12
METHODS AND DATA
Our empirical work to address Research Question 1 takes the form of explaining
variations in district fund balances, since under Michigan law, low fund balances and deficits are
the key triggers for state intervention. We focus on the districts’ general fund which includes all
financial transactions related to districts’ operations, except those required by law to be entered
in other funds, such as capital projects, long-term debt and food service. General fund revenues
and expenditure encompass the elements of district fiscal health in which we are interested. All
Michigan districts in which the state has staged emergency intervention have had serious
problems (deficits) in their general funds.
The balance on district i’s general fund in year t, FB
it,
is given by:
FB
it
=
Revenue
it
– Expenditures
it
+ FB
it-1
Conceptually, therefore, we seek to model influences on district revenues, on the one hand, and
expenditures on the other. For revenues, we focus on districts’ per-pupil foundation funding and
other non-foundation funding which is comprised mostly of state and federal categorical
revenues. We also consider changes in enrollment which affects districts’ total revenue directly
through state funding formulas, but also may influence the relationship between district revenues
and expenditures. In Michigan, when a student leaves a district all funding associated with that
student is lost, but certain costs may be fixed in the short run so expenditures decline more
slowly than revenues. So revenues are likely to decline more rapidly than costs in declining-
enrollment districts.
8
Consequently, declining-enrollment districts would be forced to reduce
8
The state’s pupil count for the allocation of state aid is based on a weighted average of enrollment on the second
Wednesday in the previous February (10%) and first Wednesday in October in the current school year (90%).
Districts losing a student any time between the previous February count day and the first October Wednesday of the
current academic year will lose 90% of funding for that student in the current year.
13
services on remaining students or draw down their fund balances.
9
In addition, we isolate the
influence of school choice policies on enrollment, since these enrollment shifts may impact
district finances differently than those associated with residential population change.
For expenditures, we focus on the density of high-cost students in districts. We also
examine a set of local district resource allocation decisions that directly influence expenditures
and possibly fund balances.
We utilize fixed effects models to estimate the influence of different factors on the fund
balances of Michigan school districts. The basic model takes the following form:
FB
it
= Revenue__it__B__1'>Revenue
it
B
1
+
β
2
Enrollment
it
+ Choice
it
B
3
+ StudentCharacteristics
it
B
4
+
ResourceAllocation
it
B
5
+
θ
i
+ u
it
(1)
In which FB
it
is the fund balance per pupil in district i in year t. Revenue
it
is a vector of two
variables reflecting revenues received by district i in year t, including Foundation revenue per
pupil
10
and Nonfoundation revenue per pupil from local, state and federal sources. We expect
district fund balances to increase with increases in both
types of revenue, although foundation
funding may have a stronger positive influence than nonfoundation funding. Whereas the use of
foundation funding is entirely discretionary to local districts, most nonfoundation funding is
categorical with conditions attached to how it must be spent.
We expect district fund balances to be a positive function of Enrollment
it
, defined as full-
time equivalent student enrollment. This implies that the marginal revenue associated with
gaining an additional student exceeds the marginal cost of additional service provision. Choice
it
9
Conversely, so long as they have excess capacity (a nearly universal condition in Michigan school districts), in
growing-enrollment districts, the additional revenue associated with gaining a student is likely to exceed the
increased cost and tend to increase fund balances.
10
In a few years covered in this study, districts did not receive the full amount of their authorized foundation
funding, due to unanticipated short-falls in state revenue collections. Our Foundation variable is adjusted to
incorporate this “pro-rationing” of funding.