the condition of urban school finance: efficient resource ... · the condition of urban school...

The Condition of Urban School Finance 61

The Condition of Urban SchoolFinance: Efficient Resource

Allocation in Urban Schools

Dale BallouUniversity of Massachusetts at Amherst

About the AuthorDale Ballou is an Associate Professor of

Economics at the University of Massachusettsat Amherst. His research specializes in theeconomics of education, with particular fo-cus on policies that effect teacher recruitmentand performance. His work, much of it incollaboration with Professor Mike Podgurskyof the University of Missouri, has appearedin the Journal of Human Resources, the Quar-terly Journal of Economics, and the Industrial

and Labor Relations Review, among other pub-lications. Their book, Teacher Pay andTeacher Quality, was published by the W.E.Upjohn Institute in 1997.

Professor Ballou received his doctoratefrom Yale University in 1988 and spent oneyear at North Carolina State University beforejoining the economics department at the Uni-versity of Massachusetts in 1989.

62 Selected Papers in School Finance, 1996


The Condition of UrbanSchool Finance: Efficient

Resource Allocation inUrban Schools

Selected

Papers in

School

Finance


The Condition of Urban SchoolFinance: Efficient ResourceAllocation in Urban Schools

Dale BallouUniversity of Massachusetts at Amherst

denced by the interest shown in these commu-nities in such alternatives to traditional publiceducation as charter schools and voucher pro-grams. This dissatisfaction suggests it wouldbe useful to look more closely at the way re-sources are allocated in urban schools to as-certain whether charges of inefficiency arewarranted.

Methodology

The starting point for this investigation ofurban education consists of several criticismsthat have been made of public, if not specifi-cally urban, schools. Policies pursued by pub-lic schools are said to be inefficient or waste-ful in the following respects:

1. Too small a share of district resourcesactually make it to the classroom (i.e., arespent on instruction as opposed to admin-istration or other support services)(Walberg, 1994).

2. Schools and school districts are toolarge. Students have been shown to learn

Urban schools have long been a focus ofpublic attention. Much of this concern hascentered on inter-district disparities in per pu-pil expenditure. However, as state govern-ments have come to play a larger role in schoolfinance, local tax bases have become a lessimportant factor in determining educationalresources. Today, per pupil expenditures inmany (though not all) urban school systemsmatch those of more affluent suburbs. Yetthe performance of urban school systems (asmeasured by such indicators as student testscores, graduation rates, and a variety of stu-dent behaviors) continues to lag behind thoseof other systems (Lippman, 1996). There isno doubt that poor educational outcomes aredue in large part to high concentrations of pov-erty and to other social and economic barriersfaced by disadvantaged minorities in urbancenters. However, critics have also chargedurban schools with waste and inefficiency(Wilson, 1992). Many of these same criti-cisms have been directed at public schools insuburban and rural locations as well. Yet pub-lic dissatisfaction with schools appears to beparticularly high in urban districts, as evi-

Introduction


above. Thus, a finding that urban schools de-vote a smaller share of total resources to in-struction than other school systems is primafacia evidence of inefficiency. To argue thatsuch a pattern of resource allocation is actu-ally the efficient one would imply that on thewhole, public schools devote too large a shareof resources to instruction.

Second, because private schools face mar-ket competition, they are under pressure to useresources efficiently. Significant differencesbetween public and private schools will rein-force the conclusion that the former are notrun efficiently; conversely, the more nearlyalike the two types of schools are, the less rea-son there is for special concern about the prac-tices of urban schools.

Share of Resources Devoted toInstruction

The National Center for Education Sta-tistics (NCES) classifies school districts’ cur-rent expenditures into three broad categories:instruction, support services, and non-instruc-tional services. Instructional expenditures in-clude salaries of teachers and teachers’ aides,and classroom materials. Support services en-compass counseling, administration, opera-tions and maintenance, business office activi-ties, and student transportation. Non-instruc-tional expenditures cover food services andadult education and other community services.

To see whether the allocation of funds var-ies by district location, total dollars spent ineach category have been summed for all ur-ban districts, suburban/large town districts,and rural/small town districts.1 (For concise-ness, these groups will henceforth be referredto as urban, suburban, and rural.) The result-ing totals are displayed as percentages of cur-rent expenditures in table 1. (Expenditures

more effectively in smaller, less imper-sonal settings, offsetting whatever econo-mies may be achieved by operating on alarge scale (Walberg and Fowler, 1987)(Fowler and Walberg, 1991).

3. Public school systems are excessivelybureaucratic (Chubb and Moe, 1990).

4. Teacher compensation is unresponsiveto market conditions and provides little ifany incentive to improve performance(Hanushek et al., 1994).

While these criticisms provide a useful fo-cus for the investigation, none of them speci-fies criteria for determining when urbanschools (or, indeed, any) are inefficient.Benchmarks for efficient performance aremissing. Consider, for example, the charge thatpublic schools systems are excessively bureau-cratic. Given that schools cannot functionwithout some bureaucratic oversight, howmuch oversight is excessive? Similarly, with-out knowing what proportion of district re-sources should be devoted to instruction, itbecomes difficult to determine when a givenpattern of resource allocation is inefficient.

In the absence of a set of benchmarks forefficient performance, this paper relies on acomparative methodology, contrasting urbanpublic schools with public schools in subur-ban and rural communities. In places the com-parison is extended to private schools. A va-riety of indicators will be examined pertain-ing to the criticisms just cited. Systematic dif-ferences unfavorable to urban schools will beevidence of inefficiency. This is not fully con-clusive, of course, for such differences mightarise because urban schools are pursuing themost efficient policies. For two reasons thiswould be unlikely. First, there is probably atleast some truth to each of the criticisms cited

. . . This paper

relies on a

comparative

methodology,

contrasting urban

public schools

with public

schools in

suburban and

rural

communities.

1 This classification scheme, which is also due to NCES, defines urban districts as those located in central cities of StandardMetropolitan Statistical Areas (SMSA). Suburban districts are located in SMSAs but are not in central cities. Large towns areoutside SMSAs but have a population of at least 25,000 and are defined as urban by the Bureau of Census. Small towns areoutside SMSAs and have populations between 2,500 and 25,000. Rural districts are found in places with a population lessthan 2,500.


Table 1.—Expenditure and staffing patterns

Percent of current expenditures allocated to: Urban Suburban Rural

Instruction 61.1 61.1 61.5All support services 33.9 34.7 33.1

Administration 7.5 7.7 8.8Staffing ratios

All staff to teachers 1.76 1.78 1.78All staff to teachers and teachers aides 1.49 1.55 1.52

Administrators to teachers .16 .19 .17

SOURCE: Common Core of Data, Agency Finance Information (Expenditures), Fiscal Year 1992, AgencyStaff Information, Academic Year 1992.

on non-instructional services, which are notdisplayed, are the residual item).

