how has teacher compensation changed?nces.ed.gov/pubs2001/2001378_2.pdf · how has teacher...

How Has Teacher Compensation Changed?

11

SelectedPapers in

SchoolFinance,2000–01

How Has TeacherCompensation Changed?

Selected Papers in School Finance, 2000–01

12


13

Dan D. GoldhaberThe Urban Institute

About the Author

Dan D. Goldhaber is a Senior Research As-sociate at the Education Policy Center, Ur-ban Institute, Washington, DC. He is aneconomist who specializes in the econom-ics of education and education finance. Hisresearch focuses on issues of educationalreform and productivity at the K–12 level,and the relationships between teacher la-bor markets and teacher quality. Topics ofrecent published work include: the effectsof teacher qualifications and quality on stu-dent achievement; the relative efficiency ofpublic and private schools; and policyanalyses of market competition, in the formof education vouchers, on K–12 schooling.Publications on these topics have appearedin the Journal of Human Resources, theJournal of Urban Economics, Economicsof Education Review, Journal of EducationFinance, Education Economics, EducationMatters, Industrial and Labor RelationsReview, Educational Evaluation and PolicyAnalysis, Education Researcher, and PhiDelta Kappan.


Dr. Goldhaber is currently conducting re-search addressing the effects of the FloridaOpportunity Scholarship (voucher) Pro-gram on schools, teachers, and students.Other research includes an impact study ofthe implementation of various comprehen-sive school reform models, and analyses ofthe effects of National Board certificationon teachers and students in North Caro-lina.

Prior to joining the Urban Institute, Dr.Goldhaber served as the Assistant Direc-tor for Education at the CNA Corporationon Alexandria, VA. Dr. Goldhaber is alsoan elected member of the Alexandria CitySchool Board where he has served since1997 and an adjunct faculty member at theGeorgetown University Public Policy Insti-tute. He received a BA from the Universityof Vermont and an MS and Ph.D. in LaborEconomics from Cornell University.


14


15

Introduction

How has teacher compensation changed?The answer to this question is of crucialpolicy significance to researchers attempt-ing to assess the relationship betweenteachers’ salaries and benefits, teacherquality, and ultimately student outcomes;and to policymakers trying to craft effec-tive teacher compensation policies. Thisquestion is also particularly pertinent atthis point in time as school systems struggleto find the large numbers of new teachersthat will be needed to account for increas-ing student enrollments and an aging teach-ing labor force (Lankford and Wyckoff1997).

It is clear that over the long term, relativeteacher compensation plays an importantrole in influencing the decision to enter andleave the profession. Expressed interest inteaching as a career tends to track closelywith fluctuations in relative teacher sala-ries. For instance, median starting salariesin education lost ground (in both real termsand relative to most other fields of study)in the 1970s through the mid-1980s (U.S.Department of Education 1997). Duringthis period, the percentage of freshman re-

porting an interest in a (elementary or sec-ondary) teaching career declined from over20 percent in the late 1960s to a low of 4.7percent in 1982 (Astin et al. 1997). Sincethen, both salary and career interest inteaching has recovered somewhat, but it hasremained significantly below the 20 percentlevel. This fact provides some indicationthat individuals are influenced by long-termlabor market incentives when making ca-reer choices. All else equal, as teachers’ sala-ries and benefits rise relative to those inother professions, teaching becomes a moreattractive field and higher ability individu-als will enter the profession.

Given that the level of compensation in anoccupation plays an important role in at-tracting or dissuading able individuals fromentering and exiting that occupation, thequestion is, “How has compensation inteaching changed?” The answer to thisquestion is more complicated than onemight initially imagine. For instance, basedon data from the U.S. Bureau of the Cen-sus (1999), we know that between 1980and 1997 teachers’ salaries rose by roughly120 percent. This statistic might indicateto many that teaching is an attractive pro-

Dan D. GoldhaberThe Urban Institute



16

1 In nominal terms, salaries rose from $17,644 in 1980–81 to $38,509 in 1996–97. Adjusted for inflation this is equivalent to a realincrease of about $6,000 (in 1996–97 dollars).

2 The focus in this paper is on public school teachers so, except where explicitly noted, teachers and teacher compensation will refer topublic school teachers and compensation in the public sector.

3 “Compensation” generally refers to both pay and benefits; here, however, unless otherwise noted, we use the terms “teachercompensation,”“teacher pay,” and “teacher salary” interchangeably.

fession in which to be. However, thoughthe increase may sound strikingly large,this fact alone does not provide a goodindication of how financially attractive itis to be a teacher. For instance, the 120percent increase in salaries is only equiva-lent to a 19 percent increase once adjust-ments are made for inflation. During thissame period salaries overall increased by29 percent (U.S. Bureau of the Census1999).1 Thus, when judged by salary alone,teaching is a less attractive occupation tobe in today than it was in 1980. This ex-ample illustrates the difficulty, given con-temporaneous changes in cost of living andcompensation in other fields, in determin-ing precisely how compensation changesfor teachers changes the incentives to en-ter or remain in teaching.

This paper compares and contrasts severalalternate measures of teacher compensa-tion over time, focusing on how teachercompensation has changed relative to com-pensation in the labor market as a wholeand changed relative to specific professionsthat compete with teaching for collegegraduates.2 Specifically, data from the Bu-reau of Labor Statistics’ employment costindex (ECI) are used to make these com-parisons over time and, where applicable,compare the findings with other data de-tailing changes in educational costs. Fur-ther, the implications of these findings inrelation to the anticipated shortage ofteachers over the next decade are dis-cussed.

This paper is laid out in four sections. Thefirst section provides motivation for thistopic by providing an overview and briefreview of the existing empirical literaturelinking teacher quality and teachercompensation.3 The second section focuseson the structure of compensation in teach-ing and how it compares with the struc-

ture of compensation in other sectors of theeconomy. This section helps to frame theissues by shedding light on the relative at-tractiveness of entering teaching for indi-viduals with different backgrounds and cre-dentials. In the third section, data onchanges over time in teacher compensationare presented and compared with changesin educational costs and to compensationin other fields. The implications of thesechanges are discussed in the final section,which also focuses on the changing incen-tives to enter and remain in teaching forindividuals with different backgrounds.

Teacher Quality andCompensation

Recent research provides strong evidencethat teacher quality is the single most im-portant school factor affecting studentachievement. Both Wright, Horn, and Sand-ers (1997) and Rivkin, Hanushek, and Kain(1998) find there is a wide range of effec-tiveness among teachers and that teacherquality accounts for a much larger share ofthe total variation in student achievementthan all other educational resources. Stud-ies tend to show that teachers who scorehigher on standardized exams and attendmore selective colleges tend to be more ef-fective (Ehrenberg and Brewer 1994, 1995;Ferguson 1991, 1998; Strauss and Sawyer1986). Research also shows that teacherpreparation in mathematics and science hasa positive impact on student achievementin those subjects (Monk and Rice 1994,Goldhaber and Brewer 1997). Thus, if weare to define “teacher quality” accordingto quantifiable attributes that impact stu-dent outcomes, three measures that wouldappear to serve as good indicators of qual-ity include degree in subject, performanceon standardized exams, and selectivity ofcollege attended.

Recent research

provides strong

evidence that

teacher quality is

the single most

important school

factor affecting

student

achievement.


17

These findings are important because theysuggest that teachers can have an impor-tant impact on students’ achievement andthere are particular attributes or creden-tials that school systems may look to intrying to recruit successful teachers. How-ever, it appears that the most importantteacher attributes are difficult to identifyin data. For instance, the ability of teach-ers to convey knowledge or their enthusi-asm for class material might have a dra-matic impact on students, but these char-acteristics are difficult to quantify and maynot be related to identifiable characteris-tics that are more easily measured.Goldhaber, Brewer, and Anderson (1999)investigated the contributions of school,teacher, and class characteristics on studentachievement. They found about 3 percentof the contribution teachers make towardexplaining student achievement is associ-ated with teacher experience, degree level,and other readily observable characteris-tics. The remaining 97 percent is made upof teacher qualities or behaviors that couldnot be separately isolated and identified.

Several studies investigate the relationshipbetween the selectivity of college attendedand the subsequent labor market outcomesof teachers. Both Ballou (1996) and Cham-bers (1998) find that teachers graduatingfrom more selective colleges receive highersalaries.4 Figlio (1997) finds that schooldistricts that offer higher salaries are morelikely to hire teachers from selective col-leges. However, Ballou (1996) finds thatin teaching, unlike in other fields (e.g.; en-gineering, accounting, business administra-tion, computer science, and psychology)graduating from a “selective” college (the

top of four college quality categories) doesnot have a statistically significant impacton the probability of finding a job in one’sfield of study. In addition, relatively fewstudies have found a direct link betweenteachers’ salaries and student outcomes(Hanushek 1986, 1997).

One possible explanation for these some-what mixed findings is that there is not astrong demand for high quality teachers.Higher salaries may attract a larger poolof qualified job applicants, but if schooladministrators do not prefer stronger jobcandidates, the quality of the teacherworkforce may not improve (Strauss et al.1998).5 For instance, despite the fact thathaving a degree in subject is a better pre-dictor of teaching proficiency than a de-gree in education (Goldhaber and Brewer1997, 1998), individuals who wish to teachin mathematics or sciences in high schoolare better off majoring in education with ateaching field in mathematics or sciencethan completing an academic major inthose subjects (Ballou and Podgursky1997). Additionally, both Ballou (1996) andManski (1987) find that increases in teach-ers’ salaries increase the applicant pool ofnew teachers but do not affect the abilitydistribution of teachers.6

It is also possible that researchers investi-gating the relationship between teachers’salaries and teacher quality and student out-comes have failed to adequately accountfor important non-pecuniary job character-istics that are positively correlated with theattractiveness of a teaching job but nega-tively correlated with teachers’ salaries.7

Ballou (1996) finds that graduates of bet-ter colleges are more likely to be found in

4 Ballou (1996) estimates a salary advantage of about 6 percent in teaching for those that fall into the highest college quality categoryrather than into the lowest.

5 Also, Ballou and Podgursky (1995) show, from a theoretical perspective, that it is possible for increases in salary to result in a decline inthe quality of the teacher workforce. They find that the pool of teacher applicants will rise with salary increases. However, currentteachers may also elect to stay longer in teaching, reducing the number of available teaching slots. This, in turn, lowers the probabilityof obtaining a job in teaching which may have the biggest impact on high quality applicants who have attractive alternatives to teaching.

6 A study of public school districts in Pennsylvania suggests that school districts may not actively search for or screen teacher applicants(Strauss et al. 1998). For instance, most school districts were not found to advertise or look outside of the local labor market to find newteachers (only 25 percent of school districts advertised outside of the state, and, on average, 60 percent of newly hired school teacherscame from institutions no more than 70 miles away from the hiring school district).

7 For instance, teachers’ salaries in New York City are significantly higher than teachers’ salaries in most parts of Iowa; however, much ofthe higher New York salary must be considered a compensating differential for a higher cost of living and a greater at-risk studentpopulation.

Several studies

investigate the

relationship

between the

selectivity of

college attended

and the subsequent

labor market

outcomes of

teachers.


18

schools with smaller percentages of poorstudents and higher percentages of studentsbound for college. Chambers (1998) showsthat there are large compensating differ-entials for community characteristics, suchas cost of living, crime rates, and weather.Loeb and Page (1998), who focus on thisissue, find a strong relationship betweenteachers’ salaries and high school dropoutrates and the likelihood of college atten-dance, only after carefully controlling fornon-pecuniary school district characteris-tics.

A final possibility is that studies have failedto adequately account for alternative la-bor market opportunities for teachers.Many “teacher salary-teacher quality”studies do not measure teachers’ salariesrelative to wages in other occupations,8

which is a potentially important omissionsince economic theory predicts that in-creases in salary would attract more quali-fied individuals into a particular field onlyif salaries in that field were rising relativeto what individuals could make in otherareas. The select few studies that do con-trol for alternative labor market opportu-nities (see, for instance, Ballou andPodgursky 1997; Flyer and Rosen 1994;Hanushek and Rivkin 1997; Loeb and Page1998; Hanushek, Kain, and Rivkin 1998)treat all teachers as if their opportunitycosts are the same regardless of field of spe-cialization or quality of college attended.

One of the primary reasons that previousstudies have failed to account for some ofthe factors that likely affect studies of therelationship between teacher compensationand teacher quality is lack of adequate data.The primary problem is that sample sizesare usually not large enough to disaggre-gate the data enough to make “ideal” com-parisons between individuals who enter

teaching to those with similar backgrounds(e.g., college quality and major) who enteran alternate occupation. Given that we willencounter some of the same difficultieshere, it is important to keep these facts inmind as we go on in subsequent sections ofthis paper to discuss how teacher compen-sation has changed relative to that in otheroccupations. In particular, although theattractiveness of teaching as a profession islikely to rise or fall with overall increasesin relative teacher compensation, it is notlikely to be the same for individuals withdifferent backgrounds. These subtleties arediscussed at greater length in subsequentsections.

The Structure of TeacherCompensation

Teacher quality is difficult to predict whenevaluating teacher candidates and perfor-mance on the job is difficult to measure.For these reasons teacher compensation isoften linked to credentials that may serveas indirect indicators of knowledge andskills, rather than determined on an indi-vidual by individual basis. In fact, most,but not all, public school systems use the“single” or “uniform” salary schedule todetermine how much a teacher will earn(Odden and Kelley 1997).9 Such a sched-ule typically sets teacher pay solely on thebasis of teaching experience and teacherdegree level.10 The average pay premiumsfor advanced degrees are about 11 percentfor a master’s degree, 14 percent for aneducation specialists degree, and 17 percentfor a doctorate degree (U.S. Department ofEducation 1996b). The premium for an ad-ditional year of service is typically between$1,000 and $1,500 (Odden and Kelley1997).

8 See Loeb and Page (1998) for a more detailed discussion on this topic.

9 The uniform salary schedule dates back to 1921 when Denver, Colorado and Des Moines, Iowa became the first cities to adopt it. By the1950s, 97 percent of school districts had adopted this pay structure (Odden and Kelley 1997) and today approximately 95 percent ofschool districts structure compensation in this manner.

10 This is despite the fact that there is relatively little quantitative evidence that teacher experience and degree level, per se, are correlatedwith student outcomes (Murnane 1996; Hanushek 1986, 1997; Ballou and Podgursky 1997; Goldhaber and Brewer 1998).

Teacher quality is

difficult to predict

when evaluating

teacher candidates

and performance

on the job is

difficult to

measure.


19

Though it is far from the norm in publicschools, there is some differentiation incompensation structure across districts. Forinstance, there has been an increase in re-cent years in the number of public schooldistricts that use pay incentives to recruit(or retain) teachers to teach in less desir-able locations or in fields of shortage.11

Some states and localities have also experi-mented with various performance-basedpay systems, such as merit pay, the formallinking of individual teachers’ salaries toindividual teacher performance.12 Thoughthis has been tried in a number of locali-ties, these efforts have been viewed bymany to be unsuccessful and have largelybeen short-lived (Hatry, Greiner, andAshford 1994; Murnane and Cohen1986).13

More recently, states and localities haveopted for a group-based approach to per-formance pay. South Carolina and Tennes-see, for example, have developed systemswhere schools are rewarded with addi-tional funding based on the performanceof students in the school (Clotfelter andLadd 1996, Odden and Kelley 1997). Al-though they differ from one another insome respects, these systems compareschool-wide gains in student scores overtime to what econometric models wouldpredict the gains to be after controlling forstudents’ demographic characteristics.14

Still, these performance-based systems rep-

resent the minority of school systems’ paystructures.

Given that pay is determined by degree andexperience level in most school systems,there is little ability for school systems todifferentiate pay at the individual level inorder to pay for teacher attributes such asperformance, field of specialization, qual-ity of college, grades in college, academicor other awards, or standardized testscores.15 For this reason, the structure ofcompensation in teaching is quite differentthan the structure of compensation in theprivate sector as a whole. For instance, sala-ries in the private sector are thought to re-flect college quality (Brewer, Eide, andEhrenberg 1999),16 standardized test scores(Murnane, Willett, and Levy 1995), and thesupply and demand conditions for particu-lar fields or occupations (Grogger and Eide1995, Freeman 1976). Pay for performanceis also quite common in some occupations,such as sales where commissions determinea large share of total compensation.17 Bretzand Milkovich (1989) estimate that 93 to99 percent of private sector firms use sometype of pay for performance plan for sala-ried employees. There has recently been in-creased use of plans, such as profit sharingand employee stock ownership, that for-mally link pay and performance for all typesof employees.18 Studies tend to show thatthese compensation strategies do lead to

11 From 1987–88 to 1993–94, the percentage nearly doubled from 8 to 15 percent for all districts (U.S. Department of Education 1998b).

12 Economists have long theorized that connecting compensation to performance should elicit higher levels of effort and productivity(Pencavel 1977, Brown 1990).

13 Despite the fact that economic theory might predict merit pay to be a successful strategy to use in schools, many argue that this privatesector model cannot work in education because both the process used and the product produced are not well defined. For instance,teachers may use many different strategies to achieve multiple educational objectives. This may make it difficult to accurately assess theirlevel of effort or value added. Additionally, pay for performance may be more difficult to implement successfully in large organizations(Lawler 1981, Deming 1986, Rainey 1990).

14 An assessment of a similar accountability program in place in the Dallas Independent School District suggests that this type of reformcan have a relatively small, but statistically significant positive impact on student achievement (Clotfelter and Ladd 1996).

15 However, this does not necessarily mean that these characteristics are not rewarded in teacher labor markets. For instance, it is possiblethat no single school system differentiates pay based on college quality, but teacher candidates from higher quality colleges are system-atically more likely to obtain jobs in high paying school districts, implying an unequal distribution of teacher skills across school districts(Ferguson 1998).

16 As a particular example, the median starting salary for an MBA from the program ranked 1st, Harvard University, was $82,000 in 1997compared with only $64,000 for the program ranked 25th, Vanderbilt University (U.S. News and World Reports 1998).

17 Economists have long theorized that connecting compensation to performance should elicit higher levels of effort and productivity(Pencavel 1977, Brown 1990). It may also assist in attracting and retaining high quality employees (Milkovich and Wigdor 1991).

18 For a review of performance-based pay, see Paying for Productivity: A Look at the Evidence (Blinder 1990).

...the structure of

compensation in

teaching is quite

different than the

structure of

compensation in

the private sector

as a whole.


20

higher levels of productivity (Mitchell,Lewin, and Lawler 1990;Weitzman andKruse 1990).

Since the broader labor market appears tofinancially reward characteristics thatschool systems do not (at least explicitly),individuals who graduate from more selec-tive colleges, and who have degrees in fieldsin high demand likely sacrifice more finan-cially to enter or remain in teaching—animportant point to consider when compar-ing compensation in teaching with compen-sation in other fields. That is, the signifi-cant differences in pay structure betweenteacher labor markets and private sectorlabor markets imply that the incentives toenter, and costs to stay in teaching careers,may vary considerably from field-to-fieldand individual-to-individual.

Several other considerations are importantto highlight prior to comparing compensa-tion in teaching with that in other fields.First, yearly compensation in teaching isbased on a contract year that is generally180 to 200 days. Thus, to be comparableto most other occupations the yearly sal-ary would have to be scaled to a standardwork year. Because the contract year forteachers has for most part remained con-stant over time, it is not necessary to makethis adjustment in order to make judge-ments on how relative teacher compensa-tion has changed.

In addition, when comparing average sala-ries in teaching with those in other occu-pations, it is worth considering the fact thataverages are influenced by both changes inthe structure of salaries (i.e., the magnitudeof raises at each salary step), as well as thehiring decisions of school systems andchanges in the demographic compositionof teachers. Some school systems maychoose to hire senior, highly-credentialedstaff, while others choose to hire more jun-ior, less costly staff. For instance, two schooldistricts may have school systems that haveequivalent salary schedules but have verydifferent average salaries based on whereteachers fall on the salary schedule in terms

of degrees and years of teaching experience.Lankford and Wyckoff (1997) show that,in increasing salaries in the 1980s and early1990s, many school districts “backloaded”salaries—that is, they allocated a great dealof the overall resources used to increasesalaries toward increasing salaries of rela-tively experienced teachers. Depending onhow new entrants into the labor marketvalue starting compensation versus com-pensation in mid- and late-career, these in-creases may have had little impact on at-tracting able college graduates into teach-ing.

Ideally, in making compensation compari-sons we would like to be able to observethe compensation of individuals who areat similar points of their career. For in-stance, a teacher who has 10 years of post-baccalaureate experience earns $X versusan engineer with 10 years of post-bacca-laureate experience who earns $Y. In prac-tice, this type of detailed comparison atmultiple points of a career is not possiblegiven data constraints; however, startingsalaries provide a good benchmark for theattractiveness of an occupation and exist-ing data do allow for a comparison be-tween occupations in starting salaries.Thus, in the following section, we focusnot only on changes in averages both withinand outside of teaching, but also on howstarting salaries have changed over time.

Changes in TeacherCompensation

Over the past two decades, a number ofhighly publicized reports, such as NationalCommission on Excellence in Education(1983), the Carnegie Forum on Educationand the Economy (1986), and the NationalCommission on Teaching and America’sFuture (1996), included among their rec-ommendations, raising teachers salaries.The argument for this recommendation isthat, all else equal, higher salaries shouldattract more qualified individuals intoteaching. Researchers and policymakersmay differ about the need for higher sala-ries based on disagreements over how ef-

Ideally, in making

compensation

comparisons we

would like to be

able to observe the

compensation of

individuals who

are at similar

points of their

career.


21

fectively additional monies spent on teach-ers have been used to attract high-qualitypeople into teaching.19 However, it is sur-prisingly difficult to gain a full understand-ing of how teachers’ salaries have changedrelative to inflation and salaries in otheroccupations, thus part of the debate maybe shaped by the comparisons that aremade in assessing changes in salaries.

In this section of the paper, data and re-search drawn from several sources presenta comprehensive picture of how teachers’salaries and benefits have changed in thepast two decades and the implications ofthis change. In particular, comparing vari-ous measures of change to see whether theytend to present the same picture of change.Where appropriate, the shortcomings ofavailable data or the approach taken arealso noted. The first subsection begins witha discussion of teacher benefits. Unfortu-nately, due to a lack of available data, rela-tively little information is available at a dis-aggregated level on the magnitude ofchange in benefits over time; thus this sub-section deals primarily with changes in ag-gregate spending on benefits. Significantly

more data are available on changes overtime in average teachers’ salaries, which isdiscussed in the next subsection. However,as noted above, averages reflect choices thatare made by school districts as well aschanges in the demographics of the teacherlabor force; thus the final subsection fo-cuses on changes over time in teachers’starting pay.

Benefits

In most school systems, benefits constitutean important part of the overall compen-sation. Figure 1 shows the amount spenton benefits for instructional staff (repre-sented by the broken line and the right handY-axis) as well as the percentage of totalinstructional expenditures for public el-ementary and secondary education that isdevoted to employee benefits (representedby the solid line and the left hand Y-axis).The figure indicates that spending on ben-efits has increased significantly throughoutthe 1990s, from about $20.7 to $30.3 mil-lion or roughly 46 percent. However, giveninflation and the fact that this correspondsto a period when more teachers were be-

19 Teachers’ salaries and benefits represent by far the largest expenditure category for K–12 schools. Teachers’ salaries alone typically makeup over half of school district budgets (Chambers 1997) and salaries and benefits combined generally make up over 80 percent (U.S.Department of Education 1998a). See Hanushek (1986) or Ballou and Podgursky (1997) for a discussion of why these expenditures maynot have attracted high quality individuals into teaching.

Figure 1.—Dollars and percentage spent on employee benefits

SOURCE: U.S. Department of Education. 1999. Digest of Education Statistics, 1998. Washington, DC: National Center for

Education Statistics (NCES 1999–036).

1995–961994–951993–941992–931991–921990–911989–90

Year

17.6

17.8

18.0

18.2

18.4

18.6

18.8

19.0

19.2

19.4

19.6Percentage

0

5

10

15

20

25

30

35Dollars (millions)

Percentage of instructional expenditures devoted to employee benefits

Total dollars spent on benefits


22

ing hired, it is not necessarily a good indi-cation of benefits levels for individuals.20

To gain a better perspective of how actualbenefits have changed, it is useful to ex-amine data from the Schools and StaffingSurvey. These data indicate that almost allpublic school districts offer teachers retire-ment packages and medical insurance.21

Dental insurance (about 67 percent), lifeinsurance (about 71 percent), and other in-kind benefits (e.g., meals, transportation,tuition reimbursement, etc.) are less widelyoffered. Data from the 1987–88, 1990–91,and 1993–94 waves of the survey showthere has been a slight increase in the per-centage of school districts that offer ben-efits, particularly retirement and in-kindbenefits.22 (However, some of the reportedincreases may be a result of rewording ofquestions on surveys).23

Based simply on the availability of cover-age (e.g., whether there is a retirement orhealth plan), the benefits available to teach-ers appear to be roughly comparable tothose available to full-time employees instate and local governments and they ex-ceed the coverage offered to those em-ployed in small private establishments orthe private sector as whole (Bureau of La-bor Statistics 1995, 1998b) For instance,in 1993–94 only 50 percent of private sec-tor workers were covered by a retirementplan (Foster 1997).

The Schools and Staffing Surveys, whichare probably the best source of informa-tion on teacher benefits, only provide in-formation on the incidence and provisionsof selected benefits, thus no information isavailable on the generosity of benefit pack-ages. For instance, we do not know whatthe breakdown is between employee andemployer contributions for health care cov-

erage or the specifics of retirement pack-ages. The Bureau of Labor Statistics col-lects slightly more detailed data in theirEmployee Benefits Survey, however, at thispoint the survey does not include teachersas a separate category.

The fact that benefits as a percentage ofinstructional expenditures have been in-creasing does suggest that, from an insti-tutional perspective, benefits represent alarger fraction of overall compensation forteachers. However, this increase may notbe perceived as an enhanced benefit byteachers given that the costs of providingsome benefits, particularly medical insur-ance, increased substantially during thistime period. For instance, between 1981and 1997, the rate of increase in benefitsoutpaced the rate of increase of wages andsalaries by 34 percentage points for all ci-vilian workers (Bureau of Labor Statistics1998a, 1998b). Thus, the increased expen-diture may not correspond well with theactual services that teachers receive.

As a result, while the existing data sourcesappear to show that benefits for teacherscompare favorably to those in the privatesector, it is not possible to directly com-pare how they have changed over time rela-tive to the private sector. Thus, there is noindication of whether teaching has becomea more or less attractive occupation givenchanges in benefits. Fortunately, far moredetailed data exists on teachers’ salaries,which are explored below.

Average Salaries

The first step in assessing how financiallyattractive it is to be a teacher today com-pared with earlier years is to see how sala-ries have changed in real terms. This infor-mation is provided in figure 2, which pre-sents average teachers’ salaries, both un-

20 Also, these figures represent expenditures for all public school employees that fall under the heading of instructional staff, not teachersonly.

21 About 99 percent of districts offer some type of retirement plan and 96 percent of districts offer medical insurance.

22 For instance, in 1987–88, 56.4 percent of school districts reported offering benefits, but this percentage rose to 69.1 percent in 1990–91.

23 For a more detailed description of changes in the survey questions, see Schools and Staffing in the United States: A Statistical Profile,1990–91 (U.S. Department of Education 1993a).

The first step in

assessing how

financially

attractive it is to be

a teacher today

compared with

earlier years is to

see how salaries

have changed in

real terms.


23

adjusted (current dollars) and adjusted(constant dollars) for inflation. Unadjustedteachers’ salaries increased by 136 percentin the last two decades (from 1978–79 to1997–98); however, in real terms (in con-stant 1997–98 dollars), this actually rep-resents only a modest 11 percent increase.24

In fact, as the trend line for inflation ad-justed salaries indicates, the real purchas-ing power of teachers’ salaries actually de-clined between 1978–79 and 1980–81. Itrecovered and rose somewhat over the nextten years but has remained relatively flatthrough much of the 1990s.

