20
In assessing factors that determine local districts’ fiscal health, Table 4 provide insight
into the relative contribution of local districts’ budgetary decision-making versus factors entirely
or largely outside their control. Model 4 in Table 4 omits the resource allocation variables that
fall within local districts’ control (P-T ratio, Teacher_Salary, and %Admin, %Health and
%Purchases). The R-square of Model 4, including the variables that are outside local districts’
control (Foundation, Nonfoundation, Enrollment, %CS, %Net_IDC, %SpecEd and %FRL) is
0.21. The R-square of Model 5 increases to 0.26 with the addition of the resource allocation
variables. Consequently, local districts’ resource allocation decisions account for only 20% of
the explained variation in district fund balances, while factors outside districts control accounted
for the remaining 80%.
Districts with High Levels of Charter Penetration
We next consider the possibility that the impact of the loss of students to charter schools
on district financial health is not linear, but rather intensifies in situations where charter
penetration reaches high and sustained levels. Such a nonlinear relationship is suggested by Ni
(2009) who found an adverse impact of Michigan charter schools on district efficiency at high
and sustained levels of charter penetration, but not low levels. Arsen & Ni (2011) also found
that high levels of charter penetration produce greater declines in Michigan districts’ revenues
than expenditures.
Table 5 displays the results for our full model in which %Charter is replaced by variables
that count the number of consecutive years that a district has reached alternative thresholds (5%,
10%, 15%, 20% and 25%) of charter penetration. Under this specification, the loss of resident
students to charter schools has strong negative impact on district fund balances. Moreover the
adverse impact on district finances increases progressively as the charter threshold increases
21
from 5% to 25% of resident students. The results indicate that in the relatively small number of
Michigan districts in which charter penetration reaches very high and sustained levels, the loss of
students to charters causes district fund balances to sharply deteriorate.
17
[Table 5 about here]
Finally we examine the possibility that charter schools have an indirect impact on district
fund balances by changing the share of students who remain in district schools and receive
special education services. Charter school enrollment would cause an increase in the share of
special education students in district schools, for instance, if a disproportionately low share of
students enrolling in charter schools received special education services. Table 6 shows estimates
of fixed-effect regressions in which special education students as a percent of district enrollment
(%SpecEd) is the dependent variable and charter school penetration (%Charter) is a predictor.
In each specification, increases in the share of a district’s resident students attending charter
schools increase the percentage of special education students among students who remain in
district schools.
To test whether the share of special education students mediates the relationship between
charter enrollment and the district fund balance, we estimated a fixed-effects regression that is
identical to Model 5 in Table 4, except excluding the %SpecEd variable. The coefficient on
%Charter becomes -52.627 (p<0.05), suggesting that one percentage increase in charter school
enrollment has an indirect effect of decreasing district fund balance by $4.9 (-52.627 + 47.684 =
-4.9) through increasing the share of special education students in district schools.
[Table 6 about here]
17
In 2012, %Charter was greater than 25% in 13 districts. Four of those districts had surpassed this charter
penetration threshold for at least five years; one district for ten years.
22
Patterns of State Intervention
We turn now to address Research Question 2: among districts that fall into sustained
deficits, which ones does the state take over? Michigan’s emergency manager law affords the
state considerable discretion in the determination of when a financial emergency exists in a local
school district or municipality. There are no well-defined triggers for state intervention (Arsen
& Mason, 2013).
Thus far, the state has appointed emergency managers in three school districts (Detroit,
Muskegon Heights, Highland Park), has dissolved two school districts (Inkster and Saginaw
Buena Vista) and established consent decrees in one (Pontiac).
18
As the top panel of Table 7
shows, students in all of these districts are predominantly poor and African American. All except
Inkster experienced large declines in enrollment between 2002 and 2012. Compared to districts
statewide, all six of these districts experienced much higher loss of resident students to charter
schools and higher shares of special education students.
[Table 7 about here]
We set out to determine whether district demographic or financial characteristics were
related to in which of the fiscally distressed districts the state intervened. To do so we examined
all Michigan school districts that were in deficit for at least two consecutive years in 2012.
Under the emergency manager law, this condition would permit emergency state intervention. In
2012 there were 34 Michigan districts that satisfied this condition. We ask whether the six
districts in which the state intervened differed significantly from the deficit district in which it
did not in terms of their demographic or financial characteristics.
19
18
The state intervened in yet another predominantly poor and African-American school district, Benton Harbor, too
recently to include in the analysis. Benton Harbor is currently operating under a consent decree.
19
The state intervened in two districts in 2012 and three in 2013. The state declared a financial emergency in
Detroit Public Schools in 2009.
Dostları ilə paylaş: |