There is virtually no difference betweenurban schools and others in the percentage ofcurrent expenditures allocated to instruction,approximately 61 percent. Suburban systemsspend slightly more on support services (andby implication, less on non-instructional ser-vices). Urban school systems actually devotea smaller share of current expenditures to ad-ministration, almost 15 percent less than ruraldistricts.

Some caution is required in interpretingthese numbers, since classification of schoolexpenditures is problematic (Raywid andShaheen, 1994). All districts do not followthe same accounting practices; there is dis-agreement even among experts on how to com-pute school expenditures. When working withdistrict-level data, the problem is compoundedby differences in types of districts. Some dis-tricts, for example, have been created exclu-sively to serve special education students. Inother districts virtually all expenditures are forsupport services provided to other school sys-tems.

It was for these reasons that table 1 wasprepared by summing expenditures within theurban, suburban, and rural categories. The ef-fect of variation in accounting practices willtend to average out in these aggregates. Inaddition, totals within these broad categoriesshould not be sensitive to the establishment ofspecial districts to perform limited functions.It is the total spending on the function (andnot whether it is one district or another thatperforms it) that determines the entries in table1.2

Given uncertainty about accounting prac-tices, it is worth seeing whether alternativeways of measuring resource allocation presentthe same picture. The lower panel of table 1displays statistics on staffing patterns: ratiosof total employees to teachers and administra-tive staff to teachers. As above, these statis-tics are computed by first totaling the numberof employees within urban, suburban, or ruraldistricts without regard to the particular dis-tricts in which they are employed. The resultsconfirm that there is little difference betweenurban public education and the other catego-ries. The ratio of all staff to teachers is virtu-ally the same across categories. When teacheraides are counted with teachers, urban schools

2 It may be wondered if a few very large districts (such as the New York City school system, with a million students) have undueinfluence on the statistics presented in table 1, distorting the picture of expenditure patterns in smaller but much more numer-ous urban districts. The three largest districts in the United States are the New York, Chicago, and Los Angeles schoolsystems. As a check on the information presented in tables 1 and 2, all statistics on urban districts were recomputed excludingthese three systems. There was a very slight change in the findings: the share of current expenditures on instruction fell to 60percent while those spent on support services rose to 35 percent. However, the ratio of all staff to teachers actually fell slightly(though by less than one-tenth). On the whole, it does not appear that the findings in tables 1 and 2 are distorted by spendingand staffing decisions in the largest systems.


are found to allocate a slightly higher propor-tion of their staffs to teaching than do othersystems, a slightly smaller share to adminis-tration.

Since urban districts serve a high propor-tion of disadvantaged students, it is of someinterest to know whether the patterns in table1 hold when urban districts are distinguishedby students’ economic status. For this pur-pose, urban districts in which more than 17percent of school-aged children live below thepoverty line have been compared to the re-maining urban systems. (Data on this break-down, not displayed in table 1, are availablefrom the author.) It turns out that the poorerdistricts employ more, not fewer, teachers rela-tive to administrators and relative to total staff.The administrative share of current expendi-tures is lower by 0.5 percentage points in theseless affluent schools. (Instruction as a shareof current expenditures is, however, the samein both groups, 61 percent.)

Scale Economies

As noted, urban schools spend proportion-ately less on administration and employ feweradministrative staff relative to teachers thaneither suburban or rural schools. Since urbandistricts tend to be larger than those elsewhere,these differences may reflect economies ofscale. To explore this hypothesis, the two vari-ables pertaining to administration in table 1—the share of administration in current expen-ditures and the ratio of administrative staff toteaching staff—have been regressed on a va-riety of district and community characteris-tics. Two measures of size were used to de-tect scale economies: district enrollment andthe average number of students per schoolwithin the district. An inverse relationshipbetween district enrollment and the share of

resources devoted to administration presum-ably reflects economies in central office op-erations and district-wide services. Increas-ing the number of students per school wouldalso be expected to save on administrationthrough consolidation of positions (e.g., prin-cipals). Other regressors control for thecommunity’s demand for certain kinds ofschool services as well as the educational needsof the school-age population. These variablesinclude the percentage of school-aged childrenliving below the poverty line, median incomeof district households with school-aged chil-dren, and the percentage of household headswith a college degree. Current expendituresper pupil were introduced to allow for the pos-sibility that spending on administration varieswith district resources.3 (For example, as thebudget grows, administrators may find addi-tional slack they can divert to their own staffs.)

Earlier remarks about variation in account-ing practices across districts are relevant here.In an effort to enhance consistency, the esti-mation sample was restricted to independentschool districts. This category excludes manydistricts that function in an auxiliary capacityby providing services to other systems andwhich therefore often exhibit extreme ratiosof administrative to other expenditures. In ad-dition, following the initial estimation, obser-vations with extreme values of the dependentvariables were dropped from the sample (5percent at each end). Since the second set ofestimates did not differ substantially from thefirst on the points of greatest interest, only thefirst is discussed here.

Regression results (table 2) confirm thaturban systems spend proportionately less onadministration than do rural systems and em-ploy fewer administrative staff relative toteachers than do suburban systems, even with

Urban schools

spend

proportionately

less on

administration

and employ fewer

administrative

staff relative to

teachers than

either suburban

or rural schools.

3 It may be wondered if the poverty rate, median income, household education, and per pupil expenditure do not represent toomany ways of measuring the same thing, with the resulting multicollinearity yielding unstable and imprecise estimates. Thesevariables are not, in fact, highly correlated. The largest pairwise correlation, between median income and education of thehousehold head, is 0.75. None of the other correlation coefficients exceeds 0.4. Correlations between the estimated coeffi-cients are generally lower. Estimates are only moderately sensitive to the exclusion of other variables from the model. Thissuggests the various regressors convey independent information.


Table 2.—Regression analysis of administrative expenditures and staff(standard errors in parentheses)

Dependent variables:Administration

Mean of percentage of current AdministrativeIndependent variables: independent variables expenditures(%) staff/teachers

Intercept 1.0 12.5 (.20) .09 (.006)Suburban .19 .10 (.13) .014 (.004)Rural .74 .51 (.12) -.002 (.004)District enrollment (1,000s) 3.15 -.001 (.002) -.0006 (.00007)Students per school (100s) .367 -.51 (.01) -.002 (.0004)Median household income (1,000s) 33.4 -.03 (.004) .0007 (.0001)Percentage of school-aged children below poverty line 17.8 .002 (.003) .0004 (.0001)Percentage of household heads with college degree 15.9 -.005 (.004) -.0005 (.0001)Per-pupil current expenditures (1,000s) 5.07 -.032 (.017) .015 (.001)R2 .18 .10Number of observations — 12,596 11,864Dependent variable mean — 9.7 .18

—Not applicable.

