Colorado's Schools: The Great Divide?
Public education is always a controversial subject among parents, educators, legislators, and courts-in Colorado and nationally. Our state government spend more on it than any other single activity. In addition, local property taxes paid to school districts are the major share of total property taxes. Public education and its financing reaches deeply into our lives, as taxpayers and as patrons. The fervor of the debate is a reflection of very basic parental and more philosophical societal concerns about education as the foundation of opportunity. Such sentiments as "our children are our future," and "education is the most important thing we can give our children," are accepted as verities. In this paper, the data on how schools are financed is examined to see how closely Colorado is meeting the commitments it has made to equal education.
These deep-seated concerns about education are reflected in our most fundamental laws. In Colorado, Article IX, Sec. 2, of the constitution says that "The General Assembly shall, . . ., provide for the establishment and maintenance of a thorough and uniform system of free public schools throughout the state, wherein all residents of the state, between the ages of six and twenty-one years, may be educated gratuitously." In Colorado, half of local property taxes (almost $2 billion in 1999)(1) are spent on public education and then the state adds an almost equal amount from its sales and income tax revenues (just over $1.9 billion in fiscal year 1998-99).(2) Even those amounts were considered by many as inadequate and last year, through the initiative process, 53% of Colorado voters approved Amendment 23 to the constitution, requiring that more state tax revenues be devoted to public education.
The Colorado scheme for funding public education essentially tries to meet the constitutional obligation to provide "uniform" public education in all school districts and, at the same time, recognize and accommodate differing local needs and desires. At the state level, attempts to accommodate different local circumstances, include transportation funding that recognizes some rural districts must spend relatively more on transporting students than more compact urban districts. Because there are some economies (and diseconomies) of scale for administering a school district, the state share of per-pupil funding is further adjusted to recognize the cost differences in administering small, medium, and large school districts. There are obvious differences among regions of the state in the cost of living and, therefore, in the costs of everything from salaries to food, so the state formula makes another adjustment to accommodate this.(3)
In addition to the state government's accommodations for local variations in priorities, the public education funding system recognizes local differences by allowing districts to opt for override levies to the property tax to provide monies that are additional to the base formula amounts.
The most significant difficulty with compromise between uniformity and local variation hinges on the fact that half the monies involved come from the property tax. To avoid the possibility of a local district reducing its property tax in order to get more than its share of state money, a minimum local property tax must be maintained.(4) This minimum levy must be equal to the smallest of three possibilities (1) the school district's mill levy for last year, (2) the maximum amount permitted by the TABOR amendment, or (3) enough to fund the total prescribed per-pupil amount of funding and categorical programs (less the state's share of per-pupil funding and specific ownership tax revenues). In addition, all school construction monies must be raised within the local district.
DIFFERENT TAXING CAPACITIES OF SCHOOL DISTRICTS ACROSS COLORADO
The formula for the minimum levy effectively recognizes that property values vary widely among school districts and that this has a direct bearing on how much money a local district can raise by property tax levies. In 2000, the total value of taxable assessed property in the Denver School District was $6.3 billion. In Edison 54JT School District (located in parts of El Paso, Lincoln, and Pueblo counties), the total taxable assessed value was more than 2,500 times less at only $2.4 million. Of course, Denver has many more pupils to educate than does Edison 54JT. With almost 70, 000 students in 1999, Denver's taxable assessed property value is $91,000 per pupil. Edison's is about one-third as much at $30,000 per pupil.
Across Colorado, the Aspen School District had the highest taxable assessed property value per pupil of $994, 855 in 2000, while the Sanford 6J School District (in Conejos and Alamosa counties) was the lowest with only $10,257. The direct consequence is that a one-mill property tax levy in Aspen would generate $995 in revenue, while in Sanford it would get only $10. Appendix A is a list of the total taxable assessed property value, the pupil enrollment, taxable assessed property value per pupil, and the percentage of residential property for each school district.