It is important to remember that pay scalesin public schools are generally set via a sal-ary schedule and are thus greatly influencedby the hiring decisions of school systems(i.e., the credentials of the teachers hired).Additionally, productivity growth in ser-vice industries, such as education, is typi-cally slower than in other sectors of theeconomy. Thus, salaries may rise (with pro-ductivity growth) in some sectors of the

economy without causing commensurateincreases in output prices (inflation). Thismeans inflation in the prices of educationalinputs may exceed that calculated by aneconomy-wide measure, such as the con-sumer price index (CPI), which is used todeflate salaries in the graph above. For in-stance, the use of a general gross domesticproduct (GDP) deflator would tend to over-state the investment in education in termsof the quality of labor purchased.25

For these reasons, the CPI may not be themost appropriate statistic to use in adjust-ing teachers’ salaries for inflation. Severalalternative price deflators are used to as-sess the degree to which the magnitude ofchange in real teacher salary is determinedby the salary deflator that is chosen. Al-though the specifics of the various defla-tors employed here are not discussed, thenet services index (NSI), Chambers’ teachercost index (TCI), and Goldhaber’s generalwage index (GWI), each is designed (usingdifferent methodologies)to account for

24 This is average compensation for all elementary and secondary teachers (U.S. Department of Education 1999).

25 For a discussion of this, see Rothstein and Mishel (1997).


Education Statistics (NCES 98–015).

Figure 2.—Average teacher salaries

Year

13

18

23

28

33

38

43

1978–79

1979–80

1980–81

1981–82

1982–83

1983–84

1984–85

1985–86

1986–87

1987–88

1988–89

1989–90

1990–91

1991–92

1992–93

1993–94

1994–95

1995–96

1996–97

1997–98

Dollars (thousands)

0

5

10

15

20

25

30

35

40

45

Dollars (thousands)

Constant (1997–98) dollars

Current dollars


24

(some or all of) the potential shortcomingsof the CPI in adjusting for inflation in teach-ers’ salaries.26

Figure 3 shows a comparison of averageteachers’ salaries (in 1987–88 dollars) de-flated using the CPI to salaries deflated us-ing the other indices. The years 1987–88,1990–91, and 1993–94 are chosen becausethey correspond to the years for which theChambers TCI27 and Goldhaber GWI areavailable.28

The use of different price deflators does, insome cases, have a relatively significantimpact on the calculation of real teachers’salaries. For instance, the difference be-tween the NSI and the Goldhaber GWI in1993–94 is about $1,000. Still, the differ-ence are not so pronounced so as to changethe overall pattern showing a rising salaryfrom 1987–88 to 1990–91 and a falling sal-ary from 1990–91 to 1993–94. These dif-

ferences, however, do indicate that theremay be important differences in produc-tivity between education and other sectorsof the economy. This, in turn, implies it isquite important when investigating therelative attractiveness of teaching to com-pare teachers’ salaries to salaries in otheroccupations that directly compete forteachers in the labor market.

Such a comparison is drawn from data col-lected for and prepared by the Bureau ofLabor Statistics (BLS) Office of Compen-sation and Working Conditions. The ECImeasures changes over time in compensa-tion costs which include wages, salaries,and employer costs for employee benefitsin organizations of all sizes in private in-dustry and the public sector throughout theUnited States (it also allows researchers toseparately obtain data on average employercosts per hour worked for wages and sala-ries). The data is organized by “series”

Figure 3.—Average salaries using different price deflators

SOURCE: Author's calculation based on data from U.S. Department of Education. 1998. Digest of Education Statistics,

1997. Washington, DC: National Center for Education Statistics (NCES 98–015); Chambers, Jay G. 1998. Geographic Variations

in Public Schools’ Costs. Washington, DC: U.S. Department of Education, National Center for Education Statistics (NCES

Working Paper 1998–04); and Goldhaber, Daniel D. 1999. “An Alternative Measure of Inflation in Teachers’ Salaries.” In

William Fowler, Jr. (ed.), Selected Papers in School Finance 1997–99, Washington, DC: U.S. Department of Education, National

Center for Education Statistics (NCES 1999–334).

Goldhaber

Chambers

NSI

CPI

26,600

26,800

27,000

27,200

27,400

27,600

27,800

28,000

28,200

28,400

28,600

28,800

Dollars

1993–941990–911987–88 Base

Year

26 For a review of this issue and description of the various methodologies, see Rothstein and Mishel (1997), Chambers (1997), andGoldhaber (1999).

27 Chambers’ TCI is the teachers’ salary component of his cost of education index.

28 These two indices were calculated using the Schools and Staffing Survey which was only available in select years.


25

which constitute different segments of theeconomy (e.g., “all workers,” “white col-lar workers,” “service occupations,” and“union workers”). One of the series in-cluded in the survey is elementary and sec-ondary schools (ESS). This series providesan excellent opportunity to compare, overa relatively long period of time, how aver-age salaries in education have changed rela-tive to those in other sectors of the economy(for readers who are not familiar with thissurvey, more information is presented inappendix A).29

The ECI data are used to show how aver-age teachers’ salaries have changed rela-tive to salaries in sectors of the economythat compete in the labor market for teach-ers. Specifically, a comparison of the indexof wages and salaries in the ESS series towages and salaries for all civilian work-ers, all workers in private industry, allworkers in state and local government,white-collar workers in private industry,white-collar workers in state and local gov-ernment, civilian workers in service occu-

pations, and private industry workers in ser-vice occupations (occupations likely todraw from the same labor pool).

Each index is normed such that (June) 1989equals 100. Thus, it measures salaries ineach series relative to the value of salariesin that series in 1989.30 This allows for acomparison of salary growth rates in dif-ferent sectors of the economy. Figure 4shows the growth rate of the wage and sal-ary index in the ESS series to the indicesfor private white-collar workers and stateand local government white-collar work-ers.

This graph shows that growth rates in theESS series exceeded those in other series formost of the 1980s and 1990s. This con-flicts slightly with the impression from fig-ure 2 which shows real teachers’ salariesstagnating for most of the latter 1980s andearly 1990s (this pattern holds for all eightseries and for total compensation, both ofwhich are shown in appendix B).31 How-ever, it corresponds closely with data re-

Figure 4.—Growth rate in wages and salaries

SOURCE: Author's calculation based on data from Bureau of Labor Statistics. 1999. Employment Cost Trends, 1999. http:/

/stats.bls.gov/ectserie.htm.

Elementary and secondary schools

State and local government—white collar

Private—white collar

0

1

2

3

4

5

6

7

8

9

10

Percentage

199819971996199519941993199219911990198919881987198619851984198319821981

Year

29 The “elementary and secondary schools” series includes a broader class of employees than teachers alone, which means it may notcorrespond well to the data presented in figure 1 on average teachers’ salaries. As shown in appendix B (figure B-1), the inflation rate inreal teachers’ salaries as calculated using the ECI consistently exceeds the inflation rate as calculated using data from the Department ofEducation. However, the year-to-year changes in inflation track very closely.

30 For example, if in 1982 salaries in the ESS series were $25,000 and salaries in the private white-collar sector were $35,000, and in 1989salaries in the ESS series were $40,000 and $50,000 in the union sector, then the values of the indices in 1982 are 62.5 for ESS and 70.

31 This indicates that salaries in most sectors of the economy stagnated during this period.


26

ported in an American Federation of Teach-ers (1999a, 1999b) report on salary trendsthat shows throughout most of the 1980sand early 1990s the rate of increase ofteachers’ salaries exceeded that of other oc-cupations (Schneider and Nelson 1998).This is prima facie evidence that, all elseequal, until very recently (1996–98), whenthis pattern shifts such that the growth rateof salaries in most other sectors of theeconomy exceeded that in education, teach-ing increasingly became a (financially)more attractive field to be in throughoutmuch of the last two decades.

As discussed in the structure of teachercompensation section, most teachers aresubject to the uniform salary schedule. Thismeans the aging in the age distribution ofteachers can greatly increase the averagesalary in the profession without making ita more attractive profession to enter. Ad-ditionally, there is evidence that teachersalaries schedules were backloaded in the1980s—that is, the higher steps on the sal-ary schedule were increased faster (andmaybe at the expense of) the beginning

steps on the schedule (Lankford andWyckoff 1997). This would further inflatethe average salary without necessarily im-pacting starting salaries. It also means in-creases in average salary may not serve asa good indicator of how attractive it is toenter teaching. Thus, in the followingsubsection, how starting salaries in educa-tion have changed over time and how theycompare relative to starting salaries in otheroccupations is examined.

Starting Salaries

Figure 5 shows how the average minimumstarting salary in teaching has changed overtime and how average minimum startingsalaries compare to average salaries. As wasthe case with average salaries, real startingsalaries stagnated for the latter half of the1980s and throughout the 1990s.

Although there is some evidence ofbackloading of salaries in the late 1980sand early 1990s when starting salaries de-clined as a percentage of average salaries,by 1997–98 starting salaries had climbed

Figure 5.—Minimum average starting teachers' salaries


Education Statistics (NCES 1999–036); U.S. Department of Education. 1998a. Digest of Education Statistics, 1997. Washington,

DC: National Center for Education Statistics (NCES 98–015); American Federation of Teachers. 1999a. Teacher Salary

Boost is One Way to Stem Shortages of Teachers. Press Release. http://www.aft.org/press/index.html; and American

Federation of Teachers. 1999b. Graveyard: AFT 50-State Teacher Salary Survey. http://www.aft.org/research/salary/stgrave/

begin/92.htm.

Minimum salaries (current dollars)

Minimum salaries (constant [1997–98] dollars)

Minimum salary (percentage of average salary)

1986–871987–88

1988–89

1989–90

1990–91

1991–92

1992–931993–94

1994–95

1995–96

1997–98

1996–97

Year

0

5

10

15

20

25

30

Dollars (thousands)

63.0

63.5

64.0

64.5

65.0

65.5

66.0

66.5

67.0

Percentage


27

back up to almost the same percentage ofaverage starting salaries (65.3 percent) asthey were a decade earlier. On the whole,starting salaries in teaching changed littlerelative to average salaries. Thus, unlessstarting salaries in other occupations rosemarkedly relative to salaries in those oc-cupations, starting salaries in teachinglikely changed little relative to starting sala-ries in other sectors of the economy.

Starting salaries in teaching have typicallylagged far behind salaries in other occupa-tions and the data presented here do notprovide reason to believe this situation haschanged (U.S. Department of Education1993b). For example, figure 6 shows start-ing salaries in teaching compared to se-lected other occupations, and figure 7shows the ratio of starting salaries in thoseoccupations to salaries in teaching (the“salary ratio”).

As is apparent from figure 6, teacher sala-ries were less than any other occupation inevery year from 1972 to 1997. Figure 7clearly shows that during the 1970s andearly 1980s starting salaries in teaching fellbehind those in other occupations. Muchof this loss was recovered by the mid-1990s; however, relative to most occupa-tions, teachers’ starting salaries again lost

ground after 1995. It is important to notethat although much of the lost ground inrelative teachers’ salaries was made up dur-ing the 1980s, in absolute terms startingsalaries in teaching were further behindother occupations in the 1990s than in the1970s. This point is illustrated by figure 8,which shows the differential between real(in 1997 dollars) starting salaries to that ofteachers in both 1978 and 1996.

Between 1978 and 1996, the salary ratiodecreased substantially for most compari-sons, implying that starting salaries in teach-ing had gained relative to other occupations.However, this can be deceptive. For in-stance, it is striking that despite decreasesin the ratio for chemistry, mathematics orstatistics, and computer science, the abso-lute gap in salaries increased. The magni-tude of the gap is also striking, particularlyfor technical fields. Starting salaries in en-gineering, chemistry, mathematics or sta-tistics, and computer sciences all exceededteaching by at least $10,000.

Implications of Findings

The level and structure of compensation inan occupation plays an important role inattracting or dissuading individuals fromentering and exiting that occupation. The

SOURCE: Schneider, Krista and Nelson, Howard. 1998. Salary & Analysis of Salary Trends 1997. AFT Research Report. http:/

/www.aft.org/research/.

Figure 6.—Starting salaries in select occupations

19971996199519941992199019881986198419821980197819761972

Year

0

5

10

15

20

25

30

35

40

45Dollars (thousands)

Engineering

Liberal ArtsChemistry

Computer Science

Sales/Marketing

Mathematics or StatisticsEconomics/Finance

Accounting

Business Administration

Teaching


28

evidence on changes in teachers’ salariesin the third section presents a somewhatmixed picture. Data from various sourcesshows that real relative starting and aver-age salaries, teaching, over the last 25 yearshas gone through a cycle, first losingground relative to other occupations,regaining lost ground during most of the1980s and early 1990s and again losingground in the late 1990s.

Figure 7.—Ratio of starting salaries in select occupations to starting salaries in teaching



Engineering

Liberal ArtsChemistry

Computer Science

Sales/Marketing

Mathematics or StatisticsEconomics/Finance

Accounting


Ratio

19971996199519941992199019881986198419821980197819761972Year

1.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

2.0

2.1

Figure 8.—Difference between starting salaries in teaching and selected other occupations



4

6

8

10

12

14

16

18

19971978

Dollars (thousands)

Year

Computer Science

Economics/Finance

Mathematics or Statistics

Chemistry

Liberal Arts


Sales/Marketing

Accounting

Engineering

In some cases increases in relative salariesmask growing gaps between salaries inteaching and those in other occupations.32

In comparing starting salaries in differentoccupations, it is clear that, for some white-collar occupations, what a new graduatecan expect to make in teaching is substan-tially less than what that individual wouldmake in other fields.

32 The implications of this are unclear and depend on whether individuals care about relative salary standing or absolute differences insalaries.


29

This raises an important issue. As discussed,few school systems explicitly provide theflexibility to differentiate pay at the indi-vidual level to pay for teacher attributessuch as field of specialization, quality ofcollege, grades in college, academic or otherawards, or standardized test scores. How-ever, the data clearly shows that differentoccupations offer different financial re-wards, and research has shown that indi-viduals who graduate from more selectivecolleges have greater labor market oppor-tunities. By contrast, researchers andpolicymakers tend to refer to teachers andteacher salaries in generic terms. Given thedata presented above on college quality andfield of specialization, this treatment is in-appropriate.

In fact, there is evidence suggesting that therigid pay structure in teaching may ad-versely affect the number of high ability in-dividuals entering or remain in the teach-ing profession (Ballou and Podgursky 1997;Ballou 1996; Ferguson 1998; Hanushek1986, 1997; Murnane and Olsen 1990).Based on several readily observable mea-sures, teacher candidates and teachers areless skilled than individuals who enter otheroccupations. As measured by standardizedtest scores (the SAT, ACT, etc.), most col-lege students selecting education majorstend to be drawn from the lower end ofthe ability distribution (Hanushek and Pace1995, U.S. Department of Education1996a). These findings are summed up inMurnane et al. (1991, 10),

College graduates with high testscores are less likely to take jobs,employed teachers with high testscores are less likely to stay, andformer teachers with high testscores are less likely to return.33

On average, the higher the quality of anindividual’s undergraduate institution, theless likely they are to chose a teaching ca-reer (Ballou 1996). In fact, over the last 25years, there has been a rather dramatic shiftaway from top tier public and private re-search universities in the share of MA andPh.D./Ed.D. degrees in education (Turner1998). Finally, more education majors thannoneducation majors report taking reme-dial mathematics and English courses (U.S.Department of Education 1996a).34

This situation is unlikely to improve un-less there is an increase in teachers’ com-pensation. The recent slowdown in growthof both average and starting teachers’ sala-ries combined with a tight labor marketimply that teaching will be less financiallyattractive to many individuals. This is par-ticularly true for individuals with techni-cal skills given there are increasing returnsto skills in society.

To make compensation for all teacherscomparable with compensation in techni-cal fields would require increases in sala-ries that outpaced the gains at any point inrecent history. This leaves school systemswith difficult choices and challenges. Theycould procure and devote unprecedentedamounts of money toward teacher compen-sation, differentiate salaries by teacherskills, and/or risk losing technically profi-cient individuals to other occupations. Thisis not a new choice and the existing evi-dence is that most school systems haveopted for the last option.

How school systems respond to thesechoices will have important, long-lastingeffects. Indeed, the National Commissionon Teaching and America’s Future(NCTAF) has estimated that “more new

33 Mean verbal SAT scores for those intending to study education were 409 in 1993–94, compared to 438 for social sciences, 452 for artsand humanities, and 500 for physical sciences (U.S. Department of Education 1998a). Mean GRE scores for those intending graduatestudy in education were 477 in 1987–88 compared with 529 for social sciences, 541 for business and 685 for engineering. By contrast,teachers tend to have higher college grades, which may simply reflect differences in the grading scales in education courses which tendto be higher than in many other college majors.

34 Fifteen percent of education majors report taking at least one remedial mathematics course versus 12 percent of noneducation majors.The corresponding figures for remedial English are 13 and 7 percent. These figures raise concern about the quality of the nation’steacher labor force.

...there is evidence

suggesting that the

rigid pay structure

in teaching may

adversely affect the

number of high

ability individuals

entering or remain

in the teaching

profession.


30

teachers will be hired in the next decadethan in any previous decade in our history”(NCTAF 1996, 76). These demographicsprovide policymakers with both a chal-lenge and a golden opportunity to greatlyinfluence the nation’s teacher workforce forgenerations to come.


31

References

American Federation of Teachers. 1999a. Teacher Salary Boost is One Way to Stem Shortages of Teachers. Pressrelease. http://www.aft.org/press/index.html.

American Federation of Teachers. 1999b. Graveyard: AFT 50-State Teacher Salary Survey. http://www.aft.org/research/salary/stgrave/begin/92.htm.

Astin, Alexander W., Parrott, Sarah A., Korn, William S., and Sax, Linda J. 1997. The American Freshman:Thirty Year Trends. Los Angeles, CA: UCLA, Higher Education Research Institute.

Ballou, Dale. 1996. “Do Public Schools Hire the Best Applicants?” Quarterly Journal of Economics. 111(1):97–134.

Ballou, Dale and Podgursky, Michael. 1997. Teacher Pay and Teacher Quality. Kalamazoo, MI: Upjohn.

Blinder, Alan (ed.). 1990. Paying for Productivity: A Look at the Evidence. Washington, DC: BrookingsInstitution.

Bretz, Robert and Milkovich, George. 1989. Performance Appraisal in Large Organizations: Practice and Re-search Implications. Ithaca, NY: Cornell University, Center for Advanced Human Resource Studies (WorkingPaper 1989–17).

Brewer, Dominic J., Eide, Eric A., and Ehrenberg, Ronald G. 1999. “Does it Pay to Attend an Elite PrivateCollege: Cross-Cohort Evidence on the Effects of College Type on Earnings.” Journal of Human Resources.34(1): 104–123.

Brown, Charles. 1990. “Firms Choice of Method of Pay.” Industrial and Labor Relations Review. 43: 165S–182S.

Bureau of Labor Statistics. 1999. Employment Cost Trends, 1999. Washington, DC: U.S. Department of Labor.http://stats.bls.gov/ectserie.htm.

Bureau of Labor Statistics. 1998a. Employee Cost Indices, 1975–97. Washington, DC: U.S. Department ofLabor. Bulletin 2504.

Bureau of Labor Statistics. 1998b. Employee Benefits in Small Private Industry Establishments, 1996. Washing-ton, DC: U.S. Department of Labor. USDL 98–240.

Bureau of Labor Statistics. 1995. Employee Benefits in State and Local Governments, 1994. Washington, DC:U.S. Department of Labor. USDL 1995–368.

Carnegie Forum on Education and the Economy. 1986. A Nation Prepared: Teachers for the 21st Century.Washington, DC: The Task Force on Teaching as a Profession.

Chambers, Jay G. 1998. Geographic Variations in Public Schools’ Costs. Washington, DC: U.S. Department ofEducation, National Center for Education Statistics (NCES Working Paper 1998–04).


32

Chambers, Jay G. 1997. Measuring Inflation in Public School Costs. Washington, DC: U.S. Department ofEducation, National Center for Education Statistics (NCES Working Paper 1997–43).

Clotfelter, Charles T. and Ladd, Helen F. 1996. “Recognizing and Rewarding Success in Public Schools.” InHelen F. Ladd (ed.), Holding Schools Accountable: Performance-Based Reform in Education. Washington, DC:The Brookings Institution.

Deming, Edward S. 1986. Out of the Crisis. Cambridge, MA: Center for Advanced Engineering.

Ehrenberg, Ronald G. and Brewer, Dominic J. 1995. “Did Teachers’ Verbal Ability and Race Matter in the1960’s? Coleman Revisited.” Economics of Education Review. 14(1): 1–21.

Ehrenberg, Ronald G. and Brewer, Dominic J. 1994. “Do School and Teacher Characteristics Matter? Evidencefrom High School and Beyond.” Economics of Education Review. 13(1): 1–17.

Ferguson, Ronald F. 1998. “Can Schools Narrow the Black-White Test Score Gap?” In Christopher Jencks andMeredith Phillips, (eds.), The Black-White Test Score Gap. Washington, DC: The Brookings Institution.

Ferguson, Ronald F. 1991. “Paying for Public Education: New Evidence on How and Why Money Matters.”Harvard Journal on Legislation. 28(2): 465–498.

Figlio, David N. 1997. “Teachers’ salaries and Teacher Quality.” Economic Letters. 55: 267–271.

Foster, Amy C. 1997. “Employee Benefits in the United States, 1993–94.” Compensation and Working Condi-tions Online. 2(1): Spring 46–50.

Flyer, Frederick and Rosen, Sherwin. 1994. The New Economics of Teaches and Education. Location: Organi-zation. NBER Working Paper No. 4828.

Freeman, Richard B. 1976. “A Cobbweb Model of the Supply and Starting Salary of New Engineers.” Indus-trial and Labor Relations Review. 29(2): 236–248.

Goldhaber, Dan D. 1999. “An Alternative Measure of Inflation in Teachers’ Salaries.” In William Fowler, Jr.(ed.), Selected Papers in School Finance 1997–99. Washington, DC: U.S. Department of Education, NationalCenter for Education Statistics (NCES 1999–334).

Goldhaber, Dan D., Brewer, Dominic J., and Anderson, Deborah J. 1999. “A Three-Way Error ComponentsAnalysis of Educational Productivity.” Forthcoming in Education Economics.

Goldhaber, Dan D. and Brewer, Dominic J. 1998. “When Should We Reward Degrees for Teachers?” Phi DeltaKappan. 80(2) October: 134, 136–138.

Goldhaber, Dan D. and Brewer, Dominic J. 1997. “Why Don’t Schools and Teachers Seem to Matter? Assessingthe Impact of Unobservables on Educational Productivity.” Journal of Human Resources, 32(3) Summer: 505–523.

Grogger, Jeff and Eide, Eric R. 1995. “Changes in College Skills and the Rise in the College Wage Premium.”Journal of Human Resources, 30(2) Spring: 280–310.


33

Hanushek, Eric A. 1997. “Assessing the Effects of School Resources on Student Performance: An Update.”Educational Evaluation and Policy Analysis, 19(2) Summer: 141–164.

Hanushek, Eric A. 1986. “The Economics of Schooling: Production and Efficiency in Public Schools.” Journalof Economic Literature, 49(3) September: 1,141–1,177.

Hanushek, Eric A. and Rivkin, Steven G. 1997. “Understanding the 20th Century Explosion in U.S. SchoolSpending.” The Journal of Human Resources, 32 Winter: 35–68.

Hanuskek, Eric A., Kain, John F., and Rivkin, Steven G. 1998. “Do Higher Salaries Buy Better Teachers?”(paper presented at the 1999 AEA meetings in New York, NY, January 1999).

Hanushek, Eric A. and Pace, Richard R. 1995. “Who Chooses to Teach (and Why)?” Economics of EducationReview. 14(2): 101–117.

Hatry, Harry P., Greiner, J.M., and Ashford, B.G. 1994. Issues and Case Studies in Teacher Incentive Plans.Washington, DC: The Urban Institute Press.

Lankford, Hamilton and Wyckoff, James 1997. “The Changing Structure of Teacher Compensation, 1970–94.” Economics of Education Review. 16(4): 371–384.

Lawler, E.E. 1981. Pay and Organization Development. Reading, MA: Addison-Wesley.

Loeb, Susanna and Page, Marianne. 1998. Examining the Link Between Wages and Quality in the TeacherWorkforce. Davis, CA: Mimeo, Department of Economics, University of California, Davis.

Manski, Charles F. 1987, “Academic Ability, Earnings, and the Decision to Become a Teacher: Evidence fromthe National Longitudinal Study of the High School Class of 1972.” In D.A. Wise (ed.), Public Sector Payrolls.Chicago: University of Chicago Press. 291–316.

Milkovich, George and Wigdor, Alexandra (eds.). 1991. Pay for Performance: Evaluating Performance Ap-praisal and Merit Pay. Washington, DC: National Academy Press.

Mitchell, Daniel J.B., Lewin, David, and Lawler, Edward E. 1990. “Alternative Pay Systems, Firm Performance,and Productivity.” In Alan Blinder (ed.), Paying for Productivity: A Look at the Evidence. Washington, DC:The Brookings Institution.

Monk, David and Rice, Jennifer King. 1994. “Multi-Level Teacher Resource Effects on Pupil Performance inSecondary Mathematics and Science: The Role of Teacher Subject Matter Preparation.” In Ronald G. Ehrenberg(ed.), Contemporary Policy Issues: Choices and Consequences in Education. Ithaca, NY: ILR Press.

Murnane, Richard J. 1996. “Staffing the Nation’s Schools with Skilled Teachers.” In Eric A. Hanushek andDale W. Jorgenson (eds.), Improving America’s Schools: The Role of Incentives. Washington, DC: NationalAcademy Press.

Murnane, Richard J. and Cohen, David, K. 1986. “Merit Pay and the Evaluation Problem: Why Most MeritPay Plans Fail and a Few Survive.” Harvard Educational Review. 56(1): 1–17.


34

Murnane, Richard J. and Olsen, R.J. 1990. “The Effects of Salaries and Opportunity Costs on Length of Stay inTeaching: Evidence form North Carolina.” Journal of Human Resources. 25(1): 106–124.

Murnane, Richard J., Singer, J.D., Willett, J.B., Kemple, J.J., and Olson, R.J. 1991. Who Will Teach? Policiesthat Matter. Cambridge, MA: Harvard University Press.

Murnane, Richard J., Willett, John B., and Levy, Frank. 1995. “The Growing Importance of Cognitive Skills inWage Determination.” Review of Economics and Statistics. 55(2): 251–266.

National Commission on Excellence in Education. 1983. A Nation At Risk: The Imperative for EducationalReform. Washington, DC: U.S. Department of Education.

National Commission on Teaching and America’s Future. 1996. What Matters Most: Teaching for America’sFuture. New York.

Odden, Alan and Kelley, C. 1997. Paying Teachers for What They Know and Do: New and Smarter Compen-sation Strategies to Improve Schools. Thousand Oaks, California: Corwin Press, Inc.

Pencavel, John. 1977. “Work, Effort, On-the-Job Screening, and Alternative Methods of Remuneration.” Re-search in Labor Economics. 1: 225–258.

Rainey, Hal G. 1990. “Public Management: Recent Developments and Current Prospects.” In Naomi Lynn andAaron Widavsky (eds.), Public Administration: The State of the Discipline. Chatham, NJ: Chatham House.

Rivkin, Steven G., Hanushek, Eric A., and Kain, John F. 1998. Teachers, Schools and Academic Achievement.Location: Organization. NBER Working Paper No. 6691.