SOURCE: Common Core of Data, Agency School Information, School Year 1991, Agency StaffInformation, 1992 School Year, Agency Finance Information, Fiscal Year 1992, Household Information,1990 Census of Population.

controls with district characteristics. How-ever, there appear to be few economies of scalein central office functions. An increase in thesize of the district by 10,000 students reducesthe share of current expenditures devoted toadministration by only 0.01 percentage points.Although this estimate is somewhat imprecise(the coefficient fails conventional tests of sta-tistical significance), all estimates within a 95percent confidence interval are likewise verysmall. By contrast, average school size doeshave a statistically significant impact on re-sources allocated to administration: an in-crease of 100 students per school reduces theshare of administrative expenditures by one-half percentage point. The impact on the ra-tio of administrators to teaching staff issmaller, at 0.2 percentage points.

Failure to detect savings in administrationas district size increases is troubling, sincesuch economies are to be expected. More-over, given evidence that student achievementtends to suffer with increases in district size

(Walberg and Fowler, 1987), if large districtscannot be justified on grounds of scale econo-mies, it may be hard to justify them at all. Itturns out that there are economies of districtsize, but they become apparent only whenseparate regressions run on subsamples ofurban, suburban and rural schools, respec-tively. (These results, not shown in table 2,are available from the author on request.) Inthe urban subsample, where average districtsize is much greater (15,000 students, com-pared to 5,000 and 1,500 students in the sub-urban and rural subsamples, respectively), co-efficients on district size are an order of mag-nitude smaller than the corresponding esti-mates for the suburban and rural samples andare statistically insignificant. This evidencestrongly suggests that urban districts by andlarge exceed the size necessary to realize scaleeconomies. The notion that there are dimin-ishing returns to increasing district size is fur-ther supported by the fact that estimated dis-trict scale economies are greater for rural dis-tricts than for the suburban districts. Thus it


Table 3.—Scale economies

Change in dependent variablefrom an increase in:

Dependent District size Average schoolDependent variable Sample variable mean (1,000 students) size (100 students)

Per pupil1 All $1,555 0.2 -106.92

Urban 1,459 0.5 -75.52

Suburban 1,565 -8.32 -0.2Rural 1,562 -15.92 -147.42

Percentage ofcurrent expenditures1 All 30.5% 0.012 -0.72

Urban 28.0 0.00 -0.272

Suburban 27.2 -0.01 -0.02

Rural 31.5 -0.03 -1.022

1 Administration, Operations/Maintenance, Business Office, Transportation, and Food Services

2 Regression coefficient significant at 1 percent.

NOTE: Estimation samples restricted to Local Education Agencies (LEAs) which are IndependentSchool Districts. Other regressors included percentage of school-aged children in households below thepoverty line, median household income, percentage of heads of households who are college graduates,and indicators of urbanicity (in the combined samples).

SOURCE: Common Core of Data, Agency Finance Information (Expenditures), Fiscal Year 1992.

would appear that scale economies at the dis-trict level are exhausted somewhere betweenthe typical suburban size (about 5,000 stu-dents) and the average urban enrollment of15,000.

To this point the discussion has consid-ered administrative expenses only. Since theremay be scale economies in other functions, itis useful to examine a broader measure thatincludes spending on operation and mainte-nance, the business office, student transpor-tation, and food services. Table 3 displaysselected results when per-pupil expenditureson these items are regressed on the districtcharacteristics mentioned above. Since thelevel of spending may be affected by districtwealth, in the second panel of table 3 the de-pendent variable is expressed as a percentageof current expenditure. A decline in this per-centage as district or school size rises signalsthe presence of scale economies and meansthat resources are freed up for instruction orpupil support services.

As table 3 shows, there are few scaleeconomies in these functions at the districtlevel (and none among urban school systems).Increasing school size does produce savings,but the amounts are small. If one takes theestimates in panel two as more reliable, in-creasing mean school size by 100 studentssaves urban districts only 0.27 percent of theircurrent per pupil expenditures, or $14 on av-erage (=.0027 times $5,076). The average sav-ings for all public school districts are $35 (.007of $5,069), only slightly more than the reduc-tion in administrative expenses reported intable 2. Whether it is worth increasing schoolsize to achieve savings of this magnitude ismuch in doubt. A growing body of researchhas found evidence that smaller schools pro-vide a superior learning environment to thelarge, impersonal, factory-like schools built ingreat numbers after World War II. In the finalanalysis, the answer turns on whether themoney saved by realizing scale economies canbe put to uses that will have a greater impacton student achievement than reductions in


school size. It should also be recognized thatthe discussion here has focused only on cur-rent expenditures and that a full considerationof scale economies must take account of po-tential savings in capital costs. Unfortunately,data limitations prevent that investigationfrom occurring here.4

Non-Teaching Faculty

Schools have been criticized for assign-ing teachers to non-instructional jobs wherethey carry out administrative or even clericaltasks. In addition, some union contracts callfor a specified number of teachers to be re-lieved of classroom teaching responsibilitiesin order to perform work for the union. Suchpractices reduce the real level of resources inthe classroom in ways that are masked by suchstatistics as aggregate student/teacher ratios.

Unfortunately, it is not easy to examinehow widespread these practices are. While ithas been suggested that a comparison of theschool-wide student/teacher ratio to the aver-age class size reveals how many teachers haveregular assignments outside the classroom(Picus and Bhimani, 1993), the comparisonis misleading: average class size exceeds theschool-wide student/teacher ratio largely be-cause teachers spend fewer hours in class eachday than do students. Discrepancies in theseratios do not mean, therefore, that some teach-ers have not been assigned regular classes ofstudents, but rather that teachers are given prepperiods and other breaks during the day thatreduce at any point in time the number ofteachers available to work with students.

