Two crises in school finance likely to distress the economy
over the long term:
1. Over
the link between family income and per-student spending: To what extent should
parents’ income determine the value spent on a child’s elementary and secondary
schooling?
- The
solution is important for two reason:
1. grounds
of distributive justice
2. for
macroeconomic growth
since it is not usually efficient to have the parents’
earnings as a basis of human-capital investment in a child. (Michele Boldrin,
1993; Roldan Benabou, 1996; Raquel Fernandez and Richard Rogerson, 1996).
2. In
spite of the fact that property tax, which is the traditional and dominant
source of revenue for public elementary and secondary education, may bear good
economic properties given optimal conditions, there is growing political
dissatisfaction with it.
The concerns of the study are the distribution of per
student expenses in the United States; the change (if any) in the link between
per student spending and property value; and lastly, the link between parents’
earnings and per student spending.
Fundamental to these concerns is a classic puzzle about
revolutions:
Are
the current crises because of the increasing system failure or of the rising
expectations on what a school-finance system should be able to accomplish?
I. Empirical
Strategy
The paper uses district-level data for Massachusetts,
Illinois, and California – these states were chosen due to the data quality and
representativeness. The data consist of records in expenditures, number of
students at local equalized property valuation.
Concentrating on fiscal school districts, which are
characterized as having considerable autonomy in revenue-raising and spending,
is essential in studying school finance, since they are the suitable units to
match with data on property values and per capita earnings. Fiscal districts
should not be mistaken for attendance districts which do not independently
raise taxes or establish spending.
The system that started the century by each state is based
mainly on local property taxes. California’s fiscal districts are found to be
consistently larger compared to Massachusetts and Illinois. Eventually the
state governments were granted more control however they continued to depend on
the property tax. In 1970, the states established a floor under per-student
spending through reallocating funds from districts having high property value
to districts with low property value per student. Moreover, during 1978 – 1980,
California was pressured to move to statewide school finance such that the
state has one fiscal district distributing funds on a strict per-student basis
over the attendance districts.
II. Results
One of the findings shows that the inequality in per-pupil
spending among districts, measured by the enrollment-weighted coefficient of
variation was fairly stable from 1900 to 1990. The general 0.2 coefficient of
variation implies that a standard deviation in per-student spending is roughly
20% of the mean. The student whose per-pupil spending is at the 95th,
90th and 75th percentile attends school that spends about
60%, 45% and 30% greater than the student at the 5th, 10th
and 25th percentile, respectively. The scale of the value of
inequality relies on whether the standard of comparison is absolute equality or
income inequality.
In addition, there is an observed obvious parallelism
between the time pattern in spending inequality and that of economy-wide income
inequality. 1930’s and 1970’s were decades where there is apparent rise in
spending and income inequality. 1940’s was marked with declining spending and
income inequality.
In spite of the three states displaying rather similar
patterns, Illinois started with a higher spending inequality, and California
always had the lowest and least volatile spending inequality, which may be due
to that fact that its huge fiscal districts included a broader array of
professions and industries.
Another result exhibited the per-student valuation using the
enrollment weighted coefficient of variation, and the equalized property
valuation in calculating aid. The per-pupil valuation is comparatively less
stable than per-pupil spending. In 1900 to 1990, a standard deviation in
per-student valuation ranges from 55% to 65% of the mean. Also, there seems to
have a long downward trend in the per-student valuation inequality, in contrast
to per-student spending inequality. The trend is most likely from the decline
in arbitrary differences in districts’ per-student valuation because of lumpy
real-estate assets, in which taxation of such assets has become increasingly
distinguished from housing property taxation. In addition, districts have
merged, distributing these assets over a large number of children.
The main idea is that the link of per-student spending with
per-pupil valuation is not rigid. Being rigid implies per-student spending
inequality would have declined remarkably between 1900 and 1990.
Next, to net out the impact of economy-wide income
inequality on the per-pupil spending and valuation inequality would allow the focus
to be on whether the school finance systems are either operating differently
over time or similarly in various environment.