Rothstein, Richard. 1995. Where’s the Money Gone? Changes in the Level and Composition of EducationSpending. Washington, DC: The Economic Policy Institute.

Rothstein, Richard and Mishel, Lawrence. 1997. “Alternative Options for Deflating Education ExpendituresOver Time.” In William J. Fowler (ed.), Developments in School Finance, 1996. Washington, DC: U.S. Depart-ment of Education, National Center for Education Statistics (NCES 97–535).

Schneider, Krista and Nelson, Howard. 1998. Salary & Analysis of Salary Trends 1997. AFT Research Report.http://www.aft.org/research/.

Strauss, Robert P. and Sawyer, Elizabeth A. 1986. “Some New Evidence on Teacher and Student Competen-cies.” Economics of Education Review. 5(1): 41–48.

Strauss, Robert P., Bowes, Lori L., Marks, Mindy S., and Plesko, Mark R. 1998. “Who Should Teach in OurPublic Schools? Implications of Pennsylvania’s Teacher Preparation and Selection Experience.” (paper presentedat the American Education Finance Association Conference, Mobile, AL, March 1998).

Turner, Sarah E. 1998. “The Training of Teachers: The Changing Degree Output in the Area of Education.”(paper presented at 1998 Association of Public Policy and Management Meetings, New York, NY).

U.S. Bureau of the Census. 1999. Statistical Abstract of the United States: 1998. Washington, DC.


35

U.S. Department of Education. 1999. Digest of Education Statistics, 1998. Washington, DC: National Centerfor Education Statistics (NCES 1999–036).

U.S. Department of Education. 1998a. Digest of Education Statistics, 1997. Washington, DC: National Centerfor Education Statistics (NCES 98–015).

U.S. Department of Education. 1998b. Public School Districts in the United States: A Statistical Profile, 1987–88 to 1993–94. Washington, DC: National Center for Education Statistics (NCES 98–203).

U.S. Department of Education. 1997. The Condition of Education, 1997. Washington, DC: National Centerfor Education Statistics (NCES 97–338).

U.S. Department of Education. 1996a. Out of the Lecture Hall and Into the Classroom: 1992–93 CollegeGraduates and Elementary/Secondary School Teaching. Washington, DC: National Center for Education Sta-tistics (NCES 96–899).

U.S. Department of Education. 1996b. The Patterns of Teacher Compensation. Washington, DC: NationalCenter for Education Statistics (NCES 95–829).

U.S. Department of Education. 1993a. Schools and Staffing in the United States: A Statistical Profile, 1990-91.Washington, DC: National Center for Education Statistics (NCES 93–146).

U.S. Department of Education. 1993b. Teachers’ Salaries—Are They Competitive? Washington, DC: NationalCenter for Education Statistics (NCES Issue Brief 93–450).

U.S. News and World Reports. 1998. “Graduate School Rankings.” http://www.usnews.com/usnews/edu/be-yond/bcrank.htm

Weitzman, Martin L. and Kruse, Douglas L. 1990. “Profit Sharing an Productivity.” In Alan Blinder (ed.),Paying for Productivity: A Look at the Evidence. Washington, DC: The Brookings Institution.

Wright, Paul S., Horn, Sandra P., and Sanders, William L. 1997. “Teacher and Classroom Context Effects onStudent Achievement: Implications for Teacher Evaluation.” Journal of Personnel Evaluation in Education.(11): 57–67.


36


37

Appendix A: Data

The employment cost index (ECI), whichis prepared by the Bureau of Labor Statis-tics (BLS) Office of Compensation andWorking Conditions, measures changesover time in compensation costs which in-clude wages, salaries, and employer costsfor employee benefits. The survey coversorganizations of all sizes in private indus-try (except farms and households) and thepublic sector (except the federal govern-ment) throughout the United States. Thedata reflects average employer costs perhour worked for wages, salaries, and spe-cific benefits for both full- and part-timeemployees. Measures of wages and salarycosts can be obtained separately from over-all compensation costs. Additionally, theemployee benefits survey (EBS), which isalso prepared by BLS, provides annual in-formation on the incidence and supply ofselected benefits provided by employers totheir employees. Data collected from ap-proximately 6,000 private, state, and localestablishments reflects the percentage ofemployees who participate in certain ben-efit programs, or as an average of the ben-efits provided relative to other indicators.Data sources include paid holidays, short-term disability, severance pay, and child-care among many others.1

The ECI utilizes a series format and allowsseries data to be differentiated by four char-acteristics: component, ownership, group,and seasonality. Where, component indi-cates the form of compensation (total com-pensation, wages and salaries, or benefits),group indicates the industry, occupation,region, and/or union status within which

the compensation is provided. As such, eachseries group is classified industry, occupa-tion, region, and/or union. Ownership in-dicates the employer classification (private,civil, or state and local), and seasonality con-trols for possible variation due to intra-yearvariation. The combination of the charac-teristic listed above (component, group,ownership, and seasonality) produces a nine-character reference code.

Here, we use three combinations of eightdata for both total compensation and wagesand salaries component groups (series datawas reported from the second quarter only,and was not adjusted for seasonality). Theyare as follows:

� All workers—civilian (ECU20001I)

� All workers—private industry(ECU20002I)

� All workers—state and local government(ECU20003I)

� White-collar—private industry(ECU21102I)

� White—collar-state and local government(ECU21103I)

� Service occupations—civil (ECU21301I)

� Service occupations—private industry(ECU21302I)

� Elementary and secondary schools—state and local government(ECU22823I)

1 Efforts are currently underway to integrate different compensation measures into a single statistical program titled the National Compen-sation Survey (Bureau of Labor Statistics 1998a).


38


39

Appendix B: Additional GraphsFigure B-1.—Percentage of change in salaries: U.S. Department of Education compared to U.S.

Department of Labor, Bureau of Labor Statistics

SOURCE: Author's calculations based on data from the U.S. Department of Education. 1999. Digest of Education Statistics,

1998. Washington, DC: National Center for Education Statistics (NCES 1999–036); and Bureau of Labor Statistics. 1999.

Employment Cost Trends, 1999. Washington, DC: U.S. Department of Labor. http://stats.bls.gov/ectserie.htm.

Figure B-2.—Growth rate in wages and salaries: All eight series

SOURCE: Author's calculations basd on data from the Bureau of Labor Statistics. 1999. Employment Cost Trends, 1999.

Washington, DC: U.S. Department of Labor. http://stats.bls.gov/ectserie.htm.

Percentage of change based on Department of Education data

Percentage of change based on employment cost index (ECI) data

-2

0

2

4

6

8

10

12

14

Percentage19

82–8

3

1997

–98

1996

–97

1995

–96

1994

–95

1993

–94

1992

–93

1991

–92

1990

–91

1989

–90

1988

–89

1983

–84

1984

–85

1985

–86

1986

–87

1987

–88

Year

Civilian—all

Private industry—service

State and local government—white collarPrivate—white collar


Civilian—service

State and local government—allPrivate industry—all

0

1

2

3

4

5

6

7

8

9

10Percentage

19981997199619951994199319921991199019891988198719861985198419831982

Year


40

Figure B-3.—Growth rate in total compensation: All eight series

SOURCE: Author's calculations basd on data from the Bureau of Labor Statistics. 1999. Employment Cost Trends, 1999.

Washington, DC: U.S. Department of Labor. http://stats.bls.gov/ectserie.htm.

181716151413121110987654321

Year

0

1

2

3

4

5

6

7

8

9

10

Percentage

Civilian—all

Private industry—service

State and local government—white collarPrivate—white collar


Civilian—service

State and local government—allPrivate industry—all

A Primer for Making Cost Adjustments in Education: An Overview

41

SelectedPapers in


A Primer for Making Cost Adjustmentsin Education: An Overview


42


43


William J. Fowler, Jr.National Center for Education Statistics

David H. MonkPennsylvania State University

About the Authors

William J. Fowler, Jr.

William J. Fowler, Jr. is the program of-ficer of the Education Finance Program inthe Elementary and Secondary Educationand Library Studies Division at the Na-tional Center for Education Statistics(NCES), U.S. Department of Education. Hespecializes in elementary and secondaryeducation finance and education produc-tivity research. He is currently involvedwith devising a fast-response survey forschool business officials to obtain an un-derstanding of how energy costs are affect-ing school district fiscal health, and report-ing the results in user-friendly language. Heis also engaged in research regarding theequity and adequacy of resources for Kin-dergartners. His great passion is designingInternet tools for the NCES education fi-nance web site at http://nces.ed.gov/edfin,as well as the graphic display of quantita-tive data.

Dr. Fowler has worked for NCES since1987, before which he served as a supervi-sor of school finance research for the NewJersey Department of Education. He hastaught school finance at Bucknell Univer-sity and the University of Illinois, andserved as a senior research associate for theCentral Education Midwestern RegionalEducational Laboratory (CEMREL) in

Chicago and for the New York Departmentof Education. He received his doctorate ineducation from Columbia University in1977.

Dr. Fowler received the Outstanding Ser-vice Award of the American Education Fi-nance Association (AEFA) in 1997, havingserved on its Board of Directors during the1992–95 term, and has been re-elected forthe 2001–04 term. He serves on the edito-rial board of the Journal of Education Fi-nance and the Educational Evaluation andPolicy Analysis Journal of the AmericanEducation Research Association. He for-merly served on the Board of Leaders ofthe Council for Excellence in Government,and was a 1997–98 Senior Fellow. He wasa member of the Governmental Account-ing Standards Board (GASB) AdvisoryCommittee charged with developing a UserGuide for Public School District FinancialStatements.

David H. Monk

David H. Monk is a professor of educa-tional administration and the dean of theCollege of Education at Pennsylvania StateUniversity. He earned his Ph.D. in 1979 atthe University of Chicago and was a mem-


44

ber of the Cornell University faculty for 20years prior to becoming dean at Penn State.He has also taught in a visiting capacity atthe University of Rochester and the Uni-versity of Burgundy in Dijon, France.

Dr. Monk is the author of Educational Fi-nance: An Economic Approach and Rais-ing Money for Education: A Guide to theProperty Tax (with Brian O. Brent) as wellas numerous articles in scholarly journals.

He serves on the editorial boards of TheEconomics of Education Review, The Jour-nal of Education Finance, EducationalPolicy, and the Journal of Research in Ru-ral Education. He consults widely on mat-ters related to educational productivity andthe organizational structuring of schoolsand school districts and is a past presidentof the American Education Finance Asso-ciation.


45

Introduction

Most people intuitively recognize geo-graphic differences in costs and in measur-ing inflation. The U.S. Bureau of LaborStatistics (BLS) created a “market basket”from expenditure information provided byfamilies and individuals on what they ac-tually bought for the Consumer Price In-dex (CPI). In 1993, 1994, and 1995, theBLS collected data by using a nationalsample of over 30,000 families who pro-vided detailed spending habit information.These data enabled BLS to construct theCPI market basket of goods and servicesand to assign each item in the market bas-ket a weight, or importance, based on to-tal family expenditures. Examples of someof the more than 200 items included in thefinal CPI market basket appear in table 1(Williams 1996).

Efforts to compare the costs of exactly thesame things in different geographic regionsinvolve comparisons of the same marketbasket of goods in two geographic areas.The difference in the prices of the samemarket basket of goods is designed to re-veal the differences in the geographic costof living. But there are immediately detect-

able difficulties with this market basket.The average market basket may not repre-sent the choices of a person in a particulargeographic area and the composition of themarket basket changes over time, as doesthe relative weight of any component.

Where, for example, are personal comput-ers, something many households are cur-rently purchasing, but did not between1982 and 1984? In addition, corporatedownsizing may have caused families tochange their spending behavior in 1995 andnot engage in as many, or as expensive,apparel and entertainment purchases andrestaurant meals as they did between 1982and 1984. Also, assessing differences in thequality of the items included in the marketbasket is difficult. For example, a 1984automobile and a 1999 automobile havesubstantially different features, even for thesame “base” price. How does one adjustthe price of the 1984 automobile for airbags or anti-lock brake systems, which didnot exist in 1984?

The BLS added an improvement to the CPIin 1987 to recognize quality adjustmentsof used car prices (Greenlees and Mason


William J. Fowler, Jr.National Center for Education Statistics



46

1996). Since January 1999, a geometricmean formula has been used to calculatemost basic indexes within the CPI; in otherwords, the prices within most item catego-ries (e.g., apples) are averaged using a geo-metric mean formula. This improvementmoves the CPI somewhat closer to a cost-of-living measure, as the geometric meanformula allows for a modest amount ofconsumer substitution as relative priceswithin item categories change.

When contemplating costs, people wish to:

1. Understand the difference in costs fromone geographic area to another (costof living).

2. Understand how costs have changedover time (inflation).

3. Recognize changes in the quality andquantity of what is being purchased.

To discern these aspects of costs, mostpeople desire an index in which to com-pare one location or time to another. TheCPI uses an index, for example, 112, whichis interpreted as meaning that a 12 per-cent increase in price has taken place be-tween the base time period [index = 100]and the year in which the index is reportedas 112. An index of 80 would be inter-preted as a 20 percent decrease in prices.Usually, the CPI base is recalculated everydecade or so.

In A Primer for Making Cost Adjustmentsin Education, the authors attempt to ex-plain the differences between educationalcosts and expenditures, explain the differ-ences in the “unit price” of teachers anddifferences over time in the level of infla-tion, examine existing indices that can beused to make judgments for these differ-ences in costs, and outline a future plan ofaction to derive a precise, stable, and accu-rate index for school administrators andpolicymakers to use. This overview sum-marizes that publication and conveys thecomplexity of what most people intuitivelyknow: there are differences in costs in dif-fering geographic locations and in measur-ing inflation. These differences are difficultenough to measure in price indices, givenitem substitution and changes in item qual-ity. However, measuring cost differences ineducation is even more difficult, since mostof the costs are in personnel, rather than insupplies.

The Difference Between Costand Expenditure

The cost of education can be defined as theminimum of what must be given up to ac-complish some result. “Expenditure” is dif-ferent from “cost” in that expenditures arenot tied to results or outcomes and can ex-ceed the minimum of what must be givenup.

Costs can be organized according to an al-location hierarchy where the lowest level

Table 1. —Items included in the CPI market basket

Item category Examples

Food and beverages Cookies; cereals; cheese; coffee; chicken; beer and ale

Housing Rent; homeowner’s costs; fuel oil; housekeeping supplies; local phone service

Apparel Men’s shirts; women’s dresses; jewelry

Transportation Airline fares; new and used cars; gasoline; auto insurance

Medical care Prescription drugs; eye care; physicians’ services; hospital rooms

Entertainment Newspapers; toys; musical instruments; admissions

Other goods and services Haircuts; college tuition; bank fees

SOURCE: U.S. Department of Commerce, Bureau of Labor Statistics. December 1996. Monthly Labor Review, Appendix

1 “Item structure,” 1987 and 1998.


47

is the unit cost of various inputs like teach-ers’ time, space, and supplies. At the nextlevel, there are costs that occur as the indi-vidual inputs are combined to form edu-cational services within classrooms andschools. Finally, at the uppermost level arethe actual outcomes of schooling wherecosts arise because of the differentiationand special needs of the students being edu-cated. Resource allocation decisions aremade at each of these levels, and it is use-ful to keep them distinct because this canallow us to determine the relative magni-tude of each source of cost.

Cost

Cost Adjustments in Education

Currently, per pupil expenditures or teachersalaries are commonly reported as nomi-nal state or school district averages, with-out correction for differences in the geo-graphic cost of living (U.S. Department ofEducation 1999). There is good empiricalevidence that geographic cost differentialsexist, however. For example, Barro (1994)states:

The fact that Florida spends 36 per-cent more than Arkansas to pro-vide virtually the same staff-to-pu-pil ratio is largely explained byFlorida’s 28 percent higher instruc-tional staff salaries (p. 7).

Most of the costs of providing public edu-cation are personnel costs, such as provid-ing employees’ salaries and fringe benefits.1

Salaries average about 65 percent of totalcurrent expenditures and employee benefitsabout another 16 percent, so that these twocategories alone are responsible for over80 percent of a school district’s expendi-tures (Fowler 1993). Purchased profes-sional services, which in part acquire theservices of professionals,2 accounts for

more personnel expenditures, as does pur-chased property services3 student transpor-tation. Supplies are truly minor in such anenterprise. While some may wish to debatethe attributes of one brand of personal com-puter diskettes over another, most personswill generally concede that they are inter-changeable.

The personnel that staff school districts,however, are certainly not interchangeable,and have vastly different attributes, even ifone compares them on such uniform char-acteristics as educational attainment andoccupational experience. These differencesmake comparing geographic differences inthe price of personnel difficult, as one mightmistakenly measure differences in the jobsthey perform or in their personal charac-teristics, such as the nature of the under-graduate institution they attended. Imag-ine, for a moment, that one school districtis located in a suburban college town, whileanother is located in a rural area. Bothspend the same per pupil, but the schooldistrict with the college offers post retire-ment positions to college faculty to teachsecondary courses and to work in adminis-trative and support services. Assuming suchretired staffs are still capable, the staffs areof vastly different quality, despite compa-rable degree status, teaching experience,and expenditures.

These quality differences in education makegeographic cost differences difficult to mea-sure. School districts can choose to employbetter-educated, more experienced staff, orto reduce class size, or to hire more spe-cialized staff, all of which are more expen-sive staff choices. They may wish to main-tain small school systems, which may bemore expensive to operate, or they maychoose to hire expensive administrators. Inshort, while school districts must adhereto numerous rules and regulations fromfederal as well as state sources, they retain

...quality differences

in education make

geographic cost

differences difficult

to measure.

1 Benefits may include retirement, Social Security contributions, medical and group life insurance, unemployment, tuition reimbursement,workman’s compensation, accrued sick leave, and professional dues and fees.

2 Examples include architects, engineers, auditors, dentists, medical doctors, lawyers, consultants, computer programmers, psycholo-gists, social workers, and accountants.

3 Examples include utility and cleaning services, snow plowing, custodial services, lawn care, and repair and maintenance.


48

a significant amount of discretion overspending, particularly spending that goesbeyond what mandates require. Just asschool districts choose to trade cost andclass size, people trade salary and benefitsfor amenities. As Chambers and Fowlerput it,

The intuitive notion underlying[the hedonic wage model] is thatindividuals care both about thequality of their work environmentas well as the monetary rewardsassociated with particular employ-ment alternatives, and that theywill seek to attain the greatest pos-sible personal satisfaction by se-lecting a job with the appropriatecombination of monetary andnonmonetary rewards. (Chambersand Fowler 1995, xv).

A cost-of-education index, therefore, mustsimultaneously take into account thosediscretionary factors that a school districtmight manipulate, such as quality andquantity of staff, and thosenondiscretionary factors that the schooldistrict cannot control, such as the cost ofliving, the competitiveness of the labormarket, and amenities, such as climate,absence of crime, and geographic location(such as proximity to water). The result-ing index might be used to determine thecost to school districts, in different geo-graphic locations, to acquire and retainsimilar qualities and quantities of staff.However, such an index does not describewhat the CPI does, that is, it does notmeasure the change over time in the pricespaid by school districts.

Geographically Based CostAdjustments

The purpose of a geographically basedteacher price index is to determine the rela-tive cost of engaging the services of com-parable teachers. Some of the necessarycomponents include: teacher characteris-tics (level of experience, training, minor-ity status, gender), cost-of-living adjust-

ments, regional amenities, employmentamenities, nonteaching wages and employ-ment opportunities in the region, union andcollective bargaining, and demand forteacher quality. Several scholars have at-tempted to define a geographically basedindex. The Teacher Attribute Model is theresult of Stephen Barro’s (1994) approach.Barro did not strive to include all of thecomponents outlined above in order to keepthe number of assumptions based on incom-plete data low. His estimate focuses on in-terstate comparisons and estimates whateach state’s average teacher’s salary wouldbe if the state employed teachers with thesame average experience and training as thatfound in the nation as a whole.

McMahon and Chang (1996) characterizedthe “market-basket” approach. This ap-proach does not focus on school personnelbut rather on costs that are outside of theschool’s control such as wages in other sec-tors of the economy and geographicallybased differences in the cost of living. Onereason for this focus is to prevent a feed-back loop rewarding schools that increasedsalaries. The basic factors in this model arethe value of housing, per capita income, thepercent change in population for the pre-ceding decade and variables representingregions of the country. It can predict cost-of-living indices at several levels of aggre-gation.

The “hedonic” model (Chambers 1998) isa more ambitious approach that deals ex-plicitly with each of the influences listedabove. The model is called hedonic becauseit is sensitive to whatever it is that teachersfind attractive or repelling about a givencareer opportunity. The Teacher Cost Index(Chambers and Fowler 1995) is an exampleof this approach. Using Schools and Staff-ing Survey (SASS) data, it includes teachercharacteristics (ethnicity, gender, education,and experience), working conditions (classsize), and salary information. Other datasources were used to assess the regionalamenities. Cost influences that the schoolhas control over were statistically controlledwhile other influences were allowed to vary.

The purpose of a

geographically based

teacher price index

is to determine the

relative cost of

engaging the services

of comparable

teachers.


49

The Geographic Cost-of-Education index(Chambers 1998) is a more recent appli-cation of this approach. In this model theindex was broadened to include other typesof inputs (school administrators,noncertified school personnel,nonpersonnel) and widened the range ofdata sources. Both approaches run the riskof relying too much on potentially ques-tionable data sources and assumptions.

The production function models are per-haps the most ambitious by focusing onthe costs associated with actually realizinggains in educational performance. Unfor-tunately there is a lack of adequate dataand complete theoretical specification forthis model to have widespread use in prac-tice. However, in recent years this modelhas been applied to several states. For anexample, see the application to New York(Duncombe, Ruggiero, and Yinger 1996).There also have been applications to Wis-consin and Texas.

A comparison of the three main models(Barro, McMahon, and Chang; and Cham-bers and Fowler) demonstrates that theindices are highly correlated at over .70.Also, the more adjustments are made, themore the degree of variation drops. Despitethe high correspondence between these in-dices, there are certain regions where thereis disagreement between the indices. Acomparison between the hedonic modeland the cost-of-living model may indicatethat this discrepancy is due to the region’sattractiveness (such as San Francisco) orunattractiveness (such as nonmetropolitanConnecticut) to most teachers.

Cost Adjustments Over Time

Adjusting for regional cost-of-living differ-ences is only one of the challenges to pro-ducing a cost-of-education index. The othermajor challenge involves adjusting for cost-of-living differences over time. Differentdeflators can lead researchers to differentconclusions.

The most common way of measuring in-flation is the method used by the Consumer

Price Index (CPI) where the cost of com-monly purchased items is tracked over time.The School Price Index is one example ofthis method that uses the urban componentof the CPI, the CPI-U. Unfortunately, thisindex can only be used at the national level.There are many problems with applying theCPI approach to education, especially thechange of relevant products over time (itemsubstitution) and the uneven growth of in-flation for different occupational areas.Education is one of those occupations thathave been strongly influenced by changesin technology. This makes it difficult totrack inflation since the supplies boughttoday (such as the computer or VCR) arenot really comparable to the supplies of afew decades ago (such as the typewriter orprojector). The second problem is that someareas have seen strong inflation (such asmedicine) while other areas have not.Rothstein and Mishel (1997) argue that dueto factors such the increase in quality dueto smaller teacher/student ratios have madeinflation greater for education. Their solu-tion is to use the Net Services Index (NSI),which measures inflation by focusing onlabor-intensive components of the CPI simi-lar to education. However, they acknowl-edge that while the NSI is an improvement,it is still an underestimate.

A second approach, the Inflationary Cost-of-Education Index (ICEI) modifies thehedonic TCI to include school administra-tors and noncertified staff. However, givendata limitations this only provides a 6-yearinflation index during the years SASS wasadministered.

The Employment Cost Index (ECI) alsoavoids the market-basket approach by mea-suring the rate of change in employee com-pensation, which includes wages, salariesand employer’s costs for employee’s ben-efits. It covers all occupations with the ex-ception of federal government workers, andis used extensively by the Federal ReserveBoard as a measure of inflation. It has aneducation subscale and has separate dataon salaries as well as fringe benefits. Of allof the indices, this one is the most attrac-

Adjusting for

regional cost-of-

living differences is

only one of the

challenges to

producing a cost-of-

education index.


50

tive because it avoids the pitfalls of itemsubstitution found in the market-basketapproach and has a large time frame (1981to 1996) available.

Using Geographic and InflationDeflators

Both geographic and inflation cost adjust-ments suffer from many flaws. Overallthere is correspondence between differentgeographic indices, however for a particu-lar area the results can be dramatically dif-ferent. Given the political nature of theseadjustments, such discrepancies can be asproblematic as they are informative. Whilethe addition of more adjustments leads toa reduction of variability and arguablygreater accuracy, the policymakers’ reluc-tance to use adjustments is understandable.

Expenditure

Education Expenditures Over Time

Per pupil expenditures from 1970 to thepresent are sometimes reported in bothcurrent and “constant” dollars (figure 1).

Inflation has been removed from these ex-penditures by using the CPI. Thus, the re-ported 1970–71 per pupil expenditure of$3,774 is reported in constant 1998–99dollars as $6,915. Unfortunately, the CPIis not specifically designed to measurechanges in education costs between timeperiods, that is, the market basket does notinclude public education costs (or taxes)(although it does include private schooltuition). In addition, some argue that theCPI consistently overestimates inflation,which will make the 1970–71 per pupilexpenditure higher than it should be; itmakes us think that the investment in edu-cation has been greater than, in fact, it was(U.S. Senate 1996 [better known as theBoskin Commission]).

What would be ideal when wishing to re-port education expenditures over timewould be a cost-of-education index thatwas computed each year (or every severalyears), that both held constant the aver-age school district discretionary costs,while measuring those costs that a schooldistrict cannot influence, including geo-graphic amenities. Such an “education in-flation index” would more accurately por-tray increases in education spending.

Figure 1.—Current per pupil expenditure in average daily attendance in publicelementary and secondary schools: 1998–99

SOURCE: U.S. Department of Education, National Center for Education Statistics, Statistics of State School Systems, Revenues

and Ex penditures for Public Elementary and Secondary Education; and Common Core of Data surveys, unpublished

tabulations.

0

1,000

2,000

3,000

4,000

5,000

6,000

$7,000

1998–991990–911985–861980–811975–761970–71

Constant 1998–99 dollars

Current dollars

Per pupil expenditures


51

Lessons to Learn and Directionsfor Future Work

There are two primary goals for the futureof geographic cost adjustments: improvethe index of cost variations as well as edu-cate the public and policymakers about anyprogress that is made. The basic challengesare: make the indices generalizable acrossdifferent levels (local, state, and region),separate and distinguish influences that arecontrollable by the school, be careful ofdouble counting when adding new adjust-ments, and address any political consider-ations.

Advice for next steps:

1. Keep the indices as simple and under-standable as possible.

2. Strive to reach consensus about howambitious you wish to be with respectto cost adjustments in full knowledgeof the flaws that remain in the avail-able tools.

3. Keep in mind that not all adjustmentsare beneficial to all parties. Be particu-larly wary of flawed adjustments thatbenefit one set of political interests overothers.

4. Provide for gradual phase-ins. Consider“quasi-leveling up” strategies and takeadvantage of inflation.

5. Place primary emphasis on supportingthe further improvement of the avail-able indices.

A more sophisticated index will allowpolicymakers to more accurately identifywhat costs are the results of regional dif-ferences and what changes in costs overtime are the result of different decisions andfactors. This will allow a more efficientallocation of educational resources. Boththe public and policymakers need to be in-formed of progress made in this area sothe index can be better utilized and a con-sensus can be reached on the appropriateapproach.