This is evident in table 4, where the stu-dent teacher ratio measured at the school level(total students/FTE teachers) is contrasted

with the average class size reported by teach-ers. As anticipated, the former ratio is alwayssmaller than the latter. However, class sizesin urban secondary schools are unusuallylarge, given the mean student-teacher ratio.The latter is smaller by 1.4 students than theratio of suburban secondary school studentsto teachers, yet urban classes are larger bynearly two students. By contrast, in rural sec-ondary schools, lower student-teacher ratiostranslate into smaller class sizes. These dis-crepancies (which are statistically significantat conventional levels) suggest that faculty inurban secondary schools are diverted fromteaching in larger numbers than elsewhere.Other explanations, while possible, receivelittle support from the data. If urban teachershad more prep periods, class sizes would risefor that reason. However, the average num-ber of classes is virtually the same for urbanas suburban secondary school teachers. If stu-dents took more classes in the urban systems,average class size would increase, but thereis no evidence of this, either.5

Teacher Effort

More than 90 percent of instructionalspending is on salaries and benefits. Teacherabsenteeism reduces the real level of class-room resources for a given dollar expenditure.Conversely, the time teachers put in outsideschool grading homework and preparing les-son plans augments these resources.

By some indications, teacher absenteeismis a greater problem in urban schools than else-where. The first rows of table 5 summarizeteacher and administrator perceptions ofteacher absenteeism in the Schools and Staff-ing Survey (SASS). The proportion of prin-cipals who believe faculty absenteeism poses

4 The Agency Finance Information file on the Common Core of Data (CCD) contains capital outlay expenditures. However,without information on the vintage of structures and equipment, such data provide a very incomplete picture of true capitalcosts. There are no imputed rental values for buildings and durable equipment that have been fully amortized. Districts thathave recently expanded or upgraded equipment will appear to have relatively high capital costs while other systems mayappear to incur no capital costs whatever.

5 The average length of the school day is the same in the two types of districts. The same number of credits are typically requiredfor graduation.

. . . Average class

size exceeds the

school-wide

student/teacher

ratio largely

because teachers

spend fewer

hours in class

each day than do

students.


Elementary 1,025 19.1 (.16) 3856 26.6 (.29)

Secondary 725 17.1 (.21) 5005 27.0 (.28)

Combined 211 9.3 (.40) 750 16.7 (.54)

Suburban

Elementary 1,051 20.2 (.16) 3738 26.9 (.25)

Secondary 904 18.5 (.29) 6264 25.4 (.22)

Combined 143 13.0 (1.1) 589 20.7 (.77)

Rural

Elementary 2,165 18.6 (.13) 7218 25.6 (.22)

Secondary 1,979 16.4 (.14) 12071 23.5 (.16)

Combined 564 14.9 (.33) 2504 21.9 (.33)

SOURCE: U.S. Department of Education, National Center for Education Statistics, 1993–94 Schools andStaffing Survey.

at least a moderate problem is 70 percentgreater in urban schools. This perception islargely shared by teachers themselves: halfagain as many urban teachers believe facultyabsenteeism is a problem as do their counter-parts in suburban districts. In light of thesebeliefs, it is somewhat surprising that actualmeasures of teacher absenteeism reported inSASS do not differ more between urban andsuburban systems.

The limitations of the data should be bornein mind: absenteeism rates in SASS refer to asingle school day (on or just prior to the sur-vey date). Clearly, absenteeism rates on anygiven day may vary considerably for a singledistrict, though in a sample of many districtsone would expect such variation to averageout. Still, systematic differences may remain,as shown by differences in the absenteeismstatistics based on the 1993–94 survey and theearlier SASS administered in the 1990–91school year. For whatever reason, absentee-ism was higher across the board in 1993–94.Teacher attendance was better in rural districtsthan elsewhere in both years, but evidence ofan urban/suburban difference is much weaker.

When 1990 teacher absentee rates are re-gressed on a set of school characteristics in-cluding size, percentage of black and Hispanicstudents, and the percentage of students eli-gible for free or reduced-price lunch (a mea-sure of the incidence of poverty), evidence ofany difference between urban schools and oth-ers, apart from that explained by these con-trols, completely disappears. This is not reas-suring, for absenteeism increases with higherpercentages of poor and minority students.Thus, absenteeism is worst in precisely thoseschools that can least afford the loss of ser-vices of regular teachers. This may help toexplain why urban teacher absenteeism is re-garded as a greater problem in urban systemseven though the measured difference is notlarge.6

The last eight rows of table 5 contain thetime teachers report spending on school-relatedactivities outside regular school hours. Re-sponses, which refer to the most recent fullweek before the survey date, are again dis-played for the 1990–91 SASS as well as the1993–94 survey. Secondary school teachersspend substantially more time with students

Table 4.—Student/teacher ratios and class size (standard errors in parentheses)

Number of Students/teachers Number of Average

schools (school) teachers class size

Urban

6 Other reasons are possible. Qualified substitutes may be in shorter supply. Urban classes may also be harder to control whenthe regular teacher is absent.


Table 5.—Teacher absenteeism, time outside class

Urban Suburban Rural

Principals (%) perceiving teacher absenteeism as:

Serious or moderate problem 17.6% 10.3 10.4

Not a problem 39.8% 50.3 48.4

Teachers (%) perceiving teacher absenteeism as:

Serious or moderate problem 19.2 12.7 11.1

Not a problem 35.9 45.2 46.7

Teacher absenteeism (%)1

1993–94 5.8 5.8 5.0

1990–91 4.9 4.4 3.9

School-related activities involving students (hours)2•

1990–91:

Elementary teachers 1.7 1.6 1.5

Secondary teachers 4.5 4.6 5.4

1993–94:

Elementary 1.8 1.6 1.9

Secondary 4.3 4.7 5.5

Other school-related activities (preparation, grading papers,parent conferences, etc.)2

1990–91:



1993–94:



1 Data refer to most recent school day. Absentees include part-time teachers.

2 Time spent outside regular school hours during most recent full week. Full-time teachers only.

SOURCE: U.S. Department of Education, National Center for Education Statistics, Schools andStaffing Surveys, 1990–91 and 1993–94.

outside school (e.g., coaching). Elementaryteachers devote approximately one more hourper week to activities that do not involve stu-dents directly (e.g., grading papers). Differ-ences by district location are less pronounced,with urban teachers occupying an intermedi-ate position. Relative to rural teachers, theyspend less time outside school in student ac-tivities, but more on other school-related tasks.When compared to suburban teachers, the pat-tern is reversed, with the biggest difference atthe elementary level. These differences shrinkslightly when controls are added for teacherexperience, marital status, number and age ofchild dependents, subject taught, and region.

Since teachers are frequently compen-sated for the time they spend in after-schoolactivities with students, hours spent on taskslike grading papers and preparing lessons maybe a truer measure of the extra effort they areputting in. The increase in this variable be-tween 1990–91 and 1993–94 suggests that re-cent efforts to raise academic standards arehaving an effect, at least where teachers areconcerned. However, while urban teacherscompare favorably with rural instructors, theyfall behind those in suburban districts.


