learning disabilities, speech and language impairments, other health impairments, and emotional disturbances. The second tercile includes orthopedic impairments, intellectual disabilities, hearing impairments, developmental delay, and traumatic brain injury. The third tercile includes autism, visual impairments, multiple disabilities, and deaf-blindness.

6. 

Ahearn's 2010 report updates the classification of special education finance formulas undertaken by Parrish et al. (2003a) for 1999–2000.

7. 

As will be discussed later, many states that have adopted census funding adjust their funding formula based on factors external to individual school districts.

8. 

These states are Alabama, California, Idaho, Massachusetts, Montana, New Jersey, and Pennsylvania.

Acknowledgments

Elizabeth Dhuey gratefully acknowledges the financial support of the Social Sciences and Humanities Research Council of Canada.

REFERENCES

REFERENCES
Ahearn
,
Eileen
.
2010
.
Financing special education: State funding formulas
.
Alexandria, VA
:
Project Forum, National Association of State Directors of Special Education
.
Augenblick
,
Palaich
, and Associates
.
2011
.
Analysis of New Jersey's census-based special education funding system.
Available www.nj.gov/education/finance/sereport.pdf.
Accessed 8 March 2013
.
Baker
,
Bruce D.
, and
Matthew J.
Ramsey
.
2010
.
What we know can't hurt us? Equity consequences of financing special education on the untested assumption of uniform needs.
Journal of Education Finance
35
(
3
):
245
75
. doi:10.1353/jef.0.0018
Baker
,
Bruce D.
,
Preston C.
Green
III, and
Matthew J.
Ramsey
.
2012
.
Financing education for children with special needs
. In
Handbook of leadership and administration for special education
,
edited by Jean B. Crockett, Bonnie Billingsley, and Mary Lynn Boscardin
, pp.
97
114
.
New York
:
Routledge
.
Chambers
,
Jay G.
,
Jamie
Shkolnik
, and
Maria
Pérez
.
2003
.
Total expenditures for students with disabilities, 1999–2000: Spending variation by disability
.
Palo Alto, CA
:
Center for Special Education Research, American Institutes for Research
.
Cullen
,
Julie Berry
.
2003
.
The impact of fiscal incentives on student disability rates.
Journal of Public Economics
87
(
7–8
):
1557
89
.
doi:10.1016/S0047-2727(01)00203-1
Dempsey
,
Samuel
, and
Douglas
Fuchs
.
1993
.
“Flat” versus “weighted” reimbursement formulas: A longitudinal analysis of statewide special education funding practices.
Exceptional Children
5
(
59
):
433
43
.
Dhuey
,
Elizabeth
, and
Stephen
Lipscomb
.
2011
.
Funding special education by capitation: Evidence from state finance reforms.
Education Finance and Policy
6
(
2
):
168
201
. doi:10.1162/EDFP_a_00031
Greene
,
Jay P.
, and
Greg
Forster
.
2002
.
Effects of funding incentives on special education enrollment
.
Manhattan Institute for Policy Research Civic Report No. 32
.
Hartman
,
William T.
2001
.
The impact of census-based special education funding in Pennsylvania.
Journal of Special Education Finance
14
(
1
):
13
20
.
Kane
,
Dennis
, and
P.
Johnson
.
1993
.
Vermont's Act 230: A new response to meeting the demands of diversity
.
Montpelier, VT
:
Vermont Department of Education
.
Kwak
,
Sally
.
2010
.
The impact of intergovernmental incentives on student disabilities rates.
Public Finance Review
38
(
1
):
41
73
. doi:10.1177/1091142109358707
Lipscomb
,
Stephen
.
2009
.
Special education financing in California: A decade after reform
.
San Francisco, CA
:
Public Policy Institute of California
.
Mahitivanichcha
,
Kanya
, and
Thomas
Parrish
.
2005
.
Do non-census funding systems encourage special education identification? Reconsidering Greene and Forster.
Journal of Special Education Leadership
18
(
1
):
38
46
.
Newhouse
,
Joseph P.
1996
.
Reimbursing health plans and health providers: Efficiency in production versus selection.
Journal of Economic Literature
34
(
3
):
1236
63
.
Odom
,
Samuel L.
,
Thomas B.
Parrish
, and
Christine
Hikido
.
2001
.
The cost of inclusive and traditional special education preschool services.
Journal of Special Education Finance
14
(
1
):
33
41
.
Parrish
,
Thomas B.
,
Daniel
Kaleba
,
Michael
Gerber
, and
Margaret
McLaughlin
.
1998
.
Special education: Study of incidence of disabilities. Final report
.
Palo Alto, CA
:
American Institutes for Research
.
Parrish
,
Thomas
.
2001
.
Special education in an era of school reform. Report for the Federal and Regional Resource Centers Network
.
Washington, DC
:
Federal Resource Center
.
Parrish
,
Thomas
,
Jennifer
Harr
,
Jennifer
Anthony
,
Amy
Merickel
, and
Phil
Esra
.
2003a
.
State special education finance systems, 1999–2000, Part I
.
Palo Alto, CA
:
Center for Special Education Finance, American Institutes for Research
.
Parrish
,
Thomas
,
Jennifer
Harr
,
Yael
Kidron
,
Leslie
Brock
, and
Priyanka
Anand
.
2003b
.
Study of the incidence adjustment in the special education funding model. Final report
.
Palo Alto, CA
:
American Institutes for Research
.
Parrish
,
Thomas
,
Leslie
Brock
,
Maria
Pérez
, and
Jamie
Shkolnik
.
2003c
.
Maryland special education expenditure project. Final report
.
Palo Alto, CA
:
Center for Special Education Finance, American Institutes for Research
.
Rose
,
Heather
, and
Ria
Sengupta
.
2007
.
Teacher compensation and local labor market conditions in California: Implications for school funding
.
San Francisco, CA
:
Public Policy Institute of California
.
Sonstelie
,
Jon
.
2007
.
Aligning school finance with academic standards: A weighted-student formula based on a survey of practitioners
.
San Francisco, CA
:
Public Policy Institute of California
.
Taylor
,
Lori L.
2006
.
Comparable wages, inflation, and school finance equity.
Education Finance and Policy
1
(
3
):
349
71
. doi:10.1162/edfp.2006.1.3.349
Winters
,
John V.
2008
.
Variations in teacher salaries in Georgia.
FRC Report No. 182
.