Conclusions

Our first conclusion is that the educationresearch community has not paid sufficientattention to both geographic and inflation-ary differences in the costs of education.In most cases, geographic cost adjustmentshave not been applied when assessing, forexample, intra-state fiscal equity. Thecourts, plaintiffs, and defendants havetended to use nominal per pupil currentexpenditures in their arguments. However,there is ample evidence that geographic costdifferences are something those contem-plating per-pupil expenditure equity shouldremove from their considerations. Gener-ally, the use of geographic cost adjustmentsreduces most measures of disparity. Al-though the equity measures show less dis-parity after cost-adjustment than before,substantial variations remain. However, forthose school districts that are acquiringhigher-quality staff, or greater numbers ofstaff (reducing pupil/teacher ratios), thecorrection of their nominal expenditureswill cause their expenditures to be evengreater than before. The most common useof geographic cost adjustments has beento give school districts in high cost-of-liv-ing areas higher state aid. However, thiscommon usage should be reconsidered,since such aid may be disequalizing, thatis, it may aid wealthy school districts tothe detriment of the poor. In addition, it isnot, we would argue, the cost of living forwhich we wish to compensate these schooldistricts. Rather, we would wish school dis-tricts be compensated for the acquisitionand retention of comparable staff, wher-ever they reside. This is why we feel moreconceptually comfortable with the hedonicrather than the market basket approach.Some school districts in tony locations witha cachet and superb facilities and a studentbody with panache may acquire and retainvery talented teachers for much less thantheir less fortunate neighbors, who can onlyattract such a staff by paying a large pre-mium. To date, educational researchershave not emphasized these differences, inpart because a suitable methodology forestimating these effects has been unavail-

The most common

use of geographic

cost adjustments has

been to give school

districts in high cost-

of-living areas

higher state aid.


52

able. The good news is that indices of thiskind are becoming available. The not sogood news is that the available indices re-main flawed because they fail to distinguishperfectly between expenditures and bonafide costs and may introduce perverse in-centive effects that could increase spend-ing on education with little resulting gain.

Our second conclusion is that existing costadjustments are frail reeds, indeed. Despitehis precision and intricate methodologies,Chambers arrives at very different geo-graphic cost adjustments for 1990–91 and1993–94 for Chicago, Philadelphia, andDetroit. These differences are pale in com-parison to differences between researchersand methodologies. What would be desir-able would be an emerging consensus aboutthe appropriateness of a given technique,and general unanimity regarding its appli-cation, at least in adjusting nominal (ac-tual) revenues or expenditures. Instead, wesee researchers still vociferously debatingthe merits of their own work, and the de-fects of the approaches of their similarlysituated brethren. Until the academic com-munity agrees in the robustness of any costadjustment, the future use of any adjust-ment seems unsustainable. If the cost ad-justments are not viewed as hardy, com-monplace and utilitarian tools, then therewill no longer continue to be an investmenton the part of the research community toattain them.

We were also unprepared for the soberingdiscovery that these worthwhile and desir-able adjustments would provoke such ran-cor. The simple use of one geographic costadjustment versus another was sufficient

for one researcher to suggest that state-aidsystems that employ such an adjustment“...encourages inefficiency and invites di-saster (McMahon 1996, 95).” Anotherhighly regarded economist interprets ananalysis of measurement issues by Misheland Rothstein on how to include the ef-fects of inflation in measuring schoolspending as providing perhaps the mostpersuasive case for a productivity collapse[in education] (Hanushek 1997, 185). Weencourage all parties to engage in spiriteddebates that are grounded in the facts athand.

Our third and final conclusion is that moreeffort needs to be devoted toward build-ing consensus in the methodologies that canbe used as geographic cost adjustments andas deflators. There is a great need in theeducation finance research community forthese mechanisms in order to better under-stand education spending in real terms. Weeven would go so far as to suggest that it isimproper to analyze education spendingwithout correction for differences in geo-graphic costs, or differences in costs overtime without correction for the effects ofinflation. However, we also find it im-proper to analyze “adjusted” figures wheredetails surrounding the nature of the ad-justments are inaccessible to the consumer.Situations like these cry out for the use ofsensitivity analyses so that analysts, policymakers, consumers, and taxpayers alikecan have an understanding of how sensi-tive the results of the analyses are to theuse of one rather than another of the pos-sible cost adjustment techniques.


53

References

Barro, Stephen M. 1994. Cost-of-Education Differentials across the States. Washington, DC: U.S. Departmentof Education, National Center for Education Statistics. (NCES Working Paper No. 94–05).

Chambers, Jay G. 1998. Geographic Variations in the Prices of Public School Inputs. Washington, DC: U.S.Department of Education, National Center for Education Statistics. (NCES Working Paper No. 98–04).

Chambers, Jay and Fowler, William J., Jr. 1995. Public School Teacher Cost Differences across the UnitedStates. Washington, DC: U.S. Department of Education, National Center for Education Statistics. (NCES 95–758).

Duncombe, William, Ruggiero, John, and Yinger, John. 1996. “Alternative Approaches to Measuring the Costof Education.” In Helen F. Ladd (ed.), Holding Schools Accountable. Washington, DC: The Brookings Institu-tion.

Fowler, William J., Jr. 1993. Public School Education Financing for School Year 1989–90. Washington, DC:U.S. Department of Education, National Center for Education Statistics. (NCES 93–095).

Greenlees, John S. and Mason, Charles C. December 1996. “Overview of the 1998 Revision of the CPI.”Monthly Labor Review. Washington, DC: U.S. Department of Labor, Bureau of Labor Statistics.

Hanushek, Eric A. 1997. “The Productivity Collapse in Schools.” In William J. Fowler, Jr. (ed.), Developmentsin School Finance, 1996. Washington, DC: U.S. Department of Education, National Center for EducationStatistics. (NCES 97–535).

McMahon, Walter W. 1996. “Intrastate Cost Adjustments.” In William J. Fowler, Jr. (ed.), Selected Papers inSchool Finance 1994. Washington, DC: U.S. Department of Education, National Center for Education Statis-tics. (NCES 96–068).

McMahon, Walter W. and Chang, S. 1991. “Geographical Cost-of-living Differences: Interstate and Intrastate,Update 1991.” McArthur/Spencer Series, No. 20. Normal, Illinois: Center for the Study of Educational Fi-nance, Illinois State University.

Rothstein, Richard and Mishel, Lawrence. 1997. “Alternative Options for Deflating Education Expendituresover Time.” In William J. Fowler, Jr. (ed.), Developments in School Finance, 1996. Washington, DC: U.S.Department of Education, National Center for Education Statistics. (NCES 97–535).

U.S. Department of Education. 1999. Digest of Education Statistics, 1998. Washington, DC: National Centerfor Education Statistics. (NCES 1999–036).

U.S. Senate, Committee on Finance. 1996. Final Report of the Advisory Commission to Study the CPI. Print104–72, 104 Cong., 2 sess., Washington, DC: Government Printing Office.

Williams, Janet L. December 1996. “The Redesign of the CPI Geographic Sample.” Monthly Labor Review.Washington, DC: U.S. Department of Commerce, Bureau of Labor Statistics.


54

Assessing the Financial Condition of Public School Districts

55

SelectedPapers in


Assessing the Financial Condition ofPublic School Districts:

Some Tools of the Trade


56


57


Some Tools of the TradeDean Michael Mead

Governmental Accounting Standards Board

About the Author

Dean Mead is a Project Manager at theGovernmental Accounting StandardsBoard (GASB). He is the author of a six-volume series of guides to state and localgovernment financial statements, writtenwith users of financial statements in mind.Dean coordinates the GASB’s outreach ef-forts to the users of governmental finan-cial statements and is co-manager of theGASB’s economic condition reportingproject. Prior to joining the GASB, Deanwas the Deputy Research Director at theCitizens Budget Commission in New York.

He was a periodic contributor to the opin-ion pages of New York Daily News andNewsday, and co-authored Power Failure:Politics and Policy in New York City Since1960 (published by Oxford UniversityPress).

Dean is also a member of the adjunct fac-ulty at New York University’s Robert F.Wagner Graduate School of Public Service,where he is completing his doctorate inpublic administration. He holds an under-graduate degree in public policy fromCornell University.


58


59

Introduction

The term “financial condition” means dif-ferent things to different people. Some con-sider it to be a school district’s financialstanding at a given point in time. Somethink of it as a district’s ability to makeends meet. Others look at it as a district’scapacity to raise revenue. The definitionof financial condition employed in thispaper is broader, encompassing aspects ofeach of those definitions:

Financial condition is the ability ofa school district to meet its obliga-tions as they come due and to fi-nance the services its constituencyrequires.

Understood in these terms, financial con-dition is a comprehensive barometer of aschool district’s overall fiscal health. Assuch, it is determined by a myriad of fac-tors, including the health of the localeconomy, the disposition of the politicalenvironment, and the prevailing wishes of

the citizenry. The principal informationabout a school district’s financial conditionis clearly financial in nature and is derivedfrom a variety of sources, including budgetdocuments, bond prospectuses, and finan-cial statements. This paper provides anoverview of the financial information toconsider when attempting to determine aschool district’s financial condition, particu-larly that information contained in adistrict’s annual financial statements.

This paper also pays particular attentionto the financial statement informationschool districts will provide when they haveimplemented Governmental AccountingStandards Board (GASB) Statement 34.1

Statement 34, issued in June 1999, dramati-cally revised the format school districts fol-low when preparing financial statementsaccording to generally accepted account-ing principles (GAAP). Among the majorfeatures of the new financial statements thatschool districts will be preparing are:


Some Tools of the TradeDean Michael Mead

Governmental Accounting Standards Board

1 Governmental Accounting Standards Board. 1999. Statement No. 34: Basic Financial Statements—and Management’s Discussion andAnalysis—for State and Local Governments. Norwalk, CT: GASB.


60

� Two new financial statements, coveringall of a district’s activities and preparedusing full accrual accounting (see tables1 and 2)

� More detailed information in the fundfinancial statements (see tables 5 and 6)

� A narrative management’s discussionand analysis (MD&A), prepared by adistrict’s finance officers, that offers asummary analysis of the financialstatements.

The body of this paper is presented in eightsections. The first section of this paper sum-marizes some important considerations re-garding the practice of financial conditionassessment and describes what categoriesof information should be considered andwhy. Then, each of the next six sectionsdiscusses one of those categories of finan-cial condition information: common-sizeratios; financial position and changes infinancial position; liquidity and solvency;fiscal capacity; risk and exposure; andother factors. The final section offers somethoughts regarding other sources of finan-cial data and their comparative advantagesand disadvantages relative to financialstatement information.

Overview of Finanical ConditionAnalysis

Financial condition analysis is not a regi-mented, strictly-defined science. There areat least two reasons why it may be consid-ered more of an individualistic art form.First, people measure financial conditionin many different ways. Second, financialcondition ultimately boils down to a sub-jective decision by the analyst, so that twopeople looking at the same financial infor-mation can come to different conclusionsabout a school district’s condition. Theseshould not be construed as weaknesses infinancial condition analysis. Financial con-dition analysis varies in form and function

because the needs and interests of analystsvary considerably.

If not a weakness, though, this inherent flex-ibility can be a threat to the credibility andreliability of the conclusions that an ana-lyst draws from his analysis. It is mucheasier to defend your conclusions about adistrict’s financial condition if your analy-sis is methodologically unassailable. Pro-tecting and enhancing the credibility of yourfinancial condition analysis can be accom-plished by following two rules.

First, financial information does not existin a vacuum. Because there are few or noabsolutes when assessing school district fi-nances, some kind of context is needed inorder to make the information meaning-ful. In other words, compare the informa-tion to a benchmark in order to knowwhether the information is telling good orbad things about a school district. Suchcomparisons typically are made with other,similar school districts, or with a district’sinformation for prior years. This compari-son can tell if financial condition is improv-ing or diminishing, or if a district’s condi-tion compares favorably or unfavorablywith other districts. For these reasons, thefirst tools considered in this paper are com-mon-size ratios, which can provide a con-text for understanding financial statementinformation and offer a common metric forcomparisons.

Second, thoroughness is next to godliness.The fewer stones left unturned, the betterthe analysis potentially may be. As we shallsee, one or two or even several financialratios cannot provide a complete pictureof a school district’s finances. A single fi-nancial factor may suggest one thing abouta school district, but when consideredalongside other information may suggest adifferent conclusion. Financial conditionanalysis is an iterative process, like peelingaway the layers of an onion. Calculatingfinancial ratios will raise questions and

Financial condition

analysis varies in

form and function

because the needs

and interests of

analysts vary

considerably.


61

identify potential issues, leading to furtherand more in-depth examination of adistrict’s finances. The effort to answerthese questions may raise further questionsstill. So, how thorough is thoroughenough? Ultimately, a decision will needto be made about how far to pursue thequestioning, by weighing the potential ben-efits of additional analysis against the ef-fort they require, and based on a personalcomfort level that the analysis has reason-ably considered all relevant factors.

The definition of financial condition usedin this paper suggests that two time dimen-sions must be considered—the present andthe future. On the one hand, we want toknow where a district stands financiallyright now—what kinds of debts and obli-gations does it face, what kinds of resourcesdoes it have available to repay them, andwhat resources will be left over? On theother hand, we need insight regarding thedistrict’s wherewithal to continue provid-ing services in the future—will it obtainsufficient resources in the near and longterm to cover its costs? Can it afford toissue debt or raise taxes? What potentialrisks are looming just beyond the horizon?

Understanding a district’s current financialstanding—its financial position—necessi-tates a comparison of the asset and liabil-ity information in its statement of net as-sets or balance sheet. Because it is impor-tant to place information in a comparativecontext, we are also interested in informa-tion in a district’s income statement thatdescribes changes in financial position.

Once we have established where a districtstands now, we must look to informationthat describes where it is going. Liquidityratios can inform us about whether a dis-trict will be able to pay its bills in the shortrun, and whether it will bring in sufficientresources to cover the recurring, annualcosts of operating a school system. Adistrict’s ability to meet its long-term obli-gations—to repay bonds, or to cover fu-

ture costs like compensated absences andleave pay—can be assessed with the use ofsolvency ratios.

Because the future is fraught with uncer-tainty, it is crucial to understand whether adistrict has the capacity or flexibility torespond to the needs of those it serves. Thispaper applies the general rubric of fiscalcapacity to the set of financial ratios thatspeak to these concerns. Can the districtraise taxes if costs rise? Can it issue bondsto finance capital needs? Can it afford evenits current levels of taxes and debt? Thesucceeding section of the paper presents ad-ditional ratios that can help ferret out thepotential financial risks a district may face,as well as its exposure to potential finan-cial problems.

Although this paper predominantly consid-ers financial condition information thatderives from annual financial statements,other factors—most notably socioeconomicand demographic data, and performancemeasures—are equally crucial to develop-ing a focused picture of school district fi-nancial health. The penultimate section ofthis paper describes some of those factorsin general.

Categories of FinancialCondition Ratios

Common-size Ratios

The most basic financial analysis tools arepercentage change and percentage distribu-tion, collectively referred to as common-size ratios because they put financial infor-mation in a form that allows meaningfulcomparisons among districts of varyingsizes. Did outstanding debts grow orshrink? Was state aid a larger or smallershare of revenue? In most cases, common-size ratios should be the first analyses youperform, because they can provide a quickoverview of the finances of a district andhow they have changed.

In most cases,

common-size

ratios should be

the first analyses

you perform,

because they can

provide a quick

overview of the

finances of a

district and how

they have changed.


62

Percentage distribution indicates the por-tion of a total category represented by in-dividual elements within the category—for example, the percentage of total ex-penses accounted for by various school dis-trict activities, such as regular education,special education, counseling, and admin-istration. Percentage distribution (see ra-tio 1 below) is calculated using unre-

stricted net assets from the district-widestatement of net assets (table 1) as an ex-ample.

Percentage change is the magnitude ofchange from year to year, which is moremeaningful for comparisons with other dis-tricts than aggregate dollar changes. It is cal-culated (ratio 2) using food services expenses

Table 1.—District-wide statement of net assets

NOTE: Italicized numbers are those used in the ratios.

SOURCE: Adapted from Mead, Dean Michael. 2000. What You Should Know about Your School District’s Finances: A Guide to

Financial Statements. Norwalk, CT: Governmental Accounting Standards Board.

Governmental Business-typeactivities activities Total

ASSETSCash and cash equivalents 106,268,980$ 7,828,243$ 114,097,223$ Property taxes receivable (net) 12,182,730 12,182,730Due from other governments 19,968,336 2,002,921 21,971,257Other receivables 2,252,919 4,081 2,257,000Internal balances 615,597 (615,597)Inventories and prepaid expenses 1,537,230 1,949,526 3,486,756Nondepreciated capital assets 32,272,411 32,272,411Depreciated capital assets 381,428,545 11,549,456 392,978,001 Less: Accumulated depreciation (98,176,725) (9,016,026) (107,192,751) Total assets 458,350,023 13,702,604 472,052,627LIABILITIESAccounts payable and other current liabilities 33,305,354 580,730 33,886,084Deferred revenues 3,117,910 723,038 3,840,948Long-term obligations Due within one year 21,569,854 21,569,854 Due beyond one year 108,793,747 108,793,747 Total liabilities 166,786,865 1,303,768 168,090,633NET ASSETSInvested in capital assets, net of related debt 231,118,669 2,533,430 233,652,099Restricted for: Debt service 4,133,180 4,133,180 School-based activities 1,396,569 1,396,569Unrestricted 54,914,740 9,865,406 64,780,146 Total net assets 291,563,158$ $ 12,398,836$ $ 303,961,994$ $

As of June 30, 2002

Example Independent School DistrictStatement of Net Assets

Ratio 1.—Percentage distribution

(individual element amount ÷ total category amount) x 100 =(unrestricted net assets ÷ total net assets) x 100 =

(64,780,146 ÷ 303,961,994) x 100 = 21.3 percent

Ratio 2.—Percentage change

[(current year amount – earlier year amount) ÷ earlier year amount] x 100 =

[(20,596,032 – 18,965,236) ÷ 18,965,236] x 100 = 8.6 percent increase


63

from the district-wide statement of activities(table 2) as an example.2

Financial Position

Financial position essentially is a district’s fi-nancial standing at a given time, based on acomparison of the resources it generally ownsor controls with the obligations it faces. Oneof the benefits of Statement 34 is that it re-quires school districts to prepare two district-wide financial statements using accrual ac-counting—the statement of net assets (table1) and the statement of activities (table 2).These statements are the first available pre-sentation of all of a school district’s assetsand liabilities and all of its revenues, ex-penses, and other changes in net assets to-gether in one place. It is the most compre-hensive accounting of financial position everavailable.

The most straightforward indicator of finan-cial position is “assets minus liabilities” ornet assets. Table 2 shows that total net as-sets for a fictional “Example” independentschool district (ISD) increased $20,341,902in fiscal year 2002. All things being equal,an increase in net assets is an improvementin financial position. Likewise, a decrease innet assets is a decline in financial position.

In order to obtain a sense of the magnitudeof the change in financial position, some ana-lysts compare the aggregate net assetsamount to the district’s overall level of fi-nancial activity, using either revenues or ex-penses. It is calculated in ratio 3 for ExampleISD using information in tables 1 and 2.

This ratio can be examined over time andcompared to other similar entities after it hasbeen placed in context. This is necessary be-cause comparing aggregate net assets or

change in net assets amounts is mislead-ing if a district’s level of financial activitychanges over time or if districts of vari-ous sizes are compared. If two districts—one with annual expenses of $20 millionand the other $80 million—each had a$1 million increase in net assets, the in-crease is obviously much more significantfor the district with the lower annual ex-penses.

In order to facilitate comparisons, con-sider building a comparison group. Acomparison group may consist of nearbyschool districts, or all districts of a simi-lar size in the same state, or similar dis-tricts in neighboring states, or some othermeaningful assemblage. Using informa-tion from those districts’ financial state-ments, calculate average or median finan-cial ratios, against which the district ofinterest may be compared. Table 3 showsthe net assets divided by expenses ratiofor 10 fictional districts in the same stateas Example ISD and calculates a relativeindex by dividing Example’s ratio by the10-district average.3 Example’s ratio ofnet assets to expenses is higher than eightof the 10 districts’ ratios, and about 14percent higher than the average for thecomparison group. (See the final sectionof this paper for further discussion ofcomparative financial information.)

Such comparisons become even moremeaningful when they are tracked overtime. Table 4 presents the trend in therelative index for the last 5 years, reveal-ing that Example’s ratio has been improv-ing relative to the other districts.

The statement of net assets divides netassets into three categories—the portionthat is related to capital assets, the por-tion that is restricted to specific uses, and

2 Comparisons of current and prior year financial information can be found in a district’s MD&A.

3 Note that the average is not calculated by adding the ratios together and dividing by 10, but by dividing the sum of all 10 districts’ netassets by the sum of all 10 districts’ expenses.

Ratio 3.—Financial position

net assets ÷ total expenses = 303,961,994 ÷ 410,123,330 = 0.741

Financial position

essentially is a

district’s financial

standing at a given

time, based on a

comparison of the

resources it

generally owns or

controls with the

obligations it faces.


64

the portion that is unrestricted. This allows more spe-cific assessments of financial position, if preferred. Thechange in just unrestricted net assets, rather than totalnet assets can be tracked using multiple years’ financialstatements. The net assets of just the governmental ac-

Table 2.—District-wide statement of activities

* This amount excludes the depreciation that is included in the direct expenses of the various programs.


SOURCE: Adapted from Mead, Dean Michael. 2000. What You Should Know about Your School District’s Finances: A Guide to Financial Statements. Norwalk, CT:

Governmental Accounting Standards Board.

tivities—the typical, bread-and-butter activities of schooldistricts, like pupil education and transportation—mayalso be examined if the resources of the business-typeactivities cannot be used to finance the district’s otheractivities.

Operating CapitalCharges for grants and grants and Net (expense)

Functions/Programs Expenses services contributions contributions revenue Governmental activities Instruction and instruction-related services 234,774,862$ 5,509,719$ 27,631,301$ (201,633,842)$ Instructional and school leadership 33,579,907 3,783,490 (29,796,417) Support services—student-based 37,311,861 2,986,172 4,203,974 (30,121,715) Administrative support services 9,365,149 1,055,183 (8,309,966) Support services—nonstudent-based 57,379,902 5,465,065 (51,914,837) Community services 2,753,346 131,297 (2,622,049) Interest on long-term debt 5,969,465 (5,969,465) Depreciation—unallocated* 6,555,053 (6,555,053) Total governmental activities 387,689,545 8,495,891 42,270,310 (336,923,344) Business-type activities Food services 20,596,032 4,750,350 15,849,235 750,000$ 753,553 Adult education 1,837,753 936,150 1,102,491 200,888 Total business-type activities 22,433,785 5,686,500 16,951,726 750,000 954,441 Total school district 410,123,330$ $ 14,182,391$ $ 59,222,036$ $ 750,000$ $ (335,968,903)$ $

Governmental Business-type Totalactivities activities district

Changes in net assets Net (expense) revenue (from above) (336,923,344)$ 954,441$ (335,968,903)$ General revenues Taxes Property taxes, levied for general purposes 154,108,322 154,108,322 Property taxes, levied for debt service 16,860,557 16,860,557 State aid—formula grants 176,265,211 176,265,211 Investment earnings 7,397,103 312,271 7,709,374Special item —gain on sale of unimproved land 1,367,341 1,367,341 Total general revenues and special item 355,998,534 312,271 356,310,805 Change in net assets 19,075,190 1,266,712 20,341,902Net assets—beginning 272,487,968 11,132,124 283,620,092Net assets—ending 291,563,158$ $ 12,398,836$ $ 303,961,994$ $

Program revenues

Example Independent School DistrictStatement of Activities

For the Year Ended June 30, 2002


65

Table 3.—Net assets-to-expenses ratio for“ Example” independent schooldistrict’s (ISD) comparison group,fiscal year 2002

SOURCE: Author’s sketch.

Sample ISD 0.619Standard ISD 0.514General ISD 0.489Regular ISD 0.567Illustrative ISD 0.687Comparison ISD 0.641Average ISD 0.682Typical ISD 0.801Common ISD 0.776Usual ISD 0.595

10-district average 0.651

Example ISD 0.741

Index (example ÷ average) 11.138

Change in financial position for a given yearcan also be placed in context by comparingit to financial activity (ratio 4).

Care should be shown in the interpretationof any financial ratio, because there are al-

most always “extenuating circum-stances.” In other words, instead of tak-ing a ratio at face value, an understand-ing of the circumstances in which the ra-tio was produced should be sought. Forexample, it is not sufficient to comparetwo governments and conclude that theone with the higher change in net assetsas a percentage of expenses was finan-cially healthier in a given year. Such a con-clusion should be based on a wider arrayof relevant information than just this oneratio.

Furthermore, an understanding of why netassets changed is needed. Growth in netassets might be the result of a healthyeconomy that generated revenues in ex-cess of expenses. On the other hand, adistrict might have ended the year withmore revenues than expenses because itfailed to provide needed services. Clearly,judgments about financial conditionwould differ in these two cases.

Depending on the aim of your analysis,you may want to limit it to common, re-curring revenues and expenses, therebyinsulating it from the distortion of finan-cial impacts related to unusual events orcircumstances. Statement 34 requires thatschool districts report special items andextraordinary items, transfers, and certainother changes in net assets separately frommore typical revenues and expenses (orexpenditures, in the case of the govern-mental funds statements).4 Tables 2 and6 reveal that Example ISD sold a parcelof land during the year.5 Because this isnot likely to be a common activity of aschool district, you might want to exclude

Table 4.—Index comparison of “ Example”independent school district’s (ISD)net assets-to-expenses ratio withthe 10-district average, fiscal years1998–2002

SOURCE: Author’s sketch.

Fiscal year 1998 0.997Fiscal year 1999 1.060Fiscal year 2000 1.112Fiscal year 2001 1.129Fiscal year 2002 1.138

Ratio 4.—Change in financial position

change in net assets ÷ total expenses = 20,341,902 ÷ 410,123,330 = 0.0496

4 Statement 34 defines extraordinary items as both unusual in nature and infrequent in occurrence, while special items are either unusualor infrequent (but not both). Further, special items are within the control of the district.

5 Table 6 is prepared using modified accrual accounting and therefore shows the total proceeds from the land sale. Table 2, however, usesaccrual accounting, and therefore shows only the gain (or loss) on the sale—proceeds minus original cost.


66

it from your analysis. Similarly, if a tor-nado were to strike the New England re-gion (an extremely rare occurrence), theensuing repairs and cleanup could be con-sidered uncommon costs that you mightwant to exclude. Additional ratios to con-sider (ratios 5–8) are listed at the bottomof the page.

Liquidity and Solvency

Liquidity

Liquidity and solvency are essentially con-cerned with a school district’s ability topay its bills in the short and long run, re-spectively. As with financial position, theassessment of liquidity and solvency typi-cally involves comparing a district’s re-sources with its outstanding debts. For ex-ample, the current ratio (ratio 9) can beused to assess a district’s ability to raiseresources to cover its obligations over thecoming year.

Current assets are those that are expectedto be liquidated within a year, and cur-rent liabilities are those that are expectedto come due within a year. This currentratio means that Example ISD has currentassets totaling approximately 2.6 timesmore than is needed to finance current li-abilities.

A more stringent approach to liquiditymight apply the quick ratio (ratio 10),

which compares only the most liquid assetsof a district—generally cash, near-cash as-sets such as money market funds, othershort-term investments, and sometimes re-ceivables—to its current liabilities.

This is a more conservative indicator of adistrict’s ability to meet obligations, assum-ing implicitly that its current liabilities willcome due so soon that certain current as-sets cannot be liquidated quickly enough topay for them.

Some districts may present their statementof net assets in a classified format, meaningthey show current assets and liabilities sepa-rately from noncurrent assets and liabilities.Many districts, such as Example ISD, do not;in such cases, difficulty may arise in discern-ing which assets and liabilities are current,and thus the governmental funds financialstatements (tables 5 and 6) might be usedalternatively. The governmental funds state-ments are prepared using the modified ac-crual basis of accounting and the current fi-nancial resources measurement focus, whichmeans that the balance sheet generally onlyshows current assets and liabilities.