Table 6.—Categorical aid and special education

Urban Suburban/town Rural

Categorical aid as a percentage of instructional expenditures 15.3 9.8 11.7

State funds for special education, as percentage of instructional expenditures 5.4 4.9 4.3

Percentage of students in special education 9.1 8.3 8.7

Predicted increase in percentage of special education students from:

25% increase in students below poverty line 0.06% 0.98%** 0.32%*

$10,000 decrease in median family income -0.15% 0.45%*** 0.66%***

10% increase in percentage of students from householdswhere English is spoken ‘not well’ or ‘not at all.’ -1.4%** -1.15%** -1.11%***

Regression R2 0.01 0.05 0.02

Number of observations in estimation samples 841 2,175 8,199

*** (**) (*) Regression coefficient significant at 1 percent (5 percent) (10 percent).

NOTE: Sample restricted to independent school districts.

SOURCE: Common Core of Data, Agency Finance Information (Expenditures and Revenues), Fiscal Year1992, Household Information (1990 Census of Population).

Excessive Bureaucratization

While we have seen that urban school sys-tems (and large systems in general) actuallydevote a smaller share of total resources to ad-ministration, this is an imperfect way of gaug-ing the degree to which schools suffer fromtop-heavy bureaucracies. To explore this mat-ter further we need to consider the qualitativeside of school management (e.g., how cum-bersome and restrictive are the rules underwhich principals and teachers must operate?).

Resources are often provided by the fed-eral government and the states in the form ofaid tied to specific programs. When revenuesarrive with strings attached, administrators aredenied the flexibility to rebudget as local cir-cumstances require. Arguing for program con-solidation in special education, McLaughlin(1996) writes:

“[T]here is a long way to go in cre-ating the types of flexible educa-tional systems that are being pro-moted in current federal and state

restructuring initiatives...[A] deeplyembedded culture of program sepa-ration appears to support turf guard-ing and reinforce the belief that ‘dif-ferent types’ of students need verydifferent educational experiences.”

Other researchers have commented on in-creasing specialization and bureaucratizationin elementary and secondary education, inwhich a proliferation of mandates and targetedprograms results in “the creation of special-ties with an ever-narrowing realm of expertisefor each specialist” (Raywid and Shaheen,1994). As responsibility for school operationsis parcelled out among a variety of adminis-trators, each focused narrowly on theprogram(s) for which he is accountable, op-portunities are diminished to balance compet-ing interests in order that reasonable tradeoffsbe made among various goals.

“Is there some way...that we can holdofficials responsible in any signifi-cant way for more than their ownoperations?... Ultimately, this is the


sort of question that must be ad-dressed if we are ever to make ma-jor improvements as to the cost-ef-fectiveness of all schools... Solu-tions are difficult, but it seems safeto conclude that minimally it willrequire removing the present incen-tives to focus narrowly and to delib-erately ignore the broader context”(Raywid and Shaheen, 1994).

Are such problems particularly serious inurban schools? Table 6 displays the propor-tion of instructional expenditures financed withcategorial or “tied” aid. Included are revenuesfrom state or federal sources for the followingprograms: special education, compensatory orbasic skills education, bilingual education, pro-grams for the gifted and talented or childrenwith disabilities, and Chapter 1 aid. Funds re-ceived for non-instructional purposes (e.g.,child nutrition, transportation) are excludedfrom this figure. As before, summary statis-tics are presented for urban schools as a group,for suburban schools, and for rural schools.Because so much attention has focused on thegrowing share of resources devoted to specialeducation, state aid received for special edu-cation is broken out in row two. For purposesof comparison, an alternative measure of therelative size of special education programs—the proportion of students with individualizededucation programs (mandated by law for allspecial education students)—is also provided.

Urban districts finance a significantlyhigher share of instructional expenditures fromcategorical aid. While one might suspect thatthis difference is due to higher concentrationsof poverty and other social problems in inner-city neighborhoods, this turns out not to be thecase. When the share of categorical aid is re-gressed on the household characteristics thatappear in table 2 plus the percentage of house-holds in which English is spoken ‘not well’ or‘not at all,’ the estimated gap between urbanand other districts widens to more than 8 per-cent. The percentage of children below thepoverty line is, of course, a strongly signifi-

cant predictor of the amount of categorical aida district receives. An increase of one stan-dard deviation in the poverty rate—about 13percent—raises the share of categorical aidby 4 percentage points. But unmeasured fac-tors contribute importantly to the amount offederal and state aid received in these catego-ries.

Growing special education expenditureshave attracted particular concern. Apart fromthe fact that special education has proven tobe enormously expensive, absorbing resourcesthat could be devoted to general education,questions have been raised about the appro-priateness of many placements. Reports inthe press have described a variety of abuses:students who are placed in special educationbecause they speak English poorly; racial andethnic minorities who are discriminatedagainst by teachers who underestimate theircognitive abilities and misread behaviorshaped by unfamiliar cultural backgrounds;districts that place large percentages of stu-dents into special education to obtain extrastate and federal revenues. To investigatethese concerns, the percentage of studentsplaced in special education was regressed onthe household characteristics in table 2 plusthe following additional regressors: the per-centage of households in which English isspoken not well or not at all, the percentageof school-age children who belong to racialor ethnic minorities (blacks, Hispanics, Na-tive Americans, Asians) and per-pupil currentexpenditures less state aid received for spe-cial education. Inclusion of this last variableallows us to examine whether districts withfewer resources apart from special educationaid respond by placing more students in spe-cial education, other things equal.

The lower panel of table 6 presents se-lected results. While there are doubtless prob-lems in some districts, these results do not sup-port the notion that special education plays adisproportionate role in the schooling of theeconomically and socially disadvantaged.Very large changes in median income or pov-

Growing special

education

expenditures

have attracted

particular

concern.


Table 7.—Principals’ influence and autonomy

Public Private

Urban Suburban Rural Urban All

Percent of principals/heads indicating they have ‘a greatdeal’ of influence over:

Curriculum 17.9 20.0 23.4 65.4 63.6

Hiring 52.5 62.0 61.4 81.6 80.2

Discipline policy 52.5 58.7 56.7 82.0 80.8

How budget is spent 36.2 36.3 28.3 63.5 63.0

Percent of principals/heads indicating school or governingboards have little or no influence over:

Curriculum 12.9 13.6 17.9 33.5 35.4

Hiring 35.6 34.1 21.0 46.5 44.7

Discipline policy 8.9 8.1 6.7 34.0 31.9

How budget is spent 15.0 12.3 6.8 25.4 25.5

Percent of principals indicating little or no influence by statedepartment of education, district staff, or school board over:

Curriculum .6 .85 .4 — —

Hiring 9.5 10.5 8.5 — —

Discipline policy 4.4 4.8 3.7 — —

How budget is spent 6.6 4.9 2.8 — —

— Not applicable.

