How Within-District Spending Inequities Help Some Schools to Fail

School district budgets are in the news. In the past year, superintendents in Seattle, Rochester, and Baltimore have all left their jobs under pressure because of unexpected deficits, and as of summer 2003 Oakland's superintendent was in similar trouble because of a $50 million deficit for the year.

The bad economy is partly responsible. These and thousands of other districts have suffered simultaneous declines in local, state, and federal revenue. But in these cases, district actions made the worst of a tough situation. Instead of adjusting expenditures as revenues declined, these districts continued spending, with some plugging their budgets (that is, inventing revenues to make the books look balanced) in the hope that things would work out in the end. ¹ Such plugging is neither new nor limited to Seattle, Rochester, Baltimore, and Oakland. As a former superintendent involved in an earlier financial meltdown elsewhere explained to one of us, "You can always find money if you are committed to doing something. You just spend it now and cover it next year when the budget goes up."

Another justification for budget plugging is uncertainty. Few districts know precisely how much money they have, and surprise surpluses are also possible. Even in these recent recessionary times, the Philadelphia public schools found $8 million it did not know it had—enough, according to the Philadelphia Inquirer, to employ 180 teachers. ²

Tracking money is a huge challenge for school districts for many reasons: Their revenues come from many sources (state, local, federal, and philanthropic) at different times. Funders require separate record-keeping for each program, and their rules about cost accounting differ. Districts therefore maintain separate accounting systems for funds from different sources, and information is often kept on separate computer systems, bought and programmed at different times, so they cannot talk to one another.

Expenditure systems are also fragmented and isolated from one another. After five years of trying, Washington, D.C., schools still cannot say how many people they have on the payroll. Philadelphia's surplus became apparent only when the district linked up its separate systems for paying employees and funding benefits, to reveal that some employees were covered by insurance multiple times.

No wonder, then, as business analyst Larry Miller has commented, a superintendent can ask five different district budget managers the same question and get five different answers. ³

With that as background, it should be no surprise that districts do not know what they spend on particular functions. San Diego superintendent Alan Bersin has tried for two years to find out what different central office services cost and he still cannot say for sure. And determining how much has been spent at any one school is even more difficult. Schools are not cost centers, so districts do not track the dollar value of resources (teachers, services, and equipment) that flow into them. District budgeting processes create big and hidden differences in school budgets. The fact that districts do not know how much is spent at one school versus another allows for serious inequities that often hurt the schools most in need of resources.

This paper focuses on one aspect of district spending ambiguity, namely, differences in per pupil spending masked by teacher salary cost averaging. It shows how an often-discussed phenomenon—that schools serving poor children get less qualified teachers than schools in the same district serving more advantaged children—is hard-wired into district policy. ⁴ It profiles the budget layering that is then created in attempts to remedy these unacceptable consequences. It also shows how more open funding and accounting practices can help re-sort the most capable teachers so that schools serving poor students can become better staffed.

Research Study

This paper is the result of five years' study of school district budgets. We did not rely, as most researchers do, on published...