One potential complication related to modi-fied accrual accounting is the liability called“deferred revenues.” There are generally tworeasons why the recognition of revenues maybe deferred:

...the assessment

of liquidity and

solvency typically

involves

comparing a

district’s resources

with its

outstanding debts.

Ratio 5.—Financial position using revenues

Net assets ÷ total revenues (or operating revenues)

Ratio 6.—Change in financial position using revenues

Change in net assets ÷ total revenues (or operating revenues)

Ratio 7.—Financial position using fund balance

Fund balance ÷ total expenditures (or revenues, operating revenues)

Ratio 8.—Financial position using unreserved fund balance

Unreserved fund balance ÷ total expenditures (or revenues, operating revenues)

Ratio 9.—Current ratio

current assets ÷ current liabilities =

(114,097,223 + 12,182,730 + 21,971,257 + 2,257,000 + 3,486,756) ÷ (33,886,084 + 3,840,948 +21,569,854) = 2.597

Ratio 10.—Quick ratio

(cash + current investments) ÷ current liabilities =

114,097,223 ÷ (33,886,084 + 3,840,948 + 21,569,854) = 1.924


67

(1) If a district receives resources that require itto provide certain services or meet specificconditions, those resources are not recognizedas revenues until the services have beenprovided or the conditions have been met.

(2) Under modified accrual, if the resources arenot available during the year or soon enoughthereafter to pay current liabilities, they aredeferred. For instance, property taxes must becollected during the year or within 60 days of

the end of the fiscal year in order to berecognized as revenue; anything expected to becollected after the 60-day period is deferred.

Deferred revenue is relevant to computing liquid-ity ratios because the comparison of liquid as-sets with current liabilities should include nei-ther assets that are not generally available to sat-isfy liabilities, nor liabilities that are not liqui-dated in the traditional way. If a district has de-

Table 5.—Governmental funds balance sheet


SOURCE: Adapted from Mead, Dean Michael. 2000. What You Should Know about Your School District’s Finances: A Guide to Financial

Statements. Norwalk, CT: Governmental Accounting Standards Board.

Other Total General Debt service governmental governmental

fund fund funds fundsASSETSCash and cash equivalents 100,864,805$ 3,294,850$ 2,109,325$ 106,268,980$ Property taxes receivable, net 10,341,512 1,841,218 12,182,730 Due from other governments 15,105,826 4,862,510 19,968,336 Accrued interest 504,757 504,757 Due from other funds 5,170,479 759,359 1,852,454 7,782,292 Other receivables 1,218,640 20,695 508,827 1,748,162 Inventories—supplies and materials 1,412,121 1,412,121 Other current assets 125,109 125,109 Total assets 134,743,249$ $ 5,916,122$ $ 9,333,116$ $ 149,992,487$ $ LIABILITIES AND FUND BALANCESLiabilities Accounts payable and accrued liabilities 30,270,632$ 8,740$ 933,434$ 31,212,806$ Due to other funds 20,845,752 5,503,492 26,349,244 Due to other governments 10,093 10,093 Due to student groups 256,183 256,183 Deferred revenue 12,283,000 1,774,202 1,243,438 15,300,640 Amounts held for granting agencies 233,035 233,035 Total liabilities 63,642,512 1,782,942 7,936,547 73,362,001 Fund balances Reserved Inventories 1,412,121 1,412,121 Retirement of long-term debt 4,133,180 4,133,180 Encumbrances 4,744,173 4,744,173 Unreserved Designated 21,347,665 21,347,665 Undesignated, reported in: General fund 43,596,778 43,596,778 Special revenue funds 1,396,569 1,396,569 Total fund balances 71,100,737 4,133,180 1,396,569 76,630,486 Total liabilities and fund balances 134,743,249$ 5,916,122$ 9,333,116$ 149,992,487$

As of June 30, 2002

Example Independent School DistrictBalance Sheet

Governmental Funds


68

6 In the case of deferred revenues related to conditions that must be met or actions that must be taken, you may also want to subtract the related assets—typically cash or receivables—if they are significant relative to total assets.

Table 6.—Governmental funds statement of revenues, expenditures, and changes in fund balances

Other Total General Debt service governmental governmental

fund fund funds fundsREVENUESProperty taxes 153,862,367$ 16,589,425$ 170,451,792$ Interest 7,077,388 194,926 124,789$ 7,397,103 Tuition charges 1,283,778 4,225,941 5,509,719 Facility rental fees 2,437,009 2,437,009 State revenues 188,019,530 6,135,833 194,155,363 Federal revenues 2,284,748 22,095,410 24,380,158 Other 107,604 441,559 549,163 Total revenues 355,072,424 16,784,351 33,023,532 404,880,307

EXPENDITURESCurrent Instruction and instructional-related services 206,958,475 25,936,202 232,894,677 Instructional and school leadership 31,485,279 1,825,705 33,310,984 Support services—student 34,010,001 3,003,049 37,013,050 Administrative support services 9,290,149 9,290,149 Support services—nonstudent-based 55,615,563 1,308,415 56,923,978 Community services 1,691,107 1,040,189 2,731,296 Debt service Principal 1,160,471 11,985,914 380,561 13,526,946 Interest 378,447 3,908,791 124,107 4,411,345 Capital outlay 922,537 8,327 930,864 Total expenditures 341,512,029 15,894,705 33,626,555 391,033,289

Excess (deficiency) of revenues over expenditures 13,560,395 889,646 (603,023) 13,847,018

OTHER FINANCING SOURCES (USES) Proceeds from capital leases 692,245 692,245 SPECIAL ITEM Proceeds from sale of unimproved land 2,601,908 2,601,908 Net change in fund balances 16,162,303 889,646 89,222 17,141,171Fund balance—Beginning 54,938,434 3,243,534 1,307,347 59,489,315 Fund balance—Ending 71,100,737$ $ 4,133,180$ $ 1,396,569$ $ 76,630,486$ $

Governmental FundsFor the Year Ended June 30, 2002

Example Independent School DistrictStatement of Revenues, Expenditures, and Changes in Fund Balances

SOURCE: Adapted from Mead, Dean Michael. 2000. What You Should Know about Your School District’s Finances: A Guide to Financial

Statements. Norwalk, CT: Governmental Accounting Standards Board.

ferred revenues that are significant relative toeither its assets or liabilities, then the deferredrevenues should be subtracted from the liabili-ties in the denominator of the ratio.6 A currentratio calculated from table 5 is shown in ratio11.

The determination of what is an acceptable levelof liquidity is subjective, but it may be said that“acceptable” depends on some cushion of re-sources above and beyond what is necessary to

Ratio 11.—Current ratio from table 5 (without deferred revenues)

149,992,487 ÷ (73,362,001 – 15,300,640) = 2.583


69

exactly cover obligations. In other words, aprecise match of resources and obligations—a ratio of 1.0—leaves no room for the un-foreseen. If certain receivables are not actu-ally received, then resources may not beavailable to finance obligations when theycome due. Consequently, you may seek a ra-tio that indicates some slack to meet con-tingencies—whether that ratio is 1.5 or 3.0or somewhere in between depends on par-ticular interests.

Proper interpretation of liquidity ratios re-quires an understanding of a district’s cashflow patterns. Because the ratios are basedon information as of the end of the fiscalyear, they may not accurately reflect theavailability of cash in all situations. For ex-ample, a district might have a low currentratio as of the end of the year, but not havea liquidity problem if it receives a substan-tial portion of its property taxes early in theyear. By contrast, a high liquidity ratio maybe the result of state aid payments that arereceived just prior to the end of the year. Aslong as the circumstances of a particular dis-trict are understood, a knowledgeable com-parison of these ratios with prior years andwith other districts is possible.

Solvency

Solvency ratios come in two general forms—leverage ratios and coverage ratios. Lever-age is the degree to which a district’s assetsare financed through borrowing and otherlong-term obligations. The debt-to-assetsratio (ratio 12) divides total liabilities by to-tal assets. The debt-to-net-assets ratio (ra-tio 13) divides total liabilities by net assets.

The first calculation produces a ratio of 0.36,which means that more than one-third ofthe district’s assets are financed with debt.

The latter ratio is 0.55, meaning for everydollar of resources the district has avail-able to use for providing public services,it owes 55 cents.

Times-interest-earned is a coverage ratiothat compares cash flows generated by op-erations to interest payments on debt.Debt service coverage compares cashflows to a district’s entire debt repayments,both interest and principal. The difficultythat will be faced will be trying to calcu-late coverage ratios without cash flow in-formation for the district’s governmentalactivities. (By contrast, accounting rulesrequire districts to prepare a cash flowstatement for their enterprise funds, whichin most cases are the same as their busi-ness-type activities.)

If cash flow information cannot be ob-tained from a district, the alternativewithin the financial statements is compli-cated. In order to calculate the ratio, themissing cash flow information for govern-mental activities must be replaced withsecond-best information—the differencebetween revenues and expenditures in thegovernmental funds statements (table 6).This is problematic on at least two levels:

(1) Because the governmental fundsinformation, as noted, is prepared on amodified accrual basis, not on a cashbasis, the information may overstate orunderstate actual cash flows.

(2) The “excess (deficiency)” amount in thegovernmental funds includesexpenditures for capital projects, as wellas for debt service, neither of whichwould be reflected in a “cash flow fromoperations” amount if one existed.Consequently, the alternative equation

Proper

interpretation of

liquidity ratios

requires an

understanding of a

district’s cash flow

patterns.

Ratio 12.—Debt-to-assets ratio

total liabilities ÷ total assets = 168,090,633 ÷ 472,052,627 = 0.356

Ratio 13.—Debt-to-net assets ratio

total liabilities ÷ total net assets = 168,090,633 ÷ 303,961,994 = 0.553


70

should include just the revenues andcurrent expenditures of the general fundand special revenue funds.7

In light of these potential problems, theuncertainties surrounding the informationproduced by coverage ratios may outweighthe effort to calculate them. The times-in-terest-earned ratio is shown in ratio 14.8

The formula for the debt service coverageratio is shown in ratio 15.9

The debt-to-assets ratio suggests whethera district has the resources needed to re-pay its long-term debts. The ratio of 0.36calculated above implies that, to repay thedebts immediately would require the liq-uidation of 36 percent of the district’s as-sets. Is this possible? For some districts,capital assets such as school buildings andbuses represent the overwhelming major-ity of assets; such assets are not easily liq-uidated, and the district could not con-tinue to operate if they were sold off. Nev-

ertheless, it is commonplace to issue long-term debt to finance capital assets and re-pay it roughly over the useful life of the as-sets. The appropriate level of outstandingdebts relative to assets is a subjective judg-ment call.

The debt-to-net-assets ratio may be consid-ered a reflection of who owns a district’sassets, the district or its creditors. All otherfactors being equal, persons who have com-pletely paid for their homes might appearbetter off than persons with 15 years left ona 30-year mortgage.

Judgments about coverage ratios vary likeliquidity ratios, depending upon what isconsidered to be a comfortable or accept-able cushion of additional resources to coverdebt. In other words, it depends on how farabove 1.0 the ratios must be to assure ananalyst that debt can be repaid regardlessof unforeseen events. Additional ratios toconsider (ratios 16 and 17) are listed at thebottom of the page.

7 Under “major fund” reporting, the general fund or its equivalent will always be shown on the face of the statement. However, only themajor special revenues funds will be shown individually. You will need to determine if there are additional special revenue funds includedin the nonmajor “other governmental funds” column. This can be easily accomplished if a school district includes a combining statementof nonmajor governmental funds in its comprehensive annual financial report.

8 Interest on long-term debt for governmental and business-type activities can be found in the district-wide statement of activities (table 2).

9 Principal repayments may be found in the required note disclosure for long-term liabilities activity. See figure 12 in Mead (2000).

Ratio 14.—Tmes-interest-earned ratio

(cash flow from operations + interest expense) ÷ interest expense =

(general fund revenues + special revenue fund revenues – general fund current expenditures – specialrevenue fund current expenditures + enterprise funds cash flow from operations + total interest onlong-term debt for governmental and business-type activities) ÷ (total interest on long-term debt forgovernmental and business-type activities)

Ratio 15.— Debt service coverage ratio

(cash flow from operations + debt service) ÷ debt service =

(general fund revenues + special revenue fund revenues – general fund current expenditures – specialrevenue fund current expenditures + enterprise funds cash flow from operations + total interest onlong-term debt for governmental and business-type activities + total principal repayments forgovernmental and business-type activities) ÷ (total interest on long-term debt for governmental andbusiness-type activities + total principal repayments for governmental and business-type activities)

Ratio 16.—Liabilities as a share of annual revenues

Liabilities ÷ total revenues (or operating revenues)

Ratio 17.—Resources devoted to repaying debt

Debt service ÷ total revenues

…the uncertainties

surrounding the

information

produced by

coverage ratios

may outweigh the

effort to calculate

them.


71

Fiscal Capacity

Fiscal capacity is a district’s ability to raiseresources to finance the provision of the ser-vices its constituency demands. The simplestmeasures of fiscal capacity combine finan-cial statement information with economicand demographic data. These ratios com-pare revenues, expenses, and outstandingdebts with indicators that imply aconstituency’s wherewithal to pay for ser-vices. Much of the information you woulduse could be found in a district’s compre-hensive annual financial report (CAFR), orcan easily be found at federal governmentInternet sites.10 Three common indicators towhich financial information may be com-pared are property values, personal income,and population (ratios 18–20).

Any of these ratios could also be calculatedwith revenues or expenses in the numeratorinstead of liabilities. Specific revenues canalso be compared with their relevant eco-nomic bases, such as property tax revenuesper $100 of assessed value. Such a calcula-tion produces an effective tax rate—theamount actually recognized relative to thevalue of property owned, rather than theamount levied.

Population is frequently used to place finan-cial information in a metric that is easilycompared with other school districts. Some-

times the population used is that of adistrict’s geographic area, such as prop-erty taxes per capita. Other times the stu-dent population is utilized, such as spend-ing per pupil or state aid revenue per pu-pil. Additional ratios to consider (ratios21–23) are listed at the bottom of the page.

Risk and Exposure

A school district’s ability to withstand fi-nancial difficulties can be as important tojudgments about financial condition as itsability to raise revenues. One measure ofsuch capacity is revenue dispersion. Thedegree of dispersion or diversity in adistrict’s sources of revenue can tell some-thing about its exposure to financial dif-ficulty if a particular revenue source driesup. By examining the individual compo-nents of total revenues, it might be foundthat a district relies on a broad range ofrevenues to support its activities, andtherefore is relatively less likely to be dra-matically affected if one type of revenuedoes not meet expectations. Alternatively,it might be found—as is often the case—that the district relies heavily on one ortwo sources of revenue, such as propertytaxes and state formula aid. Reductionsin those kinds of revenues could harm adistrict financially because it is very de-pendent on them to run its operations.

10 For example, useful information can be obtained from the U.S. Bureau of the Census, http://www.census.gov, the Bureau of EconomicAnalysis, http://www.bea.doc.gov, and the National Center for Education Statistics, http://nces.ed.gov.

A school district’s

ability to

withstand

financial

difficulties can be

as important to

judgments about

financial condition

as its ability to

raise revenues.

Ratio 18.—Debt per $100 of assessed property value

(total liabilities x 100) ÷ total assessed property value

Ratio 19.—Debt per $1,000 of personal income

(total liabilities x 1,000) ÷ total personal income

Ratio 20.—Debt per capita

total liabilities ÷ total population

Ratio 21.—Taxes per capita

Tax revenues ÷ population

Ratio 22.—Taxes as a share of personal income

Tax revenues ÷ personal income (or assessed value)

Ratio 23.—Expenses per capita

Expenses ÷ population


72

Revenue dispersion is simple to derive, be-ing just a percentage distribution calcula-tion. Each individual revenue source is di-vided by total revenues (and then multi-plied by 100) to reveal its percentage orshare of total.

More involved ratios include those devel-oped by Bowman and Calia (1997 and1999)—risk exposure ratio and tax lever-age ratio. Although their work on gov-ernmental financial analysis was appliedto fund financial statements, it can easilybe adapted to the new district-wide state-ments required by GASB Statement 34.

The risk exposure ratio (ratio 24) focuseson revenue sources that are potentiallysubject to large, abrupt changes, specifi-cally investment income and intergovern-mental aid, the latter being particularlyimportant to school districts. The ratio ex-presses the percentage increase in prop-erty taxes that would be required to makeup for a 1 percent shortfall in those twosources of funding. It is calculated in ra-tio 24 using Example ISD’s statement ofactivities (table 2): a 1 percent shortfallin those revenues would require a 1.4 per-cent increase in property taxes.

This information can be compared withthe results found by calculating measuresof fiscal capacity. If a district does not havemuch available capacity to raise taxes orto borrow because its taxes and debt arealready relatively burdensome compared

with other similar districts, then a high riskexposure ratio could be cause for concern.That district may not have the ability to ef-fectively respond to a downturn in those rev-enue sources.

The tax leverage ratio (ratio 25), like riskexposure, is expressed as a percentage in-crease in property taxes. In this case, theratio shows how much property tax revenuewould need to increase to cover a 1 percentincrease in costs. Again using the ExampleISD data shown in table 2 (in order to getoperating expenses, first remove deprecia-tion11 and interest on long-term debt), theresult shows that a 1 percent increase inoperating expenses would require an in-crease in property taxes of almost 2.3 per-cent. With the tax-raising capacity informa-tion in hand, the tax leverage ratio can beutilized to assess a district’s ability to reactto unplanned costs. Additional ratios to con-sider (ratios 26 and 27) are listed at the bot-tom of the page.

Other Factors

Of course, financial statements are only onesource of information that is relevant to as-sessing a school district’s financial state-ments. Useful financial information can alsobe found in the statistical section of adistrict’s CAFR, in budget documents, andin the official statements districts publishwhen preparing to issue bonds or notes.

11 The figure of $6,554,236 is depreciation expenses allocated directly to the functions and programs listed in the statement of activities(table 2), and can be found in the required note disclosure regarding capital asset activity during the year. See figure 11 in Mead (2000)for an example.

Ratio 24.—Risk exposure ratio

(investment revenue + intergovernmental aid) ÷ property tax revenue =

(7,709,374 + 59,222,036 + 750,000 + 176,265,211) ÷ (154,108,322 + 16,860,557) = 1.4268

Ratio 25.—Tax leverage ratio

operating expenses ÷ property tax revenue =

(410,123,330 – 6,555,053 – 6,554,236 – 5,969,465) ÷ (154,108,322 + 16,860,557) = 2.287

Ratio 26.—Magnitude of property tax receivables

Property tax receivables ÷ current assets (or property tax levy)

Ratio 27.—Magnitude of unpaid property taxes

Uncollectable property taxes ÷ property tax levy

With the tax-

raising capacity

information in

hand, the tax

leverage ratio can

be utilized to

assess a district’s

ability to react to

unplanned costs.


73

An important financial issue that may beoverlooked is district employee pensions,which often are a district’s largest long-termobligations. Financial reporting rules requireschool districts that participate in a pensionfund (or that operate their own fund) to re-port information about the size of their pen-sion obligations and the extent to which re-sources have been set aside to fund thoseobligations.12 The information can be foundin a district’s required supplementary infor-mation accompanying its financial state-ments. This information includes several ra-tios (ratios 28–30) that indicate the fundingstatus of pensions (for the last 3 or 6 years).

There is also a considerable amount of ad-ditional information that can shed light ona government’s financial health. Berne(1992) cites a lengthy list of items, includ-ing:

� Economic and demographic information,such as population, school-agepopulation, birth rates, percentage ofpopulation in poverty, employment, andindustrial structure

� Revenue base information, such asproperty values, retail sales, and personalincome

� Service performance information,including indicators of:

� Service demands—enrollment inregular and special education,percentage of non-English-speakingstudents, poverty levels

� Service efforts—numbers and typesof employees, class sizes

� Service accomplishments—percentage of students graduating,test scores

Finally, it may not be sufficient to knowabout a district’s capacity to raise resourcesand to provide services, but also its will-ingness to do so. For example, fiscal ca-pacity ratios may suggest that it is feasibleto raise additional tax revenues, but adistrict’s board members may be unwill-ing to increase tax rates. Alternatively, atax increase may require voter approval,or a state law may prevent a district fromraising taxes. Valuable information regard-ing these issues can be found in policystatements, press releases, strategic plans,newspaper and magazine articles, and cer-tainly school district Web sites. State de-partments of education typically also haveinformative Web sites with both financialand nonfinancial information. Additionalratios to consider (ratios 31–37) are listedat the bottom of the next two pages.

12 At present there are no similar rules for reporting information about other benefits provided to retirees, such as health insurance.However, the GASB’s current technical plan envisions proposing new rules early in 2002 for reporting “other postemployment benefits”(OPEB).

Ratio 28.—Pension funding status

Actuarial value of pension fund assets ÷ unfunded actuarial accrued liability

Ratio 29.—Unfunded pension liability as a share of payroll

Unfunded actuarial accrued liability ÷ covered payroll

Ratio 30.—Actual versus required pension contributions

Actual pension contribution ÷ actuarially required contribution

Ratio 31.—Employees per capita

District employees ÷ population

Ratio 32.—Employees per pupil

Teachers (or administrative staff ) ÷ student enrollment

Ratio 33.—Magnitude of unfunded pension liability

Unfunded pension liability ÷ assessed value (or revenues, personal income)

An important

financial issue that

may be overlooked

is district

employee

pensions, which

often are a

district’s largest

long-term

obligations.


74

A Note on Comparative FinancialInformation

This discussion of financial condition hasfocused predominantly on financial infor-mation drawn from school district finan-cial statements. It is suggested that the ana-lyst of school district finances build a com-parison group of similar school districtsagainst which the financial status of aschool district may be framed. However,some may consider the process of calcu-lating ratios from the financial statementsof multiple school districts over severalyears to be too cumbersome or time con-suming. An alternative source of handy fi-nancial information may be the NationalCenter for Education Statistics (NCES).

The NCES is a unit of the U.S. Depart-ment of Education that, among other ac-tivities, compiles and disseminates finan-cial and nonfinancial information aboutpublic elementary and secondary educa-tion. The greatest strengths of the NCES’sdatabases, such as the one derived fromits annual survey of school system finances(the “F-33” form) or its Common Core ofData (CCD), are their ease of use and com-parability. The information is easily ma-nipulated with just about any spreadsheetor statistical software. For instance, withlittle difficulty an analyst could sort theinformation by school district size (i.e.,number of pupils), cut and paste 10 or 20or more districts of a relevant size, aggre-gate their information and compute com-parison group average ratios.

The greatest asset of NCES’s informationis the assurance of comparability and con-sistency it provides. The categories of in-

formation requested in the underlying sur-vey for the F-33 database are based on acommon chart of accounts that reflects thedata that state education agencies collect.This means the user of the database can bereasonably certain that the information iscomparable across districts and consistentover time.

The NCES data is not without its shortcom-ings, however. First, the information is notvery timely. Analyst complaints that schooldistrict financial statements are not avail-able for some 4 to 6 months after the end ofthe fiscal year pale in light of the 2- to 3-year lag in NCES data. Thus, although theNCES information can provide valuablecross-sectional and trend series comparisons,it is not up-to-date. Alternatively, state gov-ernments generally collect standardized fi-nancial information that is fresher than theNCES data, though typically more than oneor two years old. Second, the categorizationof information in the NCES databases (aswell as that of the state-gathered informa-tion) is not the same as you will find in aschool district’s financial statements. In otherwords, the two cannot be reliably compared.

Nevertheless, these information sources andothers are very important and informativecompanions to financial statement-orientedanalyses. The relative strengths and weak-nesses of each are complementary. For itspart, audited financial statement informa-tion is more timely and offers a level of as-surance of reliability that may not be presentin other data sources. Used in conjunction,these multiple types of financial informationshould provide a more comprehensive anddefinitive picture of a school district’s finan-cial condition than either would on its own.

Ratio 34.—Magnitude of maintenance costs

Maintenance and repair expenses ÷ capital assets

Ratio 35.—Magnitude of capital investment

Capital expenditures ÷ student enrollment (or capital assets)

Ratio 36.—Annual consumption of capital assets

Depreciation expense ÷ capital assets

Ratio 37.—Expired portion of capital assets useful lives

Accumulated depreciation ÷ capital assets

Used in

conjunction, these

multiple types of

financial

information

should provide a

more

comprehensive

and definitive

picture of a school

district’s financial

condition than

either would on its

own.


75

Acknowledgments

The opinions stated in this paper are thoseof the author. Official positions of the GASBare established only after extensive due pro-cess and deliberation.

Portions of GASB documents included in thispaper, copyright by Governmental Account-ing Standards Board, 401 Merritt 7,Norwalk, Connecticut 06857, are used bypermission. Complete copies of these docu-ments are available from the GASB.


76

References

Berne, Robert. 1992. The Relationship Between Financial Reporting and the Measurement of Financial Condi-tion. Norwalk, CT: Governmental Accounting Standards Board.

Bowman, Woods and Calia, Roland. 1999. Evaluating Local Government Financial Health: Financial Indica-tors for Major Municipalities in Northeastern Illinois. Chicago: The Civic Federation.

Bowman, Woods and Calia, Roland. 1997. Evaluating Local Government Financial Health: Financial Indica-tors for Cook, DuPage, Kane, Lake, McHenry and Will Counties. Chicago: The Civic Federation.

Governmental Accounting Standards Board. 1999. Statement Number 34: Basic Financial Statements—andManagement’s Discussion and Analysis—for State and Local Governments. Norwalk, CT: GASB.

Mead, Dean Michael. 2000. What You Should Know about Your School District’s Finances: A Guide to Finan-cial Statements. Norwalk, CT: Governmental Accounting Standards Board.

The Accounting and Resource Cost Model Approaches

77

SelectedPapers in


A Synthesis of Two Approaches toSchool-Level Financial Data:

The Accounting and Resource CostModel Approaches


78


79

William T. HartmanPennsylvania State University

Denny G. BoltonOwen J. Roberts School District


About the AuthorsWilliam T. Hartman

William T. Hartman is a professor of edu-cation in the College of Education at Penn-sylvania State University and a co-founderand the executive director of the Centerfor Total Quality Schools at Penn State. Heis a recipient of the Graduate FacultyTeaching Award for the University. Priorto coming to Penn State in 1986, he wason the faculty at the University of Oregonand Stanford University. He has also beena Visiting Fellow in Education at the Uni-versity of Sussex (England). He has servedon the Board of Directors of the AmericanEducation Finance Association.

Dr. Hartman received a Bachelor of Me-chanical Engineering with High Honorsfrom the University of Florida, a Mastersof Business Administration from HarvardUniversity, and a Ph.D. in EducationalAdministration from Stanford University.His areas of research and scholarship in-clude school budgeting, special educationfinance, school finance equity, total qual-ity management in education, and micro-



computer models in educational adminis-tration. His latest books are Resource Al-location and Productivity in Education andSchool District Budgeting.

Denny G. Bolton

Denny G. Bolton is the Chief Business Of-ficial at the Owen J. Roberts School Dis-trict in Pottstown, Pennsylvania. Dr. Boltonalso serves on the Board of Directors of theAmerican Education Finance Association.He is Vice-Chair of the Editorial Board anda member of the Accounting, Auditing, andBudgeting Committee of the Association ofSchool Business Officials International. Heis also an Adjunct Associate Professor atPennsylvania State University where heteaches graduate courses in school financeand human resources management.