SOURCE: U.S. Department of Education, National Center for Education Statistics, Schools and StaffingSurvey, 1993–94.

erty rates within the district have virtually neg-ligible impacts on the percentage of specialeducation students. There is no evidence thatstudents with English language problems arebeing shunted into special education on a sys-tematic basis, either in urban districts or else-where. Except in rural systems, an increasein the percentage of minority students actu-ally reduces special education placements,though the effects are very small. The verylow R2 in each of these equations is reassur-ing, as it implies that special education en-rollments are not a function of students’ socio-economic characteristics.

In summary, while this analysis has notfound signs of systematic abuses in specialeducation placement (and certainly no evi-dence that there is more abuse in urban sys-tems than elsewhere), urban districts do re-ceive a significantly higher proportion of rev-enues as programmatic aid. If the views citedabove are correct, the regulations and over-

sight that accompany such funding may con-strain local decision makers. This is only onereason why urban administrators and teachersmight enjoy less autonomy and flexibility thaneducators elsewhere. In addition, the well-publicized problems of urban schools mayhave prompted efforts to fix the system fromabove by imposing additional rules and con-straints on teachers and principals. The sheersize of urban school systems is apt to enhancethe power and prerogatives of central districtbureaucracies. As a result, administrators atthe school level may find themselves unableto allocate funds as cost-effectively as possibleor to hire job applicants of their own choosingin a timely manner, to cite only two policy con-cerns.

Describing reforms in school finance thatwould provide a foundation for higher studentachievement, Allan Odden identifies “a focuson the school as the key organizational unit”and the “devolution of power over the budget


and personnel to schools” as key components(Odden,1994).

“Findings from multiple strands ofresearch suggest that a decentral-ized, high involvement organiza-tion and management strategy (i.e.,school-based management) shouldexplicitly be made part of systemicreform. This research concludesthat SBM would work most effec-tively if information, knowledge,power and rewards are decentral-ized to the school level.”

How far public schools are from realiz-ing this objective is shown, in part, by princi-pals’ perception of the limits of their author-ity. The top panel of table 7 displays responsesto the 1993–94 SASS on the part of publicschool principals and private school headswhen questioned about their influence overcurriculum, hiring, discipline, and the budget.While urban principals generally indicate theyhave less influence than do their counterpartsin suburban and rural districts, the most strik-ing contrast is between public principals andprivate school heads, who have substantiallymore say about the way their schools are runin each policy area.

Also important is the extent to which prin-cipals’ managerial prerogatives are con-strained by decisions taken at higher levels.The middle panel displays the percentage whoindicated that school boards (governing or di-ocesan boards in the case of private schools)exercised little or no influence over policy.Again, responses show that private schoolheads are far more likely to run their schoolswithout interference from above. In fact, theseresponses understate the magnitude of thistype of interference in the public sector, wherestate Departments of Education and centraldistrict offices also exercise regulatory over-sight and shape educational policy. The bot-tom panel of table 7 displays the percentageof public school principals who indicated thatnone of these other bodies had appreciable in-

fluence over policy in the same four areas. Asone would expect, the percentages are verysmall.

This is not to suggest that public schoolswould be better managed if school boards andDepartments of Education exercised no regu-latory oversight. Under the present system ofpublic education, this oversight is the princi-pal means by which schools financed with tax-payer dollars are held accountable to the pub-lic. What the comparison with private schoolsreveals is that alternative mechanisms for pre-serving accountability exist that offer schoolheads considerably more autonomy. The chiefmechanisms within private education are, ofcourse, the competitive market and consumersovereignty.

Much of the current interest in schoolchoice within public education derives fromthe belief that educational performance willimprove if public schools are also exposed tocompetitive market forces. By creating op-portunities for parents to select other schoolsif they are not satisfied with the school to whichtheir child was assigned by virtue of residen-tial location, choice plans put pressure on ad-ministrators and teachers to correct deficien-cies in their programs.

Responses to the 1993–94 SASS show thatnearly half of all urban school systems offerparents some form of school choice. One-fifthhave established one or more magnet schools,one-fourth offer choice of schools within thedistrict, and nearly 40 percent allow parentsto choose schools outside the district. An al-most equal percentage accept students fromother districts. All of these measures are higherthan the corresponding rates among non-ur-ban schools.

Whether these plans are likely to improveefficiency is another matter, however. Paren-tal participation rates are much less impres-sive. Only 7 percent of the students in urbansystems containing magnet schools actually at-tend one of these schools (though this is more

Much of the

current interest

in school choice

within public

education derives

from the belief

that educational

performance will

improve if public

schools are also

exposed to

competitive

market forces.


Table 8.—Salary incentives in public schools

Purpose of incentive Percent of Percent of Percent of Percent of schools and location districts schools teachers with unfilled vacancy

Shortage subject Urban 9.3 23.6 30.4 37.2 Suburban 8.8 13.5 13.6 23.8

Rural 8.3 8.7 8.9 14.2

Undesirable location

Urban 4.4 11.1 13.7 17.2

Suburban 2.8 6.9 6.7 9.4

Rural 5.2 5.3 5.4 4.9

Merit pay

Urban 15.0 16.5 16.3 14.7

Suburban 5.9 10.7 10.7 9.6

Rural 12.1 13.4 13.7 13.4


than twice the rate of suburban and rural sys-tems). Participation rates in other choice plansare still lower and do not differ systematicallyby urbanicity. In those urban districts that al-low within-district choice, only 11 percent ofstudents actually exercise it. Ratios are sub-stantially lower for inter-district plans. Thereis, moreover, a significant difference betweenurban districts and others in the direction inwhich students are likely to travel: while ur-ban systems are more likely to receive studentsfrom other systems than to see their own stu-dents leave, the reverse is true of suburban andrural districts. This may indicate that urbanstudents are at a relative disadvantage in learn-ing about opportunities outside the district orfinding transportation into neighboring com-munities. It may also show that these commu-nities have found ways to discourage the par-ticipation of inner-city students.

In sum, urban systems are more likely tooffer various types of school choice than aresuburban or rural districts. However, partici-pation rates are low. Combined with evidencethat urban students may have fewer de factoopportunities to attend schools outside theirhome districts, it seems doubtful that schoolchoice, at least in most communities where it

is found, operates on the scale needed to havea significant effect on school performance.