In theory, all of this complexity should mean that every public school student in the state would have a "uniform" amount of financial resources available for his or her education. But two exceptions to uniformity are immediately apparent: (1) money for constructing, remodeling, and modernizing school facilities is paid only by local district bond and property tax proceeds, and (2) local districts are allowed to raise more than the mandated minimum revenue through property tax override levies. The amounts of money that each of these could generate is directly dependent upon a local district's assessed property value and more indirectly upon the affluence of taxpayers and voters in the district and many other factors.
Bond monies are even more specifically limited by the statute that restricts a district's bonded indebtedness to 20% (or 25% for rapidly growing districts) of the district's total taxable assessed property value.(5) Override levy revenue is also limited to 20% of the state's "total program" amount (includes both the state share and the prescribed minimum local district share of per-pupil funding) or $200,000, whichever is greater.(6) Appendix B is a list of each school district with its total bonded debt limit (assuming all are at the 20% maximum), its per-pupil bonded debt limit, and its actual override levy, if any, plus the per-pupil revenue that would be generated from the override levy.
Once again Denver schools have the highest bonded debt limit at almost $1.3 billion, while the Edison 54JT district limit is only $488,000. On a per-pupil basis, the Aspen schools are again the highest at $199,000 and the Sanford 6JT district is the lowest at only $2,000. Only 62 of the 176 school districts have an override mill levy. The highest of these is Adams County 14 with a levy of 17.452 mills. Kit Carson School District (in Cheyenne county) generates almost $2,300 per pupil in override revenue, the most of any district.
CONSTITUTIONAL COMPLICATIONS TO THE SCHOOL FINANCE FORMULA
If school financing in Colorado is not already sufficiently complicated, the money available to the state for funding its share is limited by the TABOR amendment(7) and by a 1991 law often called the Arveschoug-Bird limit.(8) The former limits the amount of revenue the state can keep each year to the amount it kept last year plus a increase equal to the percentage of population growth and the rate of inflation. The latter limits the legislature to increasing its appropriations by only 6% each year. The passage of TABOR in 1992 gave all previously passed statutory limits, that might have been adjusted by the legislature, the permanency of constitutional status. It remains to be seen whether the more recent passage of Amendment 23 may now have changed one or both of these limits to an extent.
On the local school district side of the equation, for almost two decades, the Gallagher amendment has regularly lowered the percentage of residential property value considered taxable-from an initial 21% to 9.15% in 2001 and 2002. The TABOR amendment limits the amount of revenue that a school district may keep to last year's amount plus inflation and the percentage increase in pupil enrollment. And under a third statutory limit dating to 1915 (now effectively made a constitutional limit by TABOR), the amount of a district's property tax revenue can increase by no more than 5ï¿½ % from one year to the next.(9)
THE EFFECTS ON OPERATING REVENUES OF LOCAL SCHOOL DISTRICTS
What are the actual results of this very complicated system of financing Colorado's public schools? If absolute uniformity were the goal, then there would have to be equality of revenues among school districts. In fact, we expect a certain degree of reasonable differences to occur, because of override levies and the high level of local discretion left with each school board and its voters. As expected, considered separately, state and local revenues vary considerably from one school district to the next. As their property values vary, individual districts can generate varying amounts of local revenue. In school year 1996-97,(10) the Telluride R1 school district (in San Miguel county) enjoyed $11,638 of local revenue per pupil, while Sanford 6J (in Conejos and Alamosa counties) had only $853.(11) The statewide average was $3,692. If absolute uniformity existed in the system, one would expect that school district revenues from the state would follow the opposite pattern and Telluride would receive the least with Sanford receiving the most state monies.
The Edison 54JT district, one of the poorest, received the most per-pupil state revenue at $8,285, while Eagle County RE50, one of the wealthiest, got the least, $10. If absolute uniformity were achieved, local revenues plus state revenues would equal the same per-pupil amount for each of the state's 176 school districts. The total of state and local per-pupil revenue for each school district would be the same, but it is not. The Briggsdale RE10J district (in Weld and Morgan counties) had a total of $17,815 per pupil, while the Pueblo City 60 district had only $4,900.(12) Again, the statewide average was $7,304. Appendix C is a list of each school district with its 1996-97 per-pupil local, state, and total state and local revenues. Clearly, something accounts for more of the variation in school district revenues than just the amount of local revenues.