Dr. Bolton holds a Ph.D. from the Univer-sity of Pennsylvania, M.B.A. from theWharton School, M.Ed. from Lehigh Uni-versity, M.A. from the University of Penn-sylvania, and B.A. from Ursinus College.He has served as the President of both the


80

Pennsylvania Association of School Busi-ness Officials and the Association ofSchool Business Officials International.He has been recognized as the Outstand-ing School Business Official of the Yearat the state and national levels. Under thesupervision of Dr. Bolton, the businessstaff of the Owen J. Roberts School Dis-trict has been recognized at both the stateand national levels for the quality of fis-cal management and technological inno-vation, including the Certificate of Ex-cellence in Financial Reporting, the Cer-tificate of Achievement for Excellence inFinancial Reporting, the DistinguishedBudget Presentation Award, and theMeritorious Budget Award.

David H. Monk

David H. Monk is a professor of educa-tional administration and the Dean of theCollege of Education at Pennsylvania

State University. He earned his Ph.D. in1979 at the University of Chicago andwas a member of the Cornell Universityfaculty for 20 years prior to becomingdean at Penn State. He has also taught ina visiting capacity at the University ofRochester and the University of Burgundyin Dijon, France.

Dr. Monk is the author of EducationalFinance: An Economic Approach andRaising Money for Education: A Guideto the Property Tax (with Brian O. Brent)as well as numerous articles in scholarlyjournals. He serves on the editorialboards of The Economics of EducationReview, The Journal of Education Fi-nance, Educational Policy, and the Jour-nal of Research in Rural Education. Heconsults widely on matters related to edu-cational productivity and the organiza-tional structuring of schools and schooldistricts and is a past president of theAmerican Education Finance Association.


81

Introduction

Purpose

For the past several years, there has beenconsiderable interest in measuring educa-tional expenditures at the school level forthe purpose of linking student costs to stu-dent results. This attention has developedas a variety of interested parties and is-sues have converged around the efforts ofimproving the effectiveness and efficiencyof schools. A common concern amongschool administrators, policymakers, andresearchers is the lack of reliable school-level data on which to make effective re-source allocation decisions based on in-formed education policy. “School-leveldata on public school expenditures are notgenerally available to inform educationpolicy discussions regarding how re-sources are allocated both within andamong schools” (Issacs et al. 1998).

Two alternative approaches to the devel-opment of school-level data for financialreporting and resource analysis haveemerged to address the data needs of ad-ministrators, policymakers, and research-ers. The two approaches are:



William T. HartmanPennsylvania State University

Denny G. BoltonOwen J. Roberts School District


Accounting approach

A downward extension of the present dis-trict-level accounting system to report ex-penditures by individual school buildings.

Resource cost model (RCM) approach

Use of physical resource data as the basisof measuring resource use and translatingthe consumption of resources into costs.

The purpose of this paper is to provide ananalysis of the strengths and weakness ofthese approaches with the objective of de-veloping a synthesis of their methodologiesthat combines their best features and avoidstheir primary problems. The synthesis leadsto recommendations for the content andformat of school-level financial and re-source data to be collected and reportedby school districts.

Previous NCES Working Papers

The National Center for Education Statis-tics (NCES) has taken a lead role in theinquiry of school-level data and commis-sioned several recent studies of both major


82

1 National Center for Education Statistics website, NCES Electronic Catalog, Product Information Page, NCES 1999–07.

2 Ibid.

approaches. This paper draws on the find-ings and analyses of these studies in theexamination of the accounting and re-source cost model (RCM) approaches todeveloping and using school-level finan-cial and productivity data.

Sherman, Best, and Luskin (1996) ana-lyzed available financial data from twostates that implemented data collection atthe school level. The analysis determinedschool-level expenditure amounts for ma-jor functions and instructional programsand then examined the variations in ex-penditures among functions and programsacross schools in each state. Drawing onthe insights from working with the dataavailable from these states’ accounting sys-tems, a series of recommendations weredeveloped for the design of a model school-level data collection system.

Issacs et al. (1999) examined the feasibil-ity and difficulties in collecting more de-tailed staffing resources and expendituresat the school level through the Schools andStaffing Survey (SASS). As described in theNCES abstract:

This working paper summarizes aseries of tasks undertaken to as-sess the feasibility of extending theresource and finance data collectedin the SASS. It includes an over-view of the RCM, instruments de-signed to collect staffing data, anappraisal of earnings data from theCurrent Population Survey (CPS)to estimate salaries for school staff,options for gathering benefits data,an approach to collecting tradi-tional finance data at the schoollevel, and a discussion of the ana-lytical value of an integrated col-lection of both staffing resourceand expenditure data.1

In the third report, Chambers (1999) com-pared the two alternative approaches with

measuring resources in K–12 education thatare the focus of this study—the accountingapproach and the RCM approach.

This report focuses on two ap-proaches to measurement of re-sources in education: an account-ing approach and a resource-basedapproach. The accounting ap-proach measures resources in dol-lars of expenditure. The resource-based approach emphasizes themeasurement of resources in termsof physical ingredients, such asteaching staff. The comparison ofthe accounting and the resource-based approaches explores the dif-ferences in the way accountants andeconomists view the concepts ofcost and expenditure. The reportfocuses on the development of aframework for organizing and ana-lyzing programmatic cost, expendi-ture, and resource data for localeducational agencies serving el-ementary and secondary students.2

Stakeholders and Uses of Data

The first questions to consider are who willbe the primary users of school-level data,what new data do they need, and for whatpurposes will the data be used? Without aclear understanding of these items, there canbe no reference point from which to chooseamong competing directions. Principalstakeholders can be divided into three gen-eral groups based on their interests in anduses for school-level financial data: schoolsand school districts; state and nationalpolicymakers; and researchers and policyanalysts. A fourth and more diverse groupof stakeholders consists of the public, rep-resented by parents and taxpayers of indi-vidual school districts, special interestgroups in education, and the financial com-munity. While members of this group aregenerally not directly involved in data col-lection or analysis, they are an important

This report focuses

on two approaches

to measurement of

resources in

education: an

accounting

approach and a

resource-based

approach.


83

audience for school-level information andhave strong interests in the results, althoughfrom different perspectives.

If successful, either or both of the ap-proaches to developing school-level finan-cial and resource data will provide moredetailed information on how schools spendtheir funds and allocate their resources.However, the critical issues are: what newdata are needed for what purposes; whichapproach or combination of approachescan best obtain the data; and if the ben-efits from such data collection, reporting,and analysis are worth the cost of obtain-ing them. As a framework for understand-ing the cycle of collection, reporting, anduse of school-level data, table 1 illustratesthe interrelationships among primarystakeholders that would be involved in theprocess.

To begin, it is useful to establish that theprimary focus of each group is somewhatdifferent. While at the fundamental level,all groups are certainly interested in im-proving schools, raising student achieve-ment, and efficient operations, they ap-proach these goals from different perspec-tives related to their positions and respon-sibilities.

School and district administrators have astheir primary responsibility the operationof schools. As such, they are interested ininformation that will allow the schools andthe district to function effectively and effi-ciently in compliance with state and fed-eral laws and regulations. From a fiscalperspective, they are concerned with cre-ating a feasible budget, monitoring expen-ditures to stay within the budgetedamounts, and developing fiscal reports thatinform them of the status of their opera-tions and allow them to report the resultsto a variety of audiences. Beyond day-to-day operations, administrators are also in-terested in improving the current operat-ing conditions of their schools, which canbe done by spending comparisons withsimilar schools and from research findings.

The role of researchers and policy analystsis to examine how schools organize theirresources and operate their various pro-grams to achieve student outcomes. Theirgoal is to understand what levels of re-sources organized in what schemes are ef-fective in achieving high level results fordifferent kinds of students. As such, theyare less interested in precise dollar amountsat the school level, which are distorted byvarious factors that make expenditure datanon-comparable across schools, districts,or states. Rather they are more concernedwith measuring resources that are used inthe educational process and how they arecombined into effective and ineffective pro-grams at the school level. The studies fromthis research effort provide insight to schooland district administrators on how to im-prove their programs and guidance for stateand national policymakers in setting edu-cational fiscal and program policy.

The final group, state and nationalpolicymakers, establish requirements andguidelines for data collection and report-ing for schools and districts. They need toconsider their own informational needs foraccurate and timely data and analyses toassist them in making data-driven decisions.Additionally, the rules that they establishimpact upon the other two groups in theform of mandates for specific data collec-tion and reporting, potential omission ofcertain useful data, and, most importantly,the availability of school-level data formanagement, comparison of spending andresource use, and fiscal and programmaticanalyses.

It is important, if not essential, to recog-nize the interests and needs of the multipleaudiences and users of school-level data.Given the different interests, a single ap-proach to school-level financial and re-source data collection and reporting maynot effectively serve the needs of the vari-ous groups. For example, an accountingsystem that provides for the collection ofschool-level expenditures may be the mostuseful approach for groups that requireactual expenditure data to manage opera-

The role of

researchers and

policy analysts is

to examine how

schools organize

their resources and

operate their

various programs

to achieve student

outcomes.


84

Table 1.—Data cycle for school-level data

Data cycleelements

School and districtadministrators

Researchers andpolicy analysts

State and nationalpolicymakers

Primary focus

Data collection

Data reporting

Data use

Data-drivendecisions

Operation of schools Research Policy

School improvement School improvement School improvement

Input for operationaldata needs and forcost and feasibiilty

of school-level data

collection

Input for mostappropriate data to

collect

State requirmentsand national

guidelines for datacollection

Primary collector ofschool-level

expenditure andresource data

Secondary datacollector, special

studies and surveys

School- and district-level expendituresand resource use

Compilation ofcomparative

statewide andnational data

Review of spendingand resource use in

education

Analyses of spendingand resource

allocation practices,equity, adequacy,

school-basedmanagement, and

accountabilityInput into legislativeand policy decisions

Management ofoperations

Comparisons ofspending patterns

Improvements toimplement in

inefficient schools

Dissemination ofresearch results

Programmatic andfunding legislation to

encourage changesto improve efficiency

and effectiveness

Changes in resourceallocation patterns to

improve studentoutcomes

�

� �

� �

�

�

�

�

�

�

�

�

�

�

�

�

��

��

�

�

SOURCE: Authors’ sketch.


85

tions (administrators), and to understandspending (taxpayers, parents, financial ana-lysts), as well as to make informed deci-sions in accordance with fiscal policy, regu-lations, and laws (school boards, state andnational policymakers). However, research-ers would find only expenditure data in-sufficient to conduct reliable studies abouteffective and efficient educational pro-grams; rather they would be more inter-ested in detailed school-level staffing data,and, perhaps, not even require actual ex-penditure data. Consequently, it is impor-tant to consider and balance the interestand needs of the different groups alongwith the value and cost of obtaining theinformation in the formulation of a school-level data collection and reporting system.

Schools and School Districts

The first major group of stakeholders rep-resents schools and school districts. Thisgroup includes school administrators, dis-trict administrators (primarily superinten-dents and school business officials), andschool boards. Others, such as school staffand parents, also have interests in school-level data. Principals and their staff, alongwith school business officials at the districtoffice, would be the main producers ofschool-level data. There is a division of la-bor in most school districts between schooland district office personnel and, depend-ing on the type of expenditure, the respon-sibility for making and recording the ex-penditure may fall to either level. For ex-ample, the human resources (personnel) orbusiness office staff at the district level usu-ally handle all personnel data (salaries andbenefits), while school level staff initiallyrecord many nonpersonnel instructionalexpenditures into the accounting systemunder rules and policies set by the centraladministration and school board. Supportexpenditures related directly and entirelyto a single school can be the responsibilityof the school, but activities or programsthat serve multiple schools or the entire dis-trict will be the province of the district of-fice.

In addition to data collection, the districthas the responsibility for maintaining fis-cal records and reporting fiscal data to theschool board, public, and state. At the dis-trict level, individual school expenditurescan be maintained separately as well asbeing aggregated and combined with dis-trict-level expenditures to arrive at total dis-trict expenditure amounts. Districts also arerequired to report expenditures to the statein a standardized format so that state re-ports may be produced. With school-levelfinancial data, the district would report notonly district totals, but also required schoolexpenditure amounts, again in accordancewith a state-established format.

As noted previously, a primary responsi-bility of school and school district admin-istrators is fiscal management and finan-cial compliance. They are required to main-tain individual school and overall districtexpenditures within the budget approvedby the school board and in accordance withstate and federal regulations and laws(Hartman 1999, 11–14). To this end, theywill look to school-level financial data asinput into developing their budget and as atool for monitoring expenditures during theyear.

Beyond fiscal management, an importantuse of school-level data for this groupwould be as a tool for school improvement.These data could serve as the basis for com-parisons of expenditure patterns among dif-ferent schools within their district. Thesecomparisons would seek to identify areasof differences and possible changes thatcould be made in their resource allocationdecisions to achieve improvements at theschool and district levels. In addition tothese internally generated comparisons, allschool and district stakeholders will be in-terested in state and national summary re-ports prepared by the respective educationalagencies that provide comparisons for mea-suring their expenditures against state andnational averages.

Also, research reports by state and nationalresearchers can identify good or best prac-

...a primary

responsibility of

school and school

district

administrators is

fiscal management

and financial

compliance.


86

tices in the manner in which effectiveschools allocate resources. Such informa-tion could be used to improve the effec-tiveness and efficiency of the programs inother school districts. However, beforecommitting significant funds, time, andenergy into program changes, a district willwant to be confident that the analyses arevalid and that the improvements can beeffectively implemented in their schools.

At the same time, a significant concern ofthis group is the burden of data collectionand reporting that they would face under aschool-level data system. The burden wouldconsist of the time required of various per-sonnel at the school and district level alongwith the expenses associated with chang-ing and maintaining new data systems. Thisis particularly true if the main immediatebeneficiaries of the data would be othergroups. Therefore, if substantial adminis-trative burden and expense are incurred todevelop and maintain a school-level datasystem, there will need to be compensatingbenefits of an immediate and concrete na-ture in order for districts to participate will-ingly and conscientiously in such an effort.Otherwise, such a burden will be viewedas another unfunded mandate without anyuseful benefit to the mission of the school.

The interests and concerns of school boardswould be similar to school administrators,but emphasizing their primary role of over-sight of fiscal and program operations.They would establish local policies forschool-level data collection and reportingsystems that would be followed by admin-istrators. A significant concern among thisgroup also would be the costs of establish-ing and maintaining a school-level datasystem. For their part, board memberscould use the more detailed school-level in-formation to provide closer scrutiny of theoperations in individual schools. This scru-tiny would include monitoring spending,reviewing individual school performance,comparing schools on key measures, andidentifying appropriate improvements inschool programs and operations. They alsowould have interest in research reports pro-viding analyses of effective programs and

practices that could be compared to theircurrent programs and practices as a meansto effect positive change.

Researchers and Policy Analysts

This group includes researchers and policyanalysts primarily at the state and nationallevels. Members of this group would be themajor consumers of school-level data.Their role is to inform the policymakersand school administrators and provide avalid basis for others to make resource al-location decisions. With more extensiveschool-level information they could con-duct a variety of analyses that would im-prove the knowledge base of effective andefficient school programs and practices.Examples would be analyses of spendingpatterns among schools, relationship ofresource allocation practices and studentoutcomes, studies of equity at the schoollevel, adequacy of resources for achievingdesired educational outcomes, school-based management, and accountability.Table 2 provides a more extensive list ofpolicy issues that can be examined withgreater thoroughness using school-leveldata. With these analyses this group be-comes producers of reports that policymak-ers and school district personnel can uti-lize to improve their resource allocationdecisions.

To fulfill their role, researchers need ap-propriate data to analyze school-level re-source allocation patterns and student out-comes. Such data include expenditures,staff assignments, and student involvementin various programs and services, alongwith student outcomes. Consequently, thedata needs of researchers and policy ana-lysts may go beyond the immediate needsof school administrators, the group thatmust produce school-level data.

Researchers are also concerned with thecosts of data collection, but for them it isthe cost of getting data from schools anddistricts that is more important, not thecosts of the school data systems. If dataare already available at the school-level,

To fulfill their role,

researchers need

appropriate data

to analyze school-

level resource

allocation patterns

and student

outcomes.


87

then it is much simpler, less expensive, andprobably more accurate to obtain themdirectly from an existing school-level datasystem than if a special one-time collectionis required.

A second issue in this area is comparabil-ity of data. To be able to make meaningfulanalyses beyond a single school or district,it is necessary that data be collected andreported in a comparable fashion for allentities. This involves some degree of stan-dardization of definitions, measures, andpractices across a state or even nationally.Without a common agreement on suchquestions as what constitutes a school,

treatment of personnel benefits, or alloca-tion of personnel that serve multiple build-ings or programs, then it becomes prob-lematic to develop consistent and validanalyses. These last two points highlightthe need to have the input of researchersinto the design of any state or nationalschool-level data system to balance theirneeds with those of school personnel whowill be operating the system.

State and Federal Policymakers

This last group represents a diverse collec-tion of different constituencies with thecommon interest of understanding and im-

Table 2.—Policy issues related to school-level financial and resource data

Resource allocation and productivity

� Basic information on level of spending at school level

�Distribution of resources among different functions, objects, programs, and school/district activities

� Relationships among choice, quantity, and utilization of resources with studentoutcomes

� Analyses to inform local, state and federal policy makers for funding and resourceallocation decisions

Costs and effects of policy initiatives

� How do school reform proposals affect school-level staffing and costs?

Equity

� Are resources distributed in an equitable manner across schools within a district orwithin a state?

� Variations in per pupil expenditures among schools and the relationship with wealthand other variables

Adequacy

� Do all schools have the minimum level of resources to provide an adequate level ofeducational services for learning?

� What differential levels of resources are necessary for different student populations?School-based management

� Financial and resource allocation data to support school-level decision making

� Benchmarking information for high-performing schools to serve as models for othersAccountability

� Are resources being spent as intended?

�Are schools achieving intended outcomes for their expenditures?

Legislative and congressional interests and public inquiries

� Special requests for spending information at the school level or for special purposes

� Topical analyses for specific topics

SOURCE: Issacs, J.B., Garet, M.S., Sherman, J.D., Cullen, A., and Phelps, R. 1999. Collection of Resource and ExpenditureData on the Schools and Staffing Survey. Washington, DC: U.S. Department of Education, National Center forEducational Statistics. (Working Paper 1999–07).


88

proving education. They include statepolicymakers and staff (Legislators, legis-lative staff, Governors and staff, state De-partments of Education) and federalpolicymakers and staff (Legislators, legis-lative staff, President and staff, U.S. De-partment of Education). They are the pri-mary consumers of information and reportsproduced by researchers and policy ana-lysts. They use the information from analy-ses for a variety of purposes: to establishfunding levels; to develop resource alloca-tion legislation, policies, and guidelines; tocreate new initiatives; to encourage effec-tive learning approaches; and to answertopical questions on school spending andproductivity. New school-level data andanalyses would provide greater informa-tion for these decisions. Of particular ben-efit is that attention could be focused onschool and program levels, or even gradelevels and academic subjects. With moredetailed information, decisions would beless reliant on aggregated district-level datathat can conceal disparate operations andconditions among individual schools andwithin schools.

In addition, school-level data would allowstate and federal agencies to monitor theachievements of schools in closer detail.Rather than being restricted to district-levelmeasures and district-to-district compari-sons to measure equity, school-level datawould provide the opportunity for intra-district analyses and comparisons of simi-lar schools across a state or nationally. Suchschool-level cost data when combined withstudent outcome measures may assist indetermining what funding and resourceallocation practices work best for variouskinds of students. These analyses can iden-tify best practices for instructional deliv-ery and enable examinations of account-ability. The inquiries and comparisons canthen feed into legislation and policies thatdirect and redirect resources for education.

Finally, these groups establish the require-ments and guidelines for data productionby schools and districts. State legislatures

and departments of education generally de-termine educational accounting systems,data reporting obligations, and other datacollection practices. However, the federalgovernment, through its financial account-ing handbook series (Fowler 1997) thatgives a suggested chart of accounts and ac-counting procedures for states, has influ-enced states’ practices and encouraged ageneral uniformity among state educationalaccounting systems.

Policymakers are also mindful of the bur-dens that their data requirements imposeon school districts. While new and im-proved data for program and funding de-cisions are useful, they may not be man-dated if they are perceived as too expen-sive or detailed.

School-level Data

School-level data of interest to the majorstakeholder groups are of three main types:actual expenditures and costs,3 staff, andstudents. Schools often have different sys-tems for collecting and reporting each type,but frequently the different systems are notintegrated with each other nor even use acommon basis for categorizing their dataelements.

Expenditure Data

District expenditures can be divided intofour types, depending on where they areincurred and for what purposes. These dis-tinctions affect both who is involved in thedata collection and what types of data areavailable.

Direct expenditures identifiable with an individualschool

These are the most straightforward typesof expenditures for school-level data re-porting. They are completely and unam-biguously identified with a single school.Examples include salaries and benefits forstaff assigned full-time to a particular

3 Expenditures are the dollar amounts spent for activities, while costs represent the resources consumed by those activities.

...school-level data

would allow state

and federal

agencies to

monitor the

achievements of

schools in closer

detail.


89

school (teachers, other professional staff,administrators, and classified staff), class-room supplies, and computer hardwareand software purchased by the school. Aslong as the accounting system has an indi-vidual school code attached to these ex-penditures, they are associated with an in-dividual school.

Joint expenditures made at the school level formore than one school

Expenditures of this type are for person-nel or nonpersonnel objects that serve sev-eral schools. The primary example of jointexpenditures is salaries and benefits forinstructional or support personnel servingmultiple schools, such as speech therapists,school nurses, or librarians. Whether thesecosts are shown for each school, for a singleschool, or in some central office accountdepends on district practice. If staff assign-ments are for full-time equivalent (FTE)positions for each school, then the account-ing system can allocate personnel expen-ditures to each school based on its share ofthe FTE for each person. If this is not thecase, then these expenditures could be al-located to individual schools on some rea-sonable basis (e.g., percentage of time spentin each school, percentage of studentsserved in each school) and should reflectthe proportion of these resources devotedto the school.

Expenditures made centrally for services identifi-able by school

These expenditures are similar to the pre-vious category in that they are for pro-grams, objects, or services for students thatcan be identified by an individual schoolor groups of schools. However, they arecontrolled by and made at the central of-fice rather than the school level. The stu-dent specific expenditures could be allo-cated to individual schools based on thenumber of children or percentage of stu-dents in each school (e.g., medical services).Joint expenditures for multiple schoolscould be allocated to individual schools ona reasonable basis reflecting their portionof resources devoted to the school (e.g.,costs of elementary curriculum coordina-

tors divided among elementary schoolsbased on the number of students in eachschool or the number of elementaryschools). Expenditures that can be associ-ated with specific schools can be assigneddirectly to each school (e.g., utilities, if con-sumption can be broken out by school).

Central office expenditures made for the entire dis-trict

Some of the expenditures cannot be directlyassociated with an individual school orgroup of schools. These are the district-levelexpenditures for programs, functions, andservices for the entire district. Examples in-clude the personnel and nonpersonnel ex-penditures for the superintendent’s office,business office, and personnel office. Theseexpenditures could be allocated to indi-vidual schools on a reasonable basis (e.g.,percentage of students or personnel in eachschool). However, the reported expendi-tures would be arbitrary and represent onlyan accounting allocation to distribute cen-tralized expenditures to the school level;they would not necessarily be related toschool activities or controlled by school per-sonnel.

Staff Data

The availability of staffing data and theability to track staff information to theschool level will vary by state and district.Data on individual staff are generally main-tained in personnel records at the districtlevel. For each person they could include:

� General type of position (professional,administrative, and classified)

� Specific assignment (regular teacher,special education teacher, counselor,librarian, vice principal, secretary,custodian, business administrator, etc.)and time assigned to position (full-timeor full-time equivalent)

� Location of assignment (specific school,multiple schools, central office,transportation office, etc.)

The availability of

staffing data and

the ability to track

staff information

to the school level

will vary by state

and district.


90

� Professional qualifications (educationbackground, academic degrees andinstitutions granting them, certificationsheld, date hired, and time in service)

� Current salary and benefits

These staff data can be aggregated in a va-riety of ways to produce personnel statis-tics for the district and, in some cases, forindividual schools. For example, the num-ber of staff, by type can be summed foreach school and for the district so that thenumber of administrative, teaching, otherprofessional, and classified staff in a schoolcan be determined.

Some district information systems keepmore specific data by individual. For ex-ample, if the district currently maintainsand reports information by grade level oracademic department, then it would bepossible to identify staff resources by gradelevel or subjects taught (e.g., mathematics,English, science, etc.) within each school.Likewise, for staff in split assignments, ifthe district records the time allocated toeach assignment or FTE, then more accu-rate data on staff resources committed toindividual schools and programs can be ob-tained. However, staff information systemsthat track how various personnel allocatetheir time among different programs orservices or students are not consistentacross districts.

Student Data

Student data systems are designed to reportnumbers of students along a variety of di-mensions. The focus is on numbers of stu-dents, rather than on how they spend theirtime in schools. Enrollment data will givecounts of the numbers of students in thedistrict and in each school and, perhaps,by grade level or subject matter. The mea-sure will generally be established by statereporting requirements (e.g., average dailymembership, and average daily attendance).

Further breakdowns, by type of student(regular, special, vocational, etc.) in the dis-trict and in each school are generally pos-sible. However, beyond the general desig-nation of the type of instruction for the stu-dents, little more detailed information isusually collected about how student timeis allocated or spent among various aca-demic programs, specific courses, supportservices, etc., unless required by state re-porting or funding requirements.

Performance measures for students are re-ceiving greater attention as the demandincreases for higher achievement and moreaccountability. Some states have testingprograms where standardized tests are ad-ministered at various grade levels and theresults are reported by school.4 However,these are usually for a limited number ofgrades or students in a school or district.National testing results, such as collegeentrance examinations (SAT or ACT) oradvanced placement tests, can also be re-ported by district and school, althoughthese are for graduating high school stu-dents and not all students take these tests.Other performance data tend to be devel-oped and reported by individual schoolsor districts. As a result, the most commonperformance measures are specific to indi-vidual schools or districts and therefore notgeneralizable or even uniformly available.The measures often focus primarily on par-ticular factors in schools or districts thatillustrate or emphasize positive results.

Current Data ReportingSituation

Expenditures

At present, educational expenditure dataare available at the district level for allstates. The general format is based onguidelines provided by Fowler (1997).While they may vary in the details, stateschool accounting systems generally followthe fund-function-object-program organi-

4 See for example, the Pennsylvania State System of Assessment program that tests students in grades 5, 8, and 11 in the areas of reading,mathematics, and writing.

At present,

educational

expenditure data

are available at the

district level for all

states.


91

zation, in which expenditure data are cat-egorized by the fund (collection of ac-counts), the function (purpose), the objectof the expenditure (item), and the activity(program). An overview of the primaryfunction-object-program categories isshown in table 3. A complete list of funds,functions, objects, and programs for schooldistricts can be found in the federal finan-cial accounting handbook.

For the function dimension, the primarycategories represent the major areas of ac-tivity of a school district: instruction; sup-port services; noninstructional services; fa-cilities acquisition and construction ser-vices; and other uses. Each of these broadfunctions is further divided intosubfunctions to provide greater specificity;

in some cases, there may be four levels ofdetail within a single function.

The unfortunate exception to the detailedsubfunction categorization is instruction.The federal handbook lists no subfunctionsunder instruction; rather it uses anotherdimension, program, to differentiate typesof instructional or other related programs.The focus of the program dimension is onvarious instructional programs (e.g., regu-lar programs, special programs, and voca-tional programs), although othernoninstructional activities are also included.In practice, many states combine the func-tion and program dimensions into a singledimension so that expenditures can be clas-sified into different instructional programs.