Teacher Salaries

Teacher compensation in public schoolsis determined by salary schedules that rewardteachers for experience (and/or seniority in thedistrict) and for earning advanced degrees orcollege credits. As a rule, schedules make nodistinction by subject taught or quality ofteaching performance. Compensation for sub-ject area knowledge of teaching expertise isgenerally provided, if at all, through add-onssuch as merit pay or policies that allow ad-ministrators to make exceptions to the sched-ule (e.g., placing a teacher on a higher stepthan he would be entitled to on the basis ofeducation and experience). These special pro-visions aside, the use of single salary sched-ules to determine the compensation of allteachers in a district has been criticized for(1) inflexibility in the face of varying marketconditions; (2) rewarding attributes that bearlittle or no observed relationship to teachingeffectiveness (e.g., advanced degrees);(3) providing no incentive for improved per-formance.


As shown in table 8, the majority of schooldistricts do not use special incentives to re-cruit teachers in subjects where there is ashortage of qualified instructors, to staff po-sitions in undesirable locations (e.g., high-crime, high-poverty inner city neighbor-hoods), or to reward merit. Urban districtsare somewhat more likely to use these incen-tives than other systems. District size alsohas an important influence on whether pay in-centives are available to recruit teachers inshortage areas: although the percentage ofurban systems that use such incentives is only9 percent, fully 30 percent of urban teacherswork in these systems. Similarly, almost 14percent of urban teachers work in systems thatreward teachers for accepting a position in anundesirable location (though only 4.4 percentof districts use incentives for this purpose).Finally, schools were more likely to use theseincentives if they had one or more unfilledvacancies, suggesting that salary flexibility ismore likely to be found in districts that havetrouble recruiting.

Table 9 displays further information onthis point. Schools are distinguished not onlyby urbanicity but also by the ease with whichthey recruited teachers in the seven subjectslisted. Schools classed as D reported that theyfound it very difficult or impossible to fill avacancy in these subjects; the remainder, ND,found it easy or only moderately difficult.(Schools that did not recruit in these subjectsare omitted from the analysis.) Two thingsstand out. As a rule, schools that had troublefilling positions were more likely to use somekind of incentive pay for teachers in that sub-ject. This is especially true of urban schools.However, in no category did the use of incen-tive pay even approach 50 percent. Thus, toofew schools use these incentives, while inthose that do use them, the extra pay does not

appear to have solved the problem: recruitmentin these subjects remains a problem.

Unfortunately, SASS did not ask teacherswho received these incentives how much ex-tra compensation they obtained. As a result,this question must be investigated by estimat-ing teacher earnings equations. The estima-tion sample comprised full-time teachers fromthe 1990–91 SASS. The dependent variablewas the natural logarithm of a teacher’s basesalary plus bonuses. Independent variables in-cluded controls for starting pay within the dis-trict and for a teacher’s education and experi-ence. The data contained discrepancies: someteachers claimed to receive extra compensa-tion from districts that did not acknowledgeusing the incentive in question. Statisticalanalysis suggested that most of these cases rep-resented response error on the teachers’ part.7

As a result, only those teachers who claimedto receive extra compensation from districtsaffirming the use of such an incentive weretreated as bona fide recipients.8

Selected results are displayed in table 10.Coefficients on incentive pay in the public sec-tor equation are small and almost always sta-tistically insignificant. The largest in magni-tude, for teaching in an undesirable location,are actually of the wrong sign (though impre-cisely estimated). Only merit pay in ruralschools enters with a significant positive co-efficient.

Although there were not enough observa-tions in the private school sample to estimateseparate coefficients for urban schools, theoverall results suggest that merit pay makes asignificantly larger contribution to the salariesof private school recipients. In fact, the dif-ference is considerably understated by the co-efficients in table 10. Further analysis of re-

7 As a group, these teachers were paid no more than other instructors at the same schools (controlling for experience andeducation). In fact, there was marginally significant evidence in the case of self-styled merit pay recipients that they receivedless.

8 A second set of dummy variables identified all teachers (not just recipients) employed in districts with special incentives forteaching in shortage fields and undesirably locations, for merit, and for mentoring. These additional controls were introducedso that the coefficients on incentive recipients would not pick up purely district level effects.

. . . schools were

more likely to use

these incentives

if they had one

or more unfilled

vacancies,

suggesting that

salary flexibility

is more likely to

be found in

districts that have

trouble

recruiting.


Table 10.—Teacher earnings (standard errors in parentheses)

Percentage change in salary for Public Private1

Shortage subject 3.6 (5.4)

Urban 1.2 (1.3)

Suburban -.4 (2.5)

Rural -.3 (2.1)

Undesirable location (2)

Urban -1.1 (1.9)

Suburban -3.2 (3.3)

Rural -5.4 (3.6)

Merit pay 8.7 (2.5)***

Urban 1.3 (1.4)

Suburban .7 (1.3)

Rural 2.6 (.9)***

Elementary level -.5 (.2)*** -3.7 (.8)***

Number of observations 38,069 3,576

R2 0.76 0.69

*** Coefficient significant at 1 percent.

1 Excludes teachers employed in schools that do not use salary schedules and teachers contributingservices for less than market wages (e.g., members of religious orders).

2Not asked of private school teachers.

NOTE: Additional regressors included district’s starting pay for new teacher with a bachelor’s degree,additional pay for new teacher with master’s degree, average annual increment in pay for each additionalyear of experience (censored at 20 years), previous part-time experience, possession of sixth-yearcertificate or Ed.D., marital status, race (black), ethnicity (Hispanic), age, gender. Private school samplealso includes binary indicators for teachers receiving in-kind compensation (tuition for faculty children,meals, housing), Catholic and other-religious schools.


Table 9.—Percentage of schools using pay incentives to recruit teachers,by shortage area

Urban Suburban Rural

Subject area Not difficult Difficult Not difficult Difficult Not difficult Difficult

English as a second language 35.7 29.3 6.5 4.5 7.3 7.7Biology 3.5 27.2 2.6 0.0 1.9 2.8Physics 4.5 32.4 2.9 0.0 1.7 7.5Mathematics 7.6 30.4 3.8 7.0 2.9 8.2Special education 17.1 24.0 8.9 15.7 4.9 6.1Foreign languages 4.0 12.8 1.7 5.7 1.4 2.7Vocational education 4.4 7.2 1.8 5.5 1.9 2.3

NOTE: Schools that did not recruit in specified subjects were not used in computations.



sponses to the 1990–91 SASS shows that mostprivate schools using merit pay award it as astep increase on the salary schedule or buildit into the teacher’s base in some other man-ner. Fewer than 30 percent make one-timecash awards. By contrast, more than 60 per-cent of the public schools that use merit payaward it as a one-time cash bonus. Thus, notonly are merit awards larger at a single pointin time in the private sector, but these awardsare more likely to be received on a recurringbasis.