THE GALLAGHER AMENDMENT AND LOCAL PROPERTY TAX REVENUES
One of the reasons often cited for the overall variation in revenues has been the effects of the Gallagher amendment. The different assessment ratios (the percentage of a property's actual value that is considered taxable) that it establishes for residential-9.74% last year-and commercial properties-held at a constant 29%, might mean that school districts with a high proportion of residential property would need more money because they would have more pupils, but would not be able to generate as much local revenue as districts with more commercial property. Appendix A lists the percentage of residential property value for each school district. They range from a low of less than 3% for Kit Carson R1 (Cheyenne county) to a high of more than 75% for Platte Canyon 1 (Park county).
The results here indicate that there is some merit to the argument. As the proportion of residential property value in a school district increases, its revenues per pupil from local sources do tend to go down. Overall, a one percent increase in the proportion of a district's residential property value means a decrease of $19.62 in its local per-pupil revenues.(13) Under a goal seeking uniformity of revenues among school districts, it would be expected that state revenues would do the opposite and go up as the proportion of a district's residential property went up. Instead state revenues go down even more sharply than local revenues. For every one percent increase in a school district's proportion of residential property, its revenues from the state fall by $30.67 per pupil.(14) With both state and local revenues trending in the same direction, it follows that a district's total per-pupil revenues, also move downward as its proportion of residential property rises. And for every one percent rise in the percentage of residential property value in a district, the total of its state and local per-pupil revenues can be expected to go down by $50.29.(15)
CAPITAL CONSTRUCTION REVENUES IN LOCAL SCHOOL DISTRICTS
Operating revenues cannot be used to construct new schools or undertake major remodeling or repairs to a district's physical facilities. For these purposes, a school district must ordinarily depend upon its ability to issue bonds which are a debt secured by the district's ability to repay the loan from its own property tax revenues. In the interests of maintaining financial integrity and solvency, a state statute limits each district to a bonded indebtedness of no more than 20% (25% for a few rapidly growing districts) of its total taxable assessed property value. Even though local revenues and bond limits are both a function of total assessed property value in each district and, therefore, both are affected by the Gallagher amendment, they do not follow the same trends. While revenues decrease as the percentage of residential property goes up, total district bond limits increase. For each one percent increase in the percentage of residential property value in a district, its bond limit may be expected to rise by about $2.5 million.(16) To put this in perspective, the average bond limit among all districts is about $53 million, or $17,289 per pupil.
The results suggest that residential property values are considerably higher in districts that are heavily residential than they are in districts that have more commercial property. But it might be expected that these heavily residential districts would also have more pupils than more commercial districts and with more students comes a greater need for facilities. In fact, the proportion of residential property in a school district seems to have only a very slight bearing upon how much the district enrollment will increase. For each one percent rise in the percentage of residential property, the district enrollment can be expected to go up by much less than one student (actually by only 0.0006).(17) While the number of residences in a school district may suggest how many students the district will have, the number of students does not vary discernibly with the value of that property.(18) In fact, high value properties may often be associated with older families who no longer have school age children.
For per-pupil bond limits, the heavily residential districts do not seem to fare quite so well. There is no statistically significant relationship between a district's proportion of residential property and its per-pupil bond limit. The average per-pupil bond limit is just over $17,000 among the state's school districts.
ADJUSTMENTS TO THE OPERATING FORMULA TO REFLECT DIFFERENT NEEDS
The Colorado school finance formula attempts to account for relative wealth or affluence in school districts. In theory, as the affluence of the school districts taxpayers is less, they would be less able (and maybe less willing) to vote for larger property tax levies and revenues for schools. The state school financing plan attempts to provide more money for less wealthy districts. To do this the formula takes property values into account and uses the number of students eligible for the federal free or reduced-price lunch program as a measure of at-risk children who require a greater expenditure of school district dollars.(19)
For each one percent increase in eligible students, state revenue goes up by $37.45 per pupil,(20) but local revenues go down by $34.27 per pupil.(21) While it would appear that the financing formula works to slightly increase the funds available to districts with higher percentages of eligible students, the statistical tests show no significant relationship between the percent of eligible students and overall per-pupil funding.(22) Appendix D lists each school district and the number of students enrolled in its K-12 programs, the number of students eligible for the federal free or reduced-price lunch program, and the percentage of eligible students.