Taking this idea into consideration, the unadjusted result
shows that from 1970-1990 there is an increase in the between-district income
inequality, however once adjusted for economy-wide income inequality the rising
pattern is eliminated. This means that since 1970, with income as basis, there
has been no increased distributing of households into districts.
In the study a regression (for Massachusetts) of per-pupil
spending on per-pupil valuation, per capita income, percentage elderly,
percentage school-age, and percentage high-school graduates was run and the
result was as follows:
- Fraction
of the population over the age of 65 has a relatively positive, statistically
significant impact on per-pupil spending in 1900 and 1910, which means that an
additional 1 percent of the population having an age over 65 increases the
school spending by 0.5 percent. However, by 1990, the estimate of this
coefficient steadily reverses sign such that it has relatively negative,
statistically significant impact on per-pupil spending, which implies that an
additional 1 percent of the population having an age over 65 decreases school
spending by 0.7 percent.
- The
percentage of households having school-age children has a steady negative sign
- The
percentage of high-school graduate adults has a positive and statistically
significant impact on per-pupil spending until 1950, but became insignificant
afterwards.
- The
per-pupil valuation and per capita earning are strong determinants of per-pupil
spending
- An
observed interesting pattern is that from 1900 to 1970, per capita income is by
far a stronger determinant of per-pupil spending than per-pupil valuation. However
past 1970, the explanatory power of per capita income drops.
- Another
pattern is that in 1900 to 1990, elasticity of spending with regards to per
capita income drops from roughly 0.35 to 0.06 and 0.12 for Massachusetts and
Illinois, respectively.
- Two
explanations for the result:
1. Over
the past 20 years, grants to districts where low-income households reside have
considerably developed such that low per capita income has now an uncertain
impact on per-pupil spending.
2. Per-pupil
valuation is increasingly an indicator of the local demand for per-pupil
spending (or indicator for those elements of taste and household income that
determine the demand for per-pupil spending).
- This
is related to one of the result in which over the century there was an increase
in the explanatory strength of per-pupil valuation, and in the estimated
elasticity of per-pupil spending with regards to per-pupil evaluation.
III. Conclusions
Despite the definite dip of inequality in 1950, per-pupil
spending inequality has been relatively constant over the century. Most of the
volatilities in spending inequality were due to economy-wide modifications in
the inequality of household earnings. Spending inequality does not show the
same downward trend seen in per-pupil valuation.
Regressions of per-student spending on per capita income,
and demographic variables reflect the strengthening of the link between
spending and income in terms of explanatory power and elasticity of spending
out income from 1900 to 1950, (and weakened afterwards). By 1990, statistically
significant relation between local per capita earnings and per-pupil spending
in Massachusetts was no longer observed. In the case of Illinois, results were
more subtle. The explanatory power of per-pupil valuation rose from 1900 to 1990,
regardless of school finance equalization programs. This is possibly because
per-pupil valuation has increasingly become an indicator for the locally
favored level of per-student spending.
Looking at California’s patterns of spending inequality
(having the lowest level and smallest fluctuations from 1900-1970) it was
unexpected that the state would have a school finance crisis in the 1970’s and
would eventually discard local school finance in general. This implies that
school finance crises are not essentially an effect of a systems break down,
however it may be because of the mounting expectations about the equality a
state’s school-finance should be able to accomplish.
Two implications for economics:
1. Main
view of link between income and spending may be too basic
- To
reduce spending inequality it would perhaps be more realistic to focus on
reducing income inequality than to reallocate the rising share of existing
public-school revenues
2. Per-student
valuation does matter for per-student spending, not because of arbitrary
variations in real property tax, but due to the fact that per-pupil valuation
is an increasingly proper measure of local demand for school spending.
Future school finance reforms must consider a more
sophisticated understanding of valuation.
Source:
Caroline M. Hoxby, “How
Much Does School Spending Depend on Family Income? The Historical Origins of
the Current School Finance Dilemma”, American
Economic Review, Vol. 88, No. 2. (May 1998), pp. 309-314
0 comments:
Post a Comment