Table 10 also shows that public elemen-tary school teachers earn virtually the samesalaries as secondary teachers. This is not sur-prising, of course, given the widespread adop-tion of uniform salary schedules for all teach-ers in a district. It also shows how unrespon-sive public school salaries are to market con-ditions. By every indication schools have littledifficulty recruiting elementary teachers.Eighty-four percent of the public schools thatrecruited elementary teachers reported toSASS that it was ‘easy’ to fill these vacan-cies. By contrast, the percentage for physicswas 50 percent, for mathematics 58 percent,and for foreign languages 42 percent. Yetteachers in all subject areas are paid accord-ing to the same schedule.

On this score, compensation policies inthe private sector appear to be just about asrigid, since the estimated difference for el-ementary teachers in the private school equa-tion is also small, just under 4 percent. How-ever, the model controlled for starting pay atthe school as well as the salary increments(again at the school) for teachers who obtaina master’s degree and for an additional yearof experience. Since most of the differencebetween elementary salaries and secondarysalaries in the private sector arises betweenschools rather than within a school, the coef-ficient in table 10 substantially understates theamount by which elementary and secondarysalaries differ. This is clearly seen whenschool-level controls are removed from the

model and teacher pay is regressed on teacherexperience, education, and a dummy variablefor school level: elementary school teachersin the private sector earn an average of 16 per-cent less than secondary teachers with compa-rable degrees and experience. The differenceremains substantial (13 percent) when controlsare added for race, gender, marital status, andage. When the same equation is run for thepublic sector, the gap between elementary andsecondary pay on average is only 2 percent.With the addition of demographic variables itfalls to 1 percent.

Summary and Conclusion

In some respects, urban public schoolscompare favorably with public school systemselsewhere. The proportion of current expen-ditures allocated to instruction is no lower thanin suburban and rural districts. Urban districtsare more likely to use pay incentives to recruitteachers, particularly in areas where qualifiedinstructors are in short supply. They are alsomore likely to offer students and their parentssome form of school choice. They occupy anintermediate position between suburban andrural districts with respect to the time teachersdevote to school-related activities outside regu-lar school hours. Although a slightly largerpercentage of urban students are enrolled inspecial education, there is no evidence of sys-tematic abuses (i.e., increasing special educa-tion enrollments associated with poverty, race,ethnicity, or use of language other than En-glish at home). This is not to say that urbanschools could not accomplish more with theresources they have, only that on these countsthey appear to be following as effective a setof policies as public school systems in sub-urbs, towns, and rural communities.

By several indications, however, there areproblems with the urban policy mix. First,there is virtually no evidence that urban schoolsystems are benefitting from economies ofscale at the district level. The average districthas three times the enrollment of the average

[While], in some

respects, urban

public schools

compare

favorably with

public school

systems . . . there

are problems

with the urban

policy mix.


suburban district, yet there appear to be no sav-ings in administration or other central officeoperations. This suggests that the typical ur-ban district exceeds the size at which scaleeconomies have been realized. Similarly, whilethere is evidence of scale economies at theschool level, the savings per student is quitelow, on the order of $25 to $50. Given find-ings in the education production literature thatstudents benefit from smaller, more personallearning environments, one must questionwhether savings of this magnitude justify cur-rent school sizes.

A larger proportion of urban revenues isreceived as programmatic aid, a circumstancethat tends to increase administrative costs anddeprives local officials of flexibility. Teacherabsenteeism appears to be a greater problem,though not necessarily because absentee ratesare actually higher. Rather, urban districts mayhave more difficulty finding (or affording) ca-pable substitutes or dealing with the disrup-tions caused when regular classroom teachersare not present. There is some evidence, al-

Some of the

comparisons that

appear to favor

urban schools

turn out to be

less favorable

when one looks

beneath the

surface.

beit indirect, that urban systems also employmore teachers in non-teaching roles: classsizes tend to be larger, though aggregate stu-dent/teacher ratios are actually lower.

Some of the comparisons that appear tofavor urban schools turn out to be less favor-able when one looks beneath the surface. Al-though more urban systems have establishedschool choice programs, the proportion of stu-dents who actually participate in these pro-grams is low and not very different from thatfound in suburban and rural systems. On pa-per there is choice, but in reality few familiesexercise it. Similarly, while a much higherproportion of urban systems indicate that theyuse salary incentives to recruit teachers, es-pecially in shortage subject areas, most of thedistricts that do so continue to experience dif-ficulty recruiting. Moreover, analysis ofteacher salaries fails to find any evidence thatteachers who receive these incentives (by theirown report) are actually paid more than thosewho do not.


References

Chubb, J.E. and T.M. Moe. 1990. Politics, Markets and America’s Schools. Washington, DC: The BrookingsInstitution.

Fowler, W.J., Jr. and H.J. Walberg. 1991. “School Size, Characteristics, and Outcomes.” EducationalEvaluation and Policy Analysis, 13(2): 189–202.

Hanushek, E.A. et al. 1994. Making Schools Work: Improving Performance and Controlling Costs. Wash-ington, DC: The Brookings Institution.

Lippman, L. 1996. Urban Schools: The Challenge of Location and Poverty. Washington, DC: U.S.Department of Education, National Center for Education Statistics.

McLaughlin, M.J. 1996. “Consolidating Categorial Program Resources to Support Local School Improve-ment: Issues and Perspectives.” Journal of Education Finance, 21: 506–526.

Odden, A. 1994. “Decentralized Management and School Finance.” Theory Into Practice, 33(2): 104–111.

Picus, L.O., Bhimani, M. 1993. “Determinants of Pupil/Teacher Ratios at School Sites: Evidence from theSchools and Staffing Survey.” Paper prepared for the American Statistical Association 1993 Proceedings ofthe Social Statistics Section. Alexandria VA: ASA.

Raywid, M., Shaheen, T.A. 1994. “In Search of Cost-Effective Schools.” Theory Into Practice 33(2): 67–74.

Walberg, H.J. 1994. “Educational Productivity: Urgent Needs and New Remedies.” Theory Into Practice33(2): 75–82.

Walberg. H.J. and W.J. Fowler, Jr. 1987. “Expenditure and Size Efficiencies of Public School Districts.”Educational Researcher 16(7): 5–13.

Wilson, S.F. 1992. Reinventing the Schools: A Radical Plan for Boston. Boston, MA: The PioneerInstitute.

the condition of urban school finance: efficient resource ... · the condition of urban school...

Documents