In the operation of the state school financing plan, there is clearly an attempt to address the issue that all Colorado student should be treated equally. In addition, there are accommodations for local preferences through an allowance for limited override levies. Also, the plan attempts to address the issue of variations in local circumstances, by providing extra monies for rural districts with more transportation needs and smaller classes, as well as for districts with more at-risk students. The plan also recognizes (as noted earlier) the economies and diseconomies of scale among small and large districts and different costs of living among districts.
The unit chosen for equalization is the school district, rather than individual schools. Even with a plan that is working effectively, some differences in total funding will remain among districts. For example, there may be some wealthier districts with higher per pupil revenues because of override levies. But systematic variation among different regions of the state suggests that the financing plan has not adequately addressed the issue of equalization.
REGIONAL VARIATIONS IN SCHOOL FUNDING
The results here show that there are systematic and significant regional variations in the amounts of money that are available for educating Colorado's public school students. In a subsequent paper we will explore further the effects on quality of education and how much students learn.
In order to look at these regional variations, we have divided the state into five regions reflective of different geography, economic, and demographic characteristics. These regions are based on county commissioner districts (Figure 1).
Regional Variations in Public Education Finance
(Based on data from 135 of 176 School Districts)(23)
|Average 96-97 Total State & Local Revenue per Pupil||$6,782||$6,267||$7,141||$7,027||$7,060||$6,953|
|Average 96-97 State Revenue per Pupil||$3,343||$3,165||$3,795||$2,160||$2,663||$4,469|
|Average 96-97 Local Revenue per Pupil||$3,439||$3,102||$3,346||$4,867||$4,397||$2,484|
|Average Bond Debt Limit per Pupil||$16,800||$10,908||$11,928||$39,726||$27,137||$8,790|
|Average 2000 Override Mill Levy||1.749||2.755||0.494||2.415||1.459||1.001|
|Average 2000 Bond Redemption Levy||6.402||10.398||6.286||4.760||4.002||3.097|
|Average % Students Eligible for Federal Lunch Program||34.67||29.47||34.24||25.07||26.98||54.45|
|1999 Average Teacher's Salary||$32,228||$35,027||$28,921||$32,912||$33,149||$29,285|
|Average Pupil/Teacher Ratio||14.8||16.7||13.5||14.6||15.1||12.8|
|Average % 1999 7th Grade Students Scoring Proficient or Advanced on CSAP Reading test||55.98||54.88||59.31||58.01||58.92||51.71|
|Average % Minority Students||24.5||27.0||15.9||13.1||14.8||41.0|
Nevertheless, it is apparent that there are systematic regional variations in the financial effort that is put into public education, a suggestion that the constitutional goal of "uniformity" has not been achieved. The purpose of this paper is not to attempt an assessment of the quality of public education in Colorado or among these five regions and such an assessment is clearly not possible on the basis of scores by students in one grade, on one subject, for one year. However, the data presented above do at least suggest that quality, as well as funding varies significantly between regions of the state.
Tom Brown (B.A. 1963, University of Texas at El Paso; J.D. 1969, University of Louisville; Ph.D. 1999, University of Colorado) serves as research associate with CCPS. He has conducted extensive research in the area of Colorado local government taxation and finance. His dissertation, Constitutional Tax and Expenditure Limitation in Colorado: The Impact on Municipal Governments, reflects extensive research in local government finances. More recently, working closely with local officials across the state, Dr. Brown completed another analysis, Amendment 21: The Impact on Local Governments. His work has recently been published by Public Budgeting & Finance. Dr. Brown was previously an attorney practicing in areas of federal civil appeals, anti-trust litigation, and complex reorganization bankruptcies.