Table 3.—Primary categories of current financial accounting system for school districts

Function Object ProgramInstruction Salaries Regular programs

Support services Employee benefits Special programs

� Support services—students Professional and technical services Vocational programs

Attendance and social work Property services Other instructional programs

Speech pathology and audiology Other purchased services Nonpublic school programs

Guidance Property Community services

Health Supplies Enterprise programs

Psychological Other objects

� Support services—instructional staff Other uses of funds

Improvement of instruction

Educational media

� Support services—general administration

Board of Education

Executive administration

� Support services—school administration

� Support services—business

� Operation and maintenance of plant

� Student transportation

� Support services—central

Operation of noninstructional services

� Food service

� Other enterprise

� Community services

Facilities acquisition and construction

Other uses

� Debt service

SOURCE: Fowler, W.J. 1997. Financial Accounting for Local and State School Systems, 1990. Washington, DC: U.S.Department of Education, National Center for Education Statistics (NCES 97–096R).


92

Additionally, it is possible to use the sameaccount code dimension to include the levelof education to differentiate among elemen-tary, middle, and high school expendituresin each of the instructional programs (e.g.,instruction, regular, and high school). Forthe new federal financial accounting hand-book, it is recommended that instructionalsubfunctions be added; they should corre-spond to the major subdivisions of instruc-tional programs—regular programs, spe-cial programs, vocational programs, andother instructional programs.

Objects of expenditure are the items forwhich expenditures are made; they includesalaries, benefits, purchased services, sup-plies, property, other objects, and other usesof funds. By coding the function, object,and program dimensions for a single ex-penditure, the accounting system providesa more detailed record. For example, thespecific expenditures for salaries of specialeducation instructional personnel (perhapsseparated into professional and classifiedstaff), benefits for staff in the principal’soffice, or supplies for the guidance programcan be specified through the accountingcode structure.

Most school accounting systems also con-tain a dimension to record the operationalunit of an expenditure. This feature pro-vides the opportunity to identify expendi-ture data by school building or any othercost center that a district may designate.This makes it possible to provide detailedexpenditure data by school when combinedwith the function/object/program report-ing, at least for direct school expenditures.

However, operational unit is currently anoptional classification in the federal hand-book and in most state education account-ing systems, and is not reported by mostschool systems. Consequently, most schoolaccounting systems do not collect andrecord comprehensive expenditure data atthe school level, particularly for joint orcentralized expenditures that require an al-location procedure. For example, school ac-counting systems can report the expendi-tures for the combined computer hardware

purchases in the district, but except for in-ternal cost control purposes, may not regu-larly report these expenditures for an indi-vidual school. Further, it is not possible toaggregate such expenditures to the statelevel since school-level data are generallynot available nor do standardized report-ing protocols exist for districts to use.

Staff and Student Data

Accounting systems, by their nature and de-sign, focus on expenditures and revenuesof a school system. Consequently, studentand staff data are not available from ac-counting systems. These data are generallymaintained in separate systems with report-ing requirements and categories that do notalways align with the reported expendi-tures. A further shortfall in student infor-mation is the lack of consistent and reli-able performance data to measure studentoutcomes.

As a result, the independent accounting,personnel, and student information systemsdo not produce or report adequate data forany of the major stakeholder groups.School-level expenditure information is notreadily or completely available. Likewise,neither school-level information on studentactivity and performance nor staffing in-formation, which provides resource alloca-tion data at the school level, are routinelycollected by school districts in a uniformmanner. Average district costs, which arecurrently reported, mask variations amongschools and cannot be linked with studentactivity and student performance data orstaff data from individual schools. Unfor-tunately, these are just the kind of data thatare needed to analyze and assess school re-form efforts.

Accounting Approach

The essence of the accounting approach isthat it is a downward accounting extension(DAE) of the current district-level approachto record and report expenditures at thelevel of individual schools. This approachwould build on the existing accounting sys-

Accounting

systems, by their

nature and design,

focus on

expenditures and

revenues of a

school system.


93

tem and expand it by adding new accountsat the school level for every school to recordexpenditures at the building level. This ex-tension would be facilitated with the usethe current optional dimension for opera-tional unit to identify expenditures byschool. In fact, many school systems al-ready identify some expenditures by schoolin their internal accounting system, even ifthey are not required by the state or otheragencies. In short, this is an approach fa-miliar to school administrators and “Be-cause of the strong congruence in the re-quirements of a district-level and a school-level financial accounting system, it shouldbe possible for most states to . . . imple-ment school-level [expenditure data] col-lections in all school districts andschools” (Sherman, Best, and Luskin 1996,xvii).

Currently, 19 states report that they alreadyrequire school-level expenditure report-ing.5 For example, Pennsylvania school dis-tricts are required to report education costs,including expenditures by each school forclassroom instruction, instructional studentsupport, and facilities and plant manage-ment costs. Additionally, they also have toreport expenditures for special education,non-instructional student support, profes-sional development, and technology.6 Thistrend is anticipated to grow and intensifywith additional states requiring school-levelexpenditure data in the coming years.

There are many reasons for states to man-date school-level financial reporting.School costs are rising rapidly and moremoney is being requested at both the localand state levels to fund education. There isgreater demand for accountability that istied to the increased funding from legisla-tors, governors, school boards, parents,and taxpayers. No longer is an aggregateddistrict-level expenditure per studentamount sufficient. Rather there is an inter-est in examining down to the school level,where the funds are actually spent, how

much money is being spent on educationand if the spending is equitable, adequate,efficient, and productive. School adminis-trators also have need of actual expendi-ture data in their management and improve-ment of school operations. School-level ex-penditure data form the basis for budgetdevelopment, which is the primary resourcereallocation process in school districts. Ac-tual expenditure data are necessary to ful-fill the fiduciary responsibilities of admin-istrators and school boards to ensure thatthe funds are both legally and effectivelyused. Detailed information is also useful forthe financial community in evaluating thefiscal soundness of school districts for creditand bond ratings.

Data collected and reported by the DAEwould be the actual expenditures associatedwith individual schools. Utilizing the cur-rent financial accounting system approachwith the operational unit dimension, expen-ditures would be assigned an accountingcode to indicate the fund, function, object,program, and school for the expenditure.If greater detail were desired, other cur-rently optional expenditure dimensions areavailable that could identify the level of in-struction (elementary, middle, secondary, orindividual grade) and the subject matter(e.g., English language arts, mathematics,natural sciences, and social sciences). Thesedata would be collected at the school anddistrict office and entered into the district’scomputerized accounting system. Fromthere, sorting the basic expenditure entriesby one or more of the coded dimensionscould generate any number of reports. Ex-amples of these types of reports are illus-trated in table 4. In the first example, allexpenditures in the school are specified byinstructional program or function. In thesecond example, the costs of special educa-tion programs in the school are further re-ported by object of expenditure. The thirdexample presents school expenditures forsupplies by subject matter.

5 Based on a survey by the authors that asked states to self-report school-level expenditure reporting requirements and collected theirannual financial report documents.

6 PA Senate Bill 652, Section 613. (2000).

Currently, 19

states report that

they already

require school-

level expenditure

reporting.


94

The accounting approach to obtainingschool-level data would focus on expendi-tures and would not expand to include stu-dent or staff data. That integration wouldhave to happen outside the accounting sys-tem.

Resource Cost Model Approach

Description

The RCM uses an economic basis for es-tablishing costs of educational programs.

Table 4.—Examples of school reports from downward accounting extension data

School: Alva High School, Code: 81

1. Expenditures by instructional program and function

Code Function/description Amount

1100 Regular programs $2,310,2001200 Special programs 362,0001300 Vocational programs 384,8001400 Other instructional programs 36,9002120 Guidance services 10,5002130 Health services 23,4502140 Psychological services 12,3502210 Improvement of instructional services 5,6002220 Educational media services 74,2002410 Principal's office 147,6002610 Building services operations 98,750

Total $3,466,350

2. Expenditures for special education program, by object

Code Object/description Amount

100 Salaries $243,800200 Benefits 85,300300 Purchased professional services 7,500400 Purchased property services 0500 Other purchased services 4,400600 Supplies 14,200700 Property 5,600800 Other objects 1,200

Total $362,000

3. Expenditures of supplies, by selected subject matter areas

Code Subject/description Amount

02 Art $1,54008 Physical education 51011 Mathematics 6,30012 Music 67518 General education 26,205

Total $35,230

SOURCE: Authors‘ sketch.


95

It is based on measurement of physical re-sources employed in an activity, as opposedto the determination of actual expendi-tures. It utilizes a formal methodology oftransforming physical resources into appro-priate costs. Four steps are common to itsapplication (Issacs et al. 1999, II–1–5).

1. Specifying the structure of the servicedelivery systems and the types ofphysical ingredients (e.g., teachers,books, etc.) used in delivering services.

2. Measuring the intensity of theseresources by quantifying them.

3. Assigning prices to the specificphysical ingredients.

4. Using the price data to aggregateresources across the entire programto determine overall program costs.

As defined by Chambers (1999, 19), “Aservice delivery system is a collection orcombination of resources (i.e., inputs) thatis specifically organized to provide a cer-tain service to a target population or stu-dents or clients.” Examples given for in-structional service delivery systems are aself-contained classroom for elementarygrades 1–3, and a language-arts pull-outprogram for disadvantaged students. Otherservice delivery systems at the school levelcould include: instructional support activi-ties, such as the school library, guidancecounselor, or speech pathologist; adminis-trative activities, such as the principal’s of-fice; and operational support activities,such as custodians.

The physical ingredients that comprise aservice delivery system are those resourcesthat are necessary to carry out its activi-ties. For example, in an elementary class-room, those ingredients could include theteacher, a part-time aide, associated ben-efits, supplies, equipment, classroom space,and utilities. In a comprehensive RCM, allof the resources utilized by the service de-livery system would be included. However,“because personnel represent the predomi-

nant resource in a social service enterpriselike education and personnel can be readilymeasured in terms of some measurablephysical quantities (Chambers 1999, 51),”the focus of the RCM is frequently on per-sonnel resources to the exclusion of theother nonstaff resources.

An additional element of the service deliv-ery system is its capacity. For instructionalsystems, this is typically specified in termsof number of students (class size orworkload) that can be served by one unit.Individual service delivery units can be com-bined into larger service delivery systems.An example of this is in table 5, where thepersonnel resources of all classroom units,support units, and administrative units thatfunction in a single location are collectedinto a single school system.

The intensity for personnel resources ismeasured in terms of quantity and can takemany forms. In table 5, the unit of mea-surement is full-time and part-time posi-tions. Other measures may be used in dif-ferent school districts, such as FTE posi-tions, number of days, or number of hours.For example, rather than assuming all part-time personnel are allocated half-time totheir assignment as is done in the example,more precise FTE amounts could be uti-lized, such as 0.4 FTE for a nurse assignedto the school for two days per week (2 days/5 days = 0.4 FTE).

In order to measure the resources in mon-etary terms, it is necessary to translate thephysical quantities into dollar amounts. Inthe RCM, this is done by assigning pricesto each resource and then multiplying thequantity of each resource by its associatedprice. However, a potential difficulty arisesat this point with the choice of a price toassign to personnel resources. The two al-ternatives are the actual price (salary andbenefits) of the specific individuals whodeliver the services or a standardized pricefor all similar positions. Each choice hasits advantages and disadvantages, particu-larly when using the resultant cost data tocompare different schools or programs.

In order to

measure the

resources in

monetary terms, it

is necessary to

translate the

physical quantities

into dollar

amounts.


96

Table 5.—Illustration of resource cost model: Staff resources at Rosemont School,1 byphysical ingredients, quantities, prices, and total costs

Quantity

Physical ingredients Full-time Part-time2 Price per unit3 Total cost

Classroom teachers 15 0 $48,000 $720,000

Music/art teacher 0 2 48,000 48,000

PE teacher 1 1 48,000 72,000

Special Education teacher 1 0 48,000 48,000

Principal 1 0 75,000 75,000

Vice Principal 1 0 62,000 62,000

Librarian 0 1 47,000 23,500

Counselor 1 0 54,000 54,000

Nurse 0 1 39,000 19,500

Social Worker 0 1 50,000 25,000

Psychologist 0 1 60,000 30,000

Speech Pathologist 0 1 52,000 26,000

Library aide 0 1 24,000 12,000

Health aide 1 0 22,000 22,000

Special Education aides 2 0 21,000 42,000

Bilingual/ESL aides 3 2 21,000 84,000

Other teacher aides 3 2 21,000 84,000

Secretaries 1 1 28,000 42,000

Food service workers 0 2 19,000 19,000

Custodians 2 0 26,000 52,000

Total 32 16 $1,560,00

1 Rosemont is a hypothetical elementary school of 400 students.

2 Each part-time person is assumed to work half-time in the school.

3 Prices are based on national staff salary averages, increased by a 28 percent fringe benefit rate.

SOURCE: Issacs, J.B., Garet, M.S., Sherman, J.D., Cullen, A., and Phelps, R. 1999. Collection of Resource and ExpenditureData on the Schools and Staffing Survey. Washington, DC: U.S. Department of Education, National Center forEducation Statistics. (Working Paper 1999–07).

The last step is to total the costs of theservice delivery system by summing thecosts of the individual resources in the sys-tem. The result is the cost (actual or stan-dard, depending on the types of prices used)of the service delivery system. The criticalaspect of the RCM is that the costs are builtup from the service delivery system(s) thatmake up the school and include the par-ticular resources involved, their quantitiesand prices. This makes it possible to com-pare schools and programs along the di-mensions that make a difference in bothcosts and student outcomes. “To under-stand the factors that affect variations inthe costs of services requires an accuratedescription of how resources are combined,

allocated, and utilized to provide those ser-vices” (Chambers 1999, 22).

Data Needs

Implementation of the RCM will requirecollection of data that identify resource useat the school level and in direct support ofthe school-level activities. The basic unitof analysis is that of the service deliverysystem, so resources must be reported bythat unit. In the comprehensive implemen-tation of the RCM approach, all of the re-sources contributing to the delivery of aparticular service, both personnel andnonpersonnel, are combined. All services


97

are then consolidated to establish the fullresource consumption of the unit and thenthe individual units are further combinedinto school-level costs. However, in prac-tice, data collection may be limited to per-sonnel time only since these resources con-stitute the bulk of educational costs.Nonpersonnel resource data, if included,can be collected for the physical quantitiesutilized (e.g., types of supplies, textbooks,and computers), but it is also possible tosubstitute the expenditure amounts (actualor average) for nonpersonnel resources asa shortcut measure to resource consump-tion.

The RCM approach to obtaining school-level data would concentrate on a detailedspecification of staff time. Student datawould need to be collected by the same cat-egories as resources. These include bothstudent enrollment data by program andservice along with student outcome mea-sures. However, the student system wouldbe a separate data system and not neces-sarily part of the RCM approach.

Areas of Common Concern

Both the accounting approach and theRCM approach face common issues thatneed to be resolved regardless of which sys-tem would be utilized. These areas affecteach approach in similar ways.

Definition of a School

The first issue is the definition of what con-stitutes a school. This prerequisite, regard-less of which approach is used, necessitatesa “clear definition of what constitutes a‘school’ to which financial activities shouldbe assigned” (Sherman, Best, and Luskin1996, 18). A definition for a public schoolis available from NCES:

An institution which provides edu-cational services and has the fol-lowing characteristics (Young1999):

� Has one or more grade groups(prekindergarten through grade 12) oris ungraded;

� Has one or more teachers to giveinstruction;

� Is located in one or more buildings orsites;

� Has an assigned administrator;

� Receives public funds as primarysupport;

� Is operated by an education agency.

However, for consistency and uniformityin data collection and reporting it is neces-sary to go beyond this general definitionand establish decision rules that cover allsituations, such as multiple buildings on asingle site, multiple programs in a singlebuilding, and special centers.

Expenditures to Include

This leads to another difficult issue—whatexpenditures or costs to include at theschool level. The range of expenditures in-cludes: direct expenditures made at theschool site (e.g., salaries, benefits, and sup-plies); joint expenditures made for severalschools; expenditures made at the districtlevel identifiable with individual schools;and all district expenditures, even those thathave to be arbitrarily allocated to the schoollevel. In particular, the question arises asto whether it is appropriate or useful toallocate district office expenditures that areunrelated to specific school activities, suchas superintendent’s office, business office,and debt service, to the school level. Onone hand, when such expenditures are notallocated, the result will be an incompletepicture; on the other hand, an arbitrary al-location is simply an accounting calcula-tion for expenditures over which a schoolcannot exercise control and should not beheld accountable. An intermediate positionis to allocate only those expenditures logi-cally and reasonably identifiable to a spe-cific school. In this approach, district-wideexpenditures at the central or district levelwould not be allocated to schools. The

Both the

accounting

approach and the

RCM approach

face common

issues that need to

be resolved

regardless of which

system would be

utilized.


98

RCM approach faces the same choices, ex-cept that the allocation is based on person-nel time and standardized costs instead ofactual expenditures.

Allocation Procedures

In order to allocate some or all nondirectexpenditures to individual schools, alloca-tion procedures would need to be clear, uni-form, and utilize readily available data.There are relatively few bases for alloca-tion that meet these criteria; they includenumber of students, number of staff, stafftime, square footage of the building, and,perhaps physical quantities of selectednonpersonnel resources, such as comput-ers. Once again, the RCM approach willrequire procedures for any allocations ofnonschool resources to individual schools.

Level of Data Collection and Reporting

Another critical issue to resolve is the levelto which data collection and reporting sys-tems should go and amount of detail to in-clude in school-level records. Clearly, aschool-level data system needs to report in-formation at the individual school level. Thequestion though is whether lower levels ofreporting are reasonable and practical. Thisdecision represents a trade-off between thebenefits of increased levels of detail and thecosts of collecting such data.

The level of detail should be establishedbased on the type of program analyses thatare desired. The primary options are illus-trated in table 6; they include the majorfunctions, various programs and furthersubdivisions within instruction, andsubfunctions for the support andnoninstructional areas. Under the RCM ap-proach, similar categories for data collec-tion and reporting of resources would alsobe needed.

The existing federal and state financial ac-counting systems already have provisionsfor several of these levels of reporting(Fowler 1997, 24–26, 33–34). The programdimension is a currently required reporting

level that permits coding of expendituresby the type of program. These are speci-fied as: regular programs; special programs,including mentally retarded, physicallyhandicapped, emotionally disturbed, learn-ing disabled, culturally deprived, bilingual,and gifted and talented; vocational pro-grams; school-sponsored cocurricular ac-tivities; and school-sponsored athletics.Other classifications in the present schemeare optional and not utilized by all statesand districts. The operational unit dimen-sion can be used “to designate a budgetaryor cost center;” or “to segregate costs bybuilding structure” (Fowler 1997, 33). Thelevel of instruction dimension permits clas-sification of expenditures by grade level orgrade grouping (e.g., elementary, middle,secondary). The subject matter dimensionallows organization of expenditures by sub-ject area (e.g., English language arts, math-ematics, natural sciences, and social sci-ences). The support services expendituresat the school level for students, instruc-tional staff, and operations can be gath-ered through the current subfunctions thatcategorize these activities.

Student Data

Regardless of which approach is used tocollect school-level financial and resourcedata, any analyses will require student datato allow calculation of per pupil expendi-ture amounts. The primary requirement isthat the student data be collected in com-patible categories with the financial data.Information on the numbers of students atthe school is needed, as well as the num-bers by the subcategories used in the finan-cial data system. That is, if expendituresby instructional program are collected, thenthe number of students served by each ofthe instructional program categories alsoneeds to be collected. For example, thenumber of students in special educationneeds to be collected and reported in orderto determine the costs per pupil of that pro-gram.

The same is true of data reporting studentresults. In order to conduct analyses foreither the accounting approach or the RCM

Regardless of

which approach is

used to collect

school-level

financial and

resource data, any

analyses will

require student

data to allow

calculation of per

pupil expenditure

amounts.


99

� �

��

��

��

� �

Table 6.—Dimensions of school-level expenditure reporting

Major function

Instruction

Support services

Noninstructional services (e.g., Food service)

Type of instructional program

Regular education

Special education

Compensatory education

Vocational education

Grade level

Individual grade

Elementary

Middle/junior high school

High school

Secondary

Subject matter or discipline

Multiple options representing courses or groups of courses

Classroom

Individual teacher

Support services—students

Attendance and social work

Guidance

Pupil health

Psychological services

Speech pathology and audiology

Support services—instructional staff


Educational media

Operational support

Operation and maintenance of buildings, grounds, and equipment

Administration

School level


� �

� �

� �

� �


100

approach, student outcome measures needto be collected in a compatible format withthe expenditure or resource utilization data.

Areas of Difference Between theTwo Approaches

The accounting approach and the RCM ap-proach also have differences in the way thatthey treat or deal with important issues incollecting and reporting school-level data.These are discussed below and summarizedin table 7.

Focus of Approach

Accounting approach

The accounting approach is concerned withactual expenditures. These are preciseamounts based on the district’s financialaccounting records. All amounts are mea-sured in dollars. The use of familiar and“true” expenditure amounts lends credibil-ity to the results for school and district levelpersonnel as well as state and nationalpolicymakers.

RCM approach

The emphasis of the RCM approach is onresource consumption, rather than actualexpenditures. Consequently, the measuresfocus on physical quantities of resources.It is possible to include all personnel andnonpersonnel resources in the data collec-tion, although limiting the data to person-nel resources captures the most importantand the majority of the resources used inthe educational process. The tradeoffs in-volve the loss of potentially important re-source information, such as technology re-sources used in an instructional program,compared with the additional effort andcost to collect all resource elements.

In order to monetize the resource quanti-ties in the RCM approach, standard pricesfor each resource are applied to calculate astandard cost for the resources consumed.The RCM example shown previously intable 5 illustrates this procedure. In theexample, there are 15 classroom teachers

in the school. That quantity is multipliedby a standard salary amount of $48,000to calculate a standard salary cost of class-room teachers (15 x $48,000 = $720,000).By contrast, the accounting approachwould sum the actual salary amounts foreach teacher (likely ranging from $30,000to $60,000) to determine the actual salaryexpenditures for classroom teachers (forexample, $795,000 if the actual salaries arehigher than the average). As a result, thecosts reported for teachers under the RCMprocedure will be different from the actualexpenditure. As this example illustrates, theresulting cost information may not makesense nor seem useful or familiar to schooland district personnel since standardizedsalaries, rather than actual salaries, are usedin reporting. Further, the cost data reportedby the RCM will not correspond to the ex-penditures for the same school reported bythe accounting system, which could causecredibility problems for the RCM amongschool personnel and possibly lead to re-jection of the “theoretical” costs.

Unit of Analysis

Since the school-level data are the primaryconcern, the school is the major unit ofanalysis for both the accounting and theRCM approach. However, the two ap-proaches get to the school level from dif-ferent directions.

Accounting approach

The accounting approach starts at theschool level with its data collection proce-dures and accounts. If additional detail isdesired, there can be a further breakdownto collect and report information by func-tion, program, grade, or subject matterwithin the school.

RCM approach

By contrast, the RCM approach is muchmore of a bottom-up effort where the pri-mary unit of analysis is the service deliv-ery system (which corresponds to the pro-gram level in the accounting system). Theindividual service delivery systems that op-

Since the school-

level data are the

primary concern,

the school is the

major unit of

analysis for both

the accounting and

the RCM

approach.


101

Table 7.—Differences of approaches for school-level dataIssues Accounting approach Resource cost model approach

Focus of approach Actual expenditures Consumption of resources

Understandable, believable Use of actual expenditure data Use of standard salary and benefitexpenditure data corresponds to actual amounts and data to develop expenditures causes

enhances face validity of reported data difference between actualexpenditures and reported results;can cause confusion and rejection ofresults

Unit of analysis School, with lower levels of detail possible Service delivery system withthrough subfunctions and programs aggregation of units up to school

level

Type of data required Actual expenditures, by accounting Physical resources utilized; cancategories: function; object; and program include all personnel and

nonpersonnel items, but may belimited to personnel only

Information collected and reported Dollars spent, by accounting categories: Use of personnel and otherfunction; object; and program resources, by service delivery system;

resource combinations utilized

Approach to data collection Existing accounting system procedures Additional personnel reporting toextended to school level obtain staff time allocations among

service delivery systems

Timing of data collection Ongoing as expenditures occur Periodic on specified dates orongoing withindividual time reporting

Difficulty of implementation Use of existing district-level accounting New personnel data system cansystem should make implementation make implementation more difficulteasier

Startup costs for data collection Significant startup effort to create new Significant startup effort to createschool- level accounts and procedures new school-level procedures for

personnel time and nonpersonnelresource consumption

Ongoing burden to maintain data Less effort to operate school-level Burdensome for staff if individualsystem expenditure data collection; done time reporting is required

primarily by clerical staff

Comparative analyses Variations in total or per pupil Variations in resource use inexpenditures, by level, school, function, instructional (and other) programsprogram, and time and can link different resource mixes

with student outcomes

Usefulness for analysis Can determine variation in expenditures, More useful for understanding howbut not causes for differences among programs operate; with use ofprograms or student outcomes standard prices can determine

separate effects of price, quantity,capacity, and mix of resources usedto achieve results

Limitations for analysis Unable to separate effects of price, Without monetizing personnelquantity, capacity, and mix of resources resource use, cannot easily combineused to achieve results multiple staff services or compare

alternative program approaches

Focus on personnel data primarilyand not including nonpersonnelresources



102

erate within the school are combined tomake up the school unit.

Type of Data

Accounting approach

The accounting approach requires actualexpenditure data for the school categorizedby function, object, and program. Addi-tional levels of detail are available by in-cluding additional dimensions in the ac-counting code structure to classify expen-ditures, such as grade level in elementaryschools and subject matter in secondaryschools.

RCM approach

To match its focus, the RCM approach col-lects data on the physical resources utilizedin the school’s service delivery systems andothers if they are allocated to the school.For personnel resources, the primary basisis the time allocated to a given system,measured in counts of staff, full-timeequivalent staff, or contact time detailedin days, hours, or minutes. In order to es-tablish standard costs for the resources, itis also necessary to determine appropriatesalary levels for key positions, such as stateor national average salary amounts forclassroom teachers, instructional aides,speech therapists, and principals. For thesalaries to be comparable across geographi-cal areas, they are then indexed to com-pensate for differences in the cost of liv-ing. Consequently, it will be possible to re-port on the combinations of resources uti-lized in various service delivery systems ineither personnel time or standard costs thatare comparable across schools within a par-ticular state or in another state in the na-tion.

Nonpersonnel resources, if they are in-cluded, can be collected in physical quan-tities (e.g., number of computers, desks)and have standard prices applied to themto calculate standard costs. Alternatively,the actual expenditure amounts can be usedas a proxy for resource consumption to

avoid extensive data collection for a mi-nor portion of the budget.

Approach to Data Collection

Accounting approach

The current financial accounting systemcan be extended downward to collect andreport data at the school level. If states wereto modify their school accounting systemsto mandate the use of the operational unitdimension (which is now optional), aschool identifier code would be created intheir accounting systems. To implement thisDAE, districts would have to create andmaintain school-level expenditure accountssimilar to those at the district level. Theimplementation of a school-level account-ing system would substantially increase thenumber of accounts utilized by a schooldistrict since many accounts that now ex-ist only at the district level would have tobe duplicated for each school. While thismay be a significant undertaking, it is onewith which school personnel are familiar.In fact, many districts already account forexpenditures at this level of detail.

RCM approach

The need to collect detailed informationon the allocation of staff time to servicedelivery systems will require new proce-dures. The time of each staff member willhave to be assigned to one or more pro-grams so that 100 percent of the time ofall personnel is accounted for. The moststraightforward procedure would to use ei-ther district personnel records that main-tain actual staff assignments by FTE posi-tions by school and service delivery sys-tem (program) or budgeted staff positions.In the case of split assignments, allocationscould be made using consistent decisionrules to divide staff time among their workassignments. More precise data collectionwould involve establishing new, morecostly, and time-consuming procedures forpersonnel to report their actual time spentby service delivery system on an ongoingor periodic basis.

The need to collect

detailed

information on the

allocation of staff

time to service

delivery systems

will require new

procedures.


103

Implementation Issues

Accounting approach

Extension of an existing accounting sys-tem from the district to the school levelwould make implementation easier sincethe personnel are familiar with the primarydimensions and procedures of the system.

However, notwithstanding the familiaritywith the accounting system, there will beadditional effort and costs involved toimplement a school-level accounting sys-tem that will replicate many of the district-level accounts for each school in the dis-tricts. In order to provide necessary func-tion-object-program accounts at the schoollevel an estimated 300 new individual ac-counts would be required for each school.7

For larger districts with many schools, thiscould be a sizable task.

RCM approach

Establishment of a new personnel data sys-tem will be an additional burden on bothdistrict and school personnel. Such a taskcould be a major undertaking both in de-veloping the collection system and gainingacceptance for its use among schools. Thiswould be particularly true if the new datacollection system requires added effort onthe part of school instructional and sup-port personnel to report their time.

Effort to Maintain Data System

Accounting approach

The additional effort to operate the school-level accounting system would consist ofdata entry and maintenance of the chart ofaccounts with the new school-level expen-diture accounts. The marginal costs of en-tering an additional code for the opera-tional unit are minor since every expendi-ture is currently coded with function-ob-ject-program information. For those dis-tricts that already use the operational unitdimension, there is no additional data en-try work. However, a larger chart of ac-counts to which expenditures must be

posted will involve more effort on the partof the district business office.

RCM approach

If individual time reporting is required ofall personnel, then a substantial ongoingeffort will be necessary both to report thedata and to collect and organize them intoan appropriate system. However, if less bur-densome procedures can be utilized (e.g.,personnel assignment records, and budgetrecords), then the level of effort could besubstantially reduced.

Types of Analyses and Usefulness

Accounting approach

The accounting approach will provide dataon expenditures at the school level by func-tion, object, and program, and by grade orsubject matter if the system is extendeddown to that level. This will permit com-parative analyses of per student expendituresby school level and by any of the subcatego-ries utilized by the accounting system. Forexample, regular education expenditures perstudent can be compared across all the el-ementary schools in a district, or with stateor national averages. In a particular districtor a region, the variations in expendituresper student can be a useful initial analysisto identify outliers (low or high spenders)as schools for further investigation.

Although the accounting approach can pro-vide useful information on the magnitudeof the variation in per student spending lev-els, it does not provide any means of identi-fying the causes for such disparities. Differ-ences in reported expenditures per studentfor special education between two schoolsare the result of several factors, includingdifferent salary levels for the staff positions,different combinations of personnel re-sources, and different capacities of the pro-gram. So what might appear to be an effi-cient (lower cost) school might really be onlythe result of younger staff (and associatedlower salary levels) or a program that has

7 Based on an estimated 10 function/program accounts combined with an estimated 30 object accounts within each school.

…there will be

additional effort

and costs involved

to implement a

school-level

accounting system

that will replicate

many of the

district-level

accounts for each

school in the

districts.


104

larger class sizes. In fact, lower salary lev-els might mask underlying inefficiencies inthe manner in which the program is deliv-ered. On the other hand, the lower cost pro-gram might have achieved its results by us-ing a more efficient combination of person-nel and nonpersonnel resources. In eitherevent, the expenditure per student informa-tion from the accounting system will notallow determination of the causes of spend-ing differences.

RCM approach

The strength of the RCM approach is thedata that it provides for analyzing educa-tional programs. With data on resource usein instructional and other service deliverysystems it is possible to examine the differ-ences among similar programs in terms ofhow programs are delivered. Analyses canestablish the mixes of personnel andnonpersonnel resources that are used andpotentially link the allocation of resourcesto student outcomes.

Use of a standardized salary rate for eachstaff position has advantages for the inter-pretation of cost differences among schools.For example, since all similar positions willhave the same salary assigned to them, anyactual salary differences due to different lev-els of experience or education will be fac-tored out. Any variations in the costs ofschools or programs between or among dis-tricts would be due to differences in the or-ganization and operation of the service de-livery system (i.e., choices of resources usedin the system, quantities of resources used,and capacity of the system).

If standard prices are not used and theanalysis is based on variations in person-nel use (time spent), then comparisonsamong service delivery systems that em-ploy multiple staff in alternative combina-tions are very difficult. However, the useof standard prices permits such compari-sons, as illustrated in table 8. ElementarySchool A uses 24 teachers to achieve anaverage class size of 18.8, but uses no in-structional aides. School C makes a differ-ent choice employing 18 teachers alongwith 9 instructional aides to serve its stu-dents with an average class size of 25 stu-dents. Which school uses more resources?A RCM analysis can compare the two in-structional approaches in terms of theirchoice of personnel. Using the standardprices for personnel in table 5, the analysisindicates that School A consumes approxi-mately $100,000 more resources (in stan-dard cost terms) than School C.

Which school is more efficient? This is amore difficult question to answer. Instruc-tional programs in School C cost $100,000less to operate, but this is caused by re-placing six teachers with nine aides. Theresult is a higher student/teacher ratio inSchool C (25.0 to 18.8), but a lower stu-dent/adult ratio (16.7 to 18.8) than inSchool A. If the student outcomes weresimilar, then School C would be more effi-cient—same results for less cost. However,if School A achieved higher student out-comes (i.e., is more effective), then the is-sue becomes are the student gains worththe increased cost? While the RCM analy-sis does not answer this policy question, it

Table 8.—Comparison of resource consumption

School A School C

Position Standard cost Number Cost Number Cost

Students 450 450

Teachers $48,000 24 $1,152,000 18 $864,000

Aides $21,000 0 0 9 189,000

Total $1,152,000 $1,053,000

Student/teacher ratio 18.8 25.0

Student/adult ratio 18.8 16.7SOURCE: Authors‘ sketch.

The strength of the

RCM approach is

the data that it

provides for

analyzing

educational

programs.


105

poses it as an important issue to be re-solved. In general, the use of standard costswith resource use data permits the analystto cost out the variations in the use of re-sources and to compare the costs of eachschool and the costs of achieving the stu-dent results.

If the choice is to focus only on personnelresources, then the analysis is not completesince nonpersonnel resources are not in-cluded. This may not be a serious limita-tion in many service delivery systems wherepersonnel are the dominant resource, butit could prove misleading for programs thatemploy a high degree of nonpersonnel re-sources, such as trading a greater use oftechnology for fewer personnel.

Synthesis of Two Approaches

To reach a synthesis of the accounting andRCM approaches to the collection and re-porting of school-level financial data, thefollowing recommendations incorporatethe positive aspects of each approach whileavoiding the primary problems of the dif-ferent perspectives. Taken together, the se-lected components form a single combinedapproach that balances the differing dataneeds of the various stakeholders with thecosts and burden of obtaining such data.The synthesis of the two approaches is ac-tually the choice and definition of the ele-ments from the accounting and RCM ap-proaches that provide the framework tomeasure education expenditures at theschool level for the purpose of linking stu-dent costs to student results.

The recommendations for a synthesis of thetwo approaches are presented by majorelement and are summarized in table 9. Thefirst element, choice of approach, proposesthe accounting approach to collect expen-diture data and the Resource cost modelapproach to collect staff data. The nextthree elements, elements to include at theschool level, elements to remain at the dis-trict level, and level of reporting, providerecommendations on the reporting struc-ture for direct, multiple and centralizedexpenditures and staff data. In addition,

these elements propose the reporting ofexpenditure and staff (service delivery sys-tems) data in a structure that can be usefulfrom the school level to the national level.The synthesis of the two approaches alsodefines the reporting systems for the col-lection of staff and student data elementsin a comparable format. Finally, the analy-sis component in the school-level data sys-tem defines how the data elements can beused to determine variations in student ex-penditures and staffing patterns as well asfor productivity study.

Choice of Approach

The DAE of the financial accounting sys-tem is recommended as the approach toobtaining expenditure data at the schoollevel. The current district-level accountingsystem would be extended to the schoollevel. The operational unit for reporting ex-penditure data would become a mandatorydimension in state school accounting sys-tems. If this recommendation is made inthe new federal fiscal accounting hand-book, states would be encouraged tomodify their school accounting systems toreport school-level fiscal data. In turn, arequirement by states for districts to reportschool-level data would necessitate creat-ing school-level accounts for all expendi-tures at that level. An earlier pilot test ofcollecting school-level expenditure data inthe standardized NCES function and ob-ject categories found that districts had dif-ficulty in reporting financial data at theschool level; only half of the pilot districtscompleted the expenditure survey (Issacset al. 1999). However, the pilot districtswere in states that did not require school-level accounting and the procedures wereunfamiliar and proved very time-consum-ing. By contrast, if the reporting of school-level expenditure were mandatory, statesand school districts would develop stan-dardized operating procedures. The newprocedures would become part of the ac-counting system and would become just asroutine as the district-level expenditure datacollection and reporting system is now.

The DAE of the

financial

accounting system

is recommended as

the approach to

obtaining

expenditure data

at the school level.


106

Table 9.—Synthesis of two approaches for school-level data system

Resource cost modelElements Accounting approach (RCM) approach

Choice of approach Accounting approach for expenditures RCM approach for staffresources; nonpersonnelresources excluded

Elements to include at All direct school-level expenditures All direct school-levelschool level personnel

Allocation of joint and multiple Allocation of personnel servingexpenditures to school level multiple schools

Elements to remain at Centralized expenditures not identifiable No allocation of time ofdistrict level by school; maintained by subfunction central office personnel

and object

Level of reporting All school-related expenditures, by Service delivery system,function/object/program corresponding to program

structure in accountingKey data elements reported to higher approachlevels via a data pyramid

Staff Separate system gatheringbudgeted time allocated bystaff to school activities;organized by service deliverysystem (program/function)

Students Expenditure function/program Separate system to collectcategories compatible with student number of students (full-timedata categories equivalent), by service delivery

system (program)

Outcome measures, by student and Outcome measures, by studentprogram and program

Analyses Variation in expenditures per student Variation in staffing patterns, byby school, function, and program service delivery system and

school

Application by analysts ofstandardprices to staffcategories to estimatestandard costs and causes ofvariation

Productivity analysescombining staff, standard costs,and student outcomes


The RCM approach is recommended forcollecting school-level staff data. Informa-tion on the deployment of staff resourcesis critical to school districts since person-nel are their most important and mostcostly resource. The recommended data col-lection process would use district person-

nel records that report actual staff assign-ments by FTE positions by school andwithin the school by instructional programor support function. These basic recordswould be extended, if necessary, by allo-cating the time of any staff assigned to morethan a single area in the school.


107

Elements to Include

All direct school-level expenditures shouldbe the first component to include. Theseare all objects of expenditure (staff sala-ries, benefits, purchased services, supplies,property, and other objects) that are di-rectly identified with an individual school.In recording these expenditures into the ac-counting records, the operational unit codefor the school would be included.

Secondly, all joint and multiple expendi-tures that are made for instruction, instruc-tional support, or operational support ac-tivities for two or more schools should beallocated to those schools in a reasonableand consistent way. Examples of these ser-vices are: attendance and social work, guid-ance, health, psychological, speech pathol-ogy and audiology, improvement of instruc-tion, educational media, school adminis-tration, and operation of maintenance andplant. They are consistent with the supportservices subfunctions in federal and stateschool accounting manuals.

To make the allocated expenditure datacomparable across schools and districts, auniform set of allocation procedures wouldneed to be established. The rules for allo-cation should be related to the cause of theexpenditure, easy to understand, and uti-lize readily available data. As previouslynoted, only a few allocation methodolo-gies meet these criteria; they include num-ber of students, number of staff, staff time,square footage a building, and physicalquantities of important nonpersonnel re-sources, such as computers. Choice of abasis for allocation of a particular expen-diture should be logical and related to thetype of expenditure. For example: the per-sonnel costs for a speech therapist servingstudents in multiple schools could be allo-cated on the basis of the number of stu-dents served in each school; the personnelcosts for a school nurse who spends twodays per week in one school and three daysper week in another could be allocated 40percent (2 days/5 days) to the first schooland 60 percent (3 days/5 days) to the other;

and the costs of a cleaning contract with amaintenance firm could be allocated to theschools on the basis of square footage ofeach building.

Staff resource utilization data should fol-low the same pattern as expenditures. Thetime for all personnel who work full timein a school should be reported at thatschool. Any personnel serving multipleschools should have their time allocatedbetween the schools on the same basis.

Consistency is a critical consideration forreporting school-level data. While theremay be legitimate differences of opinionabout which specific expenditures or staffto include at the school level or to remainat the district level or about the most ap-propriate allocation methodology for agiven joint or multiple expenditure, it is im-portant that those decisions be made thesame way by all schools and districts in astate and, to the extent possible, by allstates. Only by establishing consistent pro-cedures through state school accountingpolicies can the resulting data be reliableand useful for comparative analyses.

Elements to Remain at District Level

It is further recommended that district levelexpenditures, which are not directly iden-tifiable with one or more schools, be ac-counted for at the district level and not al-located to the schools. These include ser-vices such as governing board, executiveservices, business services, centralized op-eration and maintenance services, transpor-tation, centralized support (planning, infor-mation, staff, and data processing), com-munity services, and debt payments. Again,these categories are consistent with exist-ing federal and state educational account-ing systems.

School administrators and other school-level personnel have little authority over orresponsibility for centralized district-levelexpenditures. So, rather than load schoolswith nonschool-level expenditures, theschool-level data should reflect only school-level operations.

Only by

establishing

consistent

procedures

through state

school accounting

policies can the

resulting data be

reliable and useful

for comparative

analyses.


108

With this division, centralized expenditurescan be maintained and reported at the levelof those responsible for the expenditures.This also permits a more useful district-levelanalysis of these expenditures, in which ex-penditures per student for centralized ser-vices can be compared across districts,variations identified, and areas for man-agement attention noted. In a similar fash-ion and for the same reasons, staff time forpersonnel in centralized office activitiesshould not be allocated to the school level.

Level of Reporting

Expenditures should be reported at theschool level by function, object, and pro-gram. This follows the current practice atthe district level. This level of reporting al-lows individual school expenditure data tobe collected on instructional and supportactivities, by personnel and nonpersonnel

items, by type of instructional program, andby all combinations of these dimensions.An example of the type of expenditure datathat could be reported is shown in table10. Four main groupings are given: instruc-tion, instructional support, operational sup-port, and administration. Where appropri-ate, subcategories are used to provide ad-ditional detail. These functions capture themain aspects of educational activities at theschool level and would be reported for alldirect expenditures and those allocated tothe school from joint, multiple, or centraloffice expenditures specifically identifiablewith the school. At the object of expendi-ture level, salaries and benefits would beshown separately, but all nonpersonnelitems would be reported in a single group-ing. However, if additional detail were de-sired for a comparative or analytical pur-pose, the data would be available.

Table 10.—Expenditures reported at school level

Instructional programs

Regular programs

Salaries*

Employee benefits*

Nonpersonnel expenditures*

Special programs

Vocational programs

Other Instructional programs

Instructional support functions

Attendance and social work

Guidance

Health

Psychological

Speech pathology and audiology


Educational media

Operational support

Administration

* These three types of objects will be reported for all programs and functions.



109

Data Pyramid

With detailed expenditure data collectedat the school level, a data pyramid can becreated in which key elements would bereported from one level to the next—schoollevel to district level, district level to statelevel, and state level to national level. Itwould function as a reduced form report-ing system in which only selected datawould be reported going up the pyramidfrom one level to the next and those datawould be selected for policy and decisionmaking relevance. The diagram for the datapyramid is given in table 11.

At each level, there would be a summarythat reports the averages for that level andselected data for the levels below. To func-tion effectively, there would have to be alimited number of data elements reportedfrom one level to the next and the report-ing system would need to include not only

expenditure data, but also staffing and stu-dent data in the compatible categories. Thisapproach is illustrated in table 12. For ex-ample, for expenditures at the state level,there would be:

1. a summary of expenditure datacategories for the state (total and by eachtype of school);

2. selected data for each district that wouldinclude both central office and school-level expenditures; and

3. selected expenditure data for eachschool.

However, the school-level data reported atthe state level would only be the key indi-cators relevant for educational policymak-ers and not all of the expenditure and re-source use data collected at each school.

Table 11.—Illustration of data pyramid

National

National summarySelected data for 50 states

State

State summarySelected data for all districtsSelected data for all schools

District

District summarySelected data for each school in district

School

Complete data for each schoolExpenditures Staff Students



110

Table 12.—Elements in data pyramid at each level

National

Summarized for nation: total and averages, by type of school (elementary, middle, high)

Selected data for each state

Expenditures (cost per student)1 Staff (full-time equivalent (FTE), student/FTE)2 Students (enrollments)

Total Total Total

By instructional program By instructional program By instructional program

Regular programs Regular programs Regular programs

Special programs Special programs Special programs

Instructional support Instructional support

Operational support Operational support Student outcomes

Administration Administration Regular programs

State

Data summarized for state: total and averages, by type of school (elementary, middle, high)

Selected data for each district

Selected school-level data for each school

Expenditures (cost per student)1 Staff (FTE, student/FTE)2 Students (enrollments)

Total Total Total







District

Data summarized for district : total and averages, by type of school (elementary, middle, high)

Selected school-level data for each school

District-level expenditures (not allocated to schools)Expenditures (cost and cost per student)1 Staff (FTE, student/FTE)2 Students (enrollments)

Total Total Total




Compensatory programs Compensatory programs Compensatory programs

Vocational programs Vocational programs Vocational programs

Other programs Other programs Other programs





111

The optional dimensions of the current ac-counting system of level of instruction andsubject matter will remain available forthose states and districts that wish to col-lect and report expenditures at lower lev-els. Coding expenditures by these dimen-sions would allow a district to compare thecosts of third grade classrooms within aschool or across all schools in the districtor make a similar comparison of the costsof 11th grade mathematics. However, thereis a substantial cost to creating such infor-mation, both in the development of the pro-cedures and structure to collect and reportthe data as well as the effort for individu-als to maintain the system. At this time itdoes not appear that the benefits outweighthe cost of a further downward extensionof the accounting system and it is recom-mended that these dimensions not becomemandatory for all districts.

Since the RCM focuses on the service de-livery system as the unit of analysis, the

level of staff resource utilization should bereported at this level. However, for com-patibility with the expenditure data, it isnecessary that the service delivery systemdefinitions correspond to the programstructure (for instruction) and functionstructure (for support activities) of the ac-counting system. Otherwise, data from thetwo systems would not be organized in thesame categories and their use for analysiswould be compromised.

Staff

Staff data would be collected through theRCM approach. The data should be orga-nized by service delivery system, but as pre-viously indicated these categories should becompatible with the program categories uti-lized for expenditures in the accounting sys-tem. Within each service delivery system,staff data should be collected and reported

Table 12.—Elements in data pyramid at each level–Continued

School

Data for each school in districtExpenditures (cost and cost per student)1 Staff (FTE, student/FTE)2 Students (enrollments)

Total Total Total




Compensatory programs Compensatory programs Compensatory programs

Vocational programs Vocational programs Vocational programs

Other programs Other programs Other programs

By instructional support function By instructional support function

Library Library Student outcomes

Media Media By instructional program

Technology Technology Regular programs

Guidance Guidance Special programs

Psychological services Psychological services Compensatory programs

Speech pathology and audiology Speech pathology and audiology Vocational programs

Social work Social work Other programs

Pupil health Pupil health

Operational support Operational support

Administration Administration

1 Expenditure data reported by personnel, benefits, and nonpersonnel.

2 FTE for professional and nonprofessional staff.



112

by the major types of staff in a school sys-tem. The staff positions should be compat-ible with the standardized listing of the staffcategories available from the NCES StaffData Handbook (Malitz 1995); it specifiesten major categories with more detailedbreakdowns of subcategories within eachone. The major categories and their descrip-tions are shown in table 13.

The staff data elements are the time for allschool personnel, reported as FTEs. Timeof personnel serving several schools wouldbe allocated appropriately among theschools involved.

Table 13.—Staff categories

Official–administrativePerforms management activities that require developing broad policies and executing those policies throughdirection of individuals at all levels. This includes high-level administrative activities performed directly for policymakers.

Professional–educationalPerforms duties requiring a high degree of knowledge and skills generally acquired through at least abaccalaureate degree (or its equivalent obtained through special study and/or experience) including skills in thefield of education, educational psychology, educational social work, or an education therapy field.

Professional–otherPerforms assignments requiring a high degree of knowledge and skills usually acquired through at least abaccalaureate degree (or its equivalent obtained through special study and/or experience) but not necessarilyrequiring skills in the field of education.

ParaprofessionalsWorks alongside and assists professional individuals.

TechnicalPerforms tasks requiring a combination of basic scientific knowledge and manual skills which can be obtainedthrough approximately two years of postsecondary education such as that which is offered in community/juniorcolleges and technical institutes, or through equivalent special study and/or on-the-job training.

Office/clerical/administrative supportPerforms the activities of preparing, transferring, transcribing, systematizing, or preserving communications,records, and transactions.

Crafts and tradesPerforms tasks requiring high manual skill level which is acquired through on-the-job training and experience orthrough apprenticeship or other formal training programs. This assignment requires considerable judgment anda thorough and comprehensive knowledge of the processes involved in the work.

OperativePerforms tasks requiring intermediate level manual skills which can be mastered in a few weeks through limitedtraining to operate machines. This includes bus drivers and vehicle operators.

LaborerPerforms tasks requiring some manual skills which can be conducted with no special training. This includesindividuals performing lifting, digging, mixing, loading, and pulling operations.

Service workPerforms tasks regardless of level of difficulty which relates to both protective and nonprotective supportiveservices.

SOURCE: Malitz, G. 1995. Staff Data Handbook: Elementary, Secondary, and Early Childhood Education. NCESHandbook. Washington, DC: U.S. Department of Education, National Center for Educational Statistics (NCES 95–327).


113

Students

Student data collection would not be acomponent of either the accounting orRCM approach. Rather, it should be donein a separate system compatible with theaccounting program structure and theRCM service delivery categories. Studentsshould be measured in numbers of students,preferably FTE students in each program.

Student outcomes measures or perfor-mance indicators are necessary for effec-tiveness and accountability analyses. Suchmeasures and indicators also are not a partof either the accounting or RCM approach,but they should be collected by student inthe same format as the accounting programstructure and the RCM service delivery cat-egories.

Analyses

With the accounting approach to the col-lection of expenditure data combined with

student participation data, it will be pos-sible to calculate the expenditure per stu-dent along several dimensions, includingtotal school, instructional programs withina school, and functions within a school.This will permit analyses of the variationin expenditures per student along these di-mensions. It will also permit comparisonsof the expenditures among schools in asingle district or among schools across thestate or nation. With these analyses, highand low expenditure schools can be identi-fied and targeted for further investigationon the reasons for their deviation.

To illustrate an analysis utilizing data fromthe accounting approach, a single districtexample is shown in table 14 that comparesexpenditures across three elementaryschools of the same size. The expendituredata would be available from the account-ing approach. The expenditure data arecombined with the student data to calcu-late expenditure per student. The three

Table 14.—Expenditure comparisons among elementary schools in a single district usingdata from the accounting approach

Students Aster Bluebell Camellia District

Regular programs 450 450 450 1,350

Special programs 45 40 50 135

Function Description

1100 Regular programs 1,779,900 1,483,000 1,164,600 4,427,500Cost per student 3,955 3,296 2,588 3,280

1200 Special programs 290,650 227,700 238,450 756,800Cost per student 6,459 5,693 4,769 5,606

1400 Other instructional programs 315,300 344,300 349,900 1,009,500Cost per student 701 765 778 748

2100 Student support services 33,800 29,900 24,050 87,750Cost per student 68 61 48 59

2200 Instructional staff support 135,200 149,500 96,200 380,900Cost per student 273 305 192 256

2410 Principal’s office 258,900 233,700 207,150 699,750Cost per student 523 477 414 471

610 Building operation 281,500 254,600 277,750 813,850Cost per student 569 520 556 548

School total $3,095,250 $2,722,700 $2,358,100 $8,176,050Cost per student $6,253 $5,557 $4,716 $5,506



114

schools have widely differing costs per stu-dent, ranging from a high of $6,253 in As-ter Elementary to a low of $4,716 in Ca-mellia Elementary. Each of the main in-structional programs and supportsubfunctions also show varying costs perstudent among the schools. While the ex-penditure comparisons will not explain whythe costs are different among the schools,they do indicate that there are substantialdifferences and suggest further investiga-tion to arrive at the causes of the differ-ences.

With the addition of staffing data thatwould be available from the RCM ap-proach, some of the reasons for the differ-ences among the schools can be investi-gated. Staffing data in FTE positions as-signed to the schools are shown in table 15along the same categories as the expendi-tures. Although the schools are the samesize, they have chosen to utilize differentquantities and types of staff, particularlyin regular program. The staff data can thenbe used in a more detailed analysis.

To continue the illustrative analysis, regu-lar programs, the main instructional pro-gram and largest single expenditure area,is selected. The reasons for the differencesin per student expenditures in regular pro-grams are explored in table 16. For regularprograms, Aster has the highest per studentcost, $3,955 while Camellia remains thelowest at $2,588. The data required for this

analysis come from the three primarygroups—expenditures, staff utilization, andstudents. The school-level expenditureswould be available from the accounting ap-proach, the number of teachers in eachschool by program would be available fromRCM staff assignment information, andthe numbers of students in each schoolcould be obtained from existing student re-porting systems. Table 16 shows the ex-penditure data for regular programs ingreater detail, in which salaries for teach-ers and aides, benefits, and nonpersonnelexpenditures are reported separately. Thereason for the suggested focus on person-nel expenditures (salaries and benefits) isdemonstrated by the high percentage oftotal program expenditures that they rep-resent—from 88 to 91 percent across allschools.

Examining the staff data for regular pro-grams (table 15), Aster uses only teachersin its regular program, while the other twoschools have varying combinations ofteachers and instructional aides. Using thenumbers of teachers and aides, the aver-age salary for each group is calculated. Thisanalysis uncovers a primary cause of theper student differences. Aster has an aver-age salary per teacher of $52,000, whilethe salaries of teachers at Camellia aver-age $37,000, a $15,000 difference perteacher. The most likely explanation forthese differences would be the greater se-niority or education level of teachers at

Table 15.—Staffing comparisons among elementary schools in a single district using datafrom the resource cost model approach

Function Description Aster Bluebell Camellia District

1100 Regular programs–teachers 24.0 21.0 18.0 63.0

Regular programs–aides 0.0 4.0 9.0 13.0

1200 Special programs 4.0 3.5 4.5 12.0

1400 Other instructional programs 3.0 3.5 4.0 10.5

2100 Student support services 0.5 0.5 0.5 1.5

2200 Instructional staff support 2.0 2.5 2.0 6.5

2410 Principal’s office 1.5 1.5 1.5 4.5

2610 Building operation 2.5 2.0 2.5 7.0

School totals 37.5 38.5 42.0 118.0



115

Aster and the relatively low seniority oreducation level of teachers at Camellia.8

Lower variations are present in the aver-age aide salaries.

The use of standard prices for personnelresources (an RCM analytical procedure)permits the calculation of standard costsof service delivery systems and the impactof different salary levels as a cause of varia-tions among schools. This is also illustratedin table 16 where the cost per student ateach school is recalculated using the dis-trict average teacher salary for all schools.This removes the average salary differen-

tials as a cause of variation. This lessensthe per student differences, but does noteliminate them; the variation per studenthas been reduced from $1,367 with actualsalary levels to $463 using average salarylevels.

However, salaries are not the only cause ofper student expenditure differences. An-other cause is the different staffing ratiosamong the schools. The student/teacher andstudent/aide ratios, calculated from staffand student data at each school, also pro-duce per student expenditure differencesamong the schools. The low student/teacher

Table 16.—Analysis of expenditure differences in regular programs using data from theaccounting and resource cost model approaches

Description Aster Bluebell Camellia District