- Division of Property Tax, Department of Local Affairs, 1999 Annual Report to the General Assembly, Denver, CO, 2000.
- Office of the state Comptroller, Department of Personnel, Colorado Comprehensive Annual Report for the year ended June 30, 1999, Denver, CO, 1999.
- Some districts contend that the cost-of-living adjustment is flawed and inaccurate. Senate Bill 01-137 includes, among other provisions, changes in the calculations of this adjustment. During the 2001 Regular Session of the General Assembly, the bill was postponed indefinitely by the Senate Appropriations Committee, effectively killing it.
- The Colorado constitution, Article X, particularly sections 3 and 15, establishes an elaborate procedure by which statewide property assessments are monitored to assure that they are equalized among the counties. Behind the legal requirement is the concern that unequal assessments not be used as a way to avoid the local minimum tax effort for school funding.
- Colorado Revised Statutes, 22-42-104.
- Colorado Revised Statutes, 22-54-108(3)(b).
- Colorado Constitution, Article X, Sec. 20.
- Colorado Revised Statutes, 24-75-201.1(II)(B). Although only a statute, the TABOR amendment probably prohibits the legislature from raising the limit without approval of the voters.
- Colorado Revised Statutes, 29-1-301.
- This is the latest data available on the Colorado Department of Education's website.
- Actually the Fountain 8 school district in El Paso county had only $801 of local revenue, but it receives considerable federal monies from maintaining schools, both on and off Fort Carson, with high populations of students from Army families.
- Again, the Fountain 8 district was lower, with only $4,838, not including its federal revenues.
- The regression coefficient is -19.62, t statistic of -2.40, and probability of 0.0173. At a 95% confidence level, the coefficient ranges from -35.74 to -3.50.
- The regression coefficient is -30.67, t statistic of -4.81, and probability of 0.0000. At a 95% confidence level, the coefficient ranges from -43.26 to -18.08.
- The regression coefficient is -50.29, t statistic of -6.59, and probability of 0.0000. At a 95% confidence level, the coefficient ranges from -65.35 to -35.23.
- The regression coefficient is $2,515,217, t statistic of 4.32, and probability of 0.0000. At a 95% confidence level, the coefficient ranges from $1,366,239 to $3,664,195.
- The regression coefficient is 0.0006 students, t statistic of only 5.05, and probability of 0.000. At a 95% confidence level, the coefficient ranges from 0.0004 to 0.009 students.
- Regressing 1999 district pupil enrollment on district total residential property value, yields a coefficient of 0.000, a t-statistic of 38.83, with a probability of 0.000. The adjusted coefficient of determination (r2) is 89.6.
- See Federal Register, Vol 64, No. 63, April 2, 1999, pages 15951 and 15952 for the regulations and tables of eligible family incomes. Those students from families with incomes of less than 185% of the federal poverty guideline are eligible for reduced price lunches. Those with family incomes of less 130% of the poverty guideline are eligible for free lunches. In 1999, the 130% limit for a family of four was weekly income of less than $418. The 185% limit for a family of four was weekly income of less than $595.
- The regression coefficient is $37.45, with a t-statistic of 5.94, and a probability of 0.000. At a 95% confidence level, the coefficient ranges from $25.01 to $49.89.
- The regression coefficient is minus $34.27, with a t-statistic of -4.36, and a probability of 0.000. At a 95% confidence level, the coefficient ranges from minus $49.77 to minus $18.77.
- Regressing 1996-97 per pupil total state and local revenues per pupil on district percentage of students eligible for the federal free or reduced-price lunch program, yields a coefficient of 3.18, a t-statistic of 0.37, with a probability of 0.709.
- Data from 41 districts was missing from the CSAP scores, the school lunch program, or teacher salaries. The Colorado Department of Education does not report CSAP scores from any school in which fewer than 16 students took the test. For some small districts, this means that there are no scores available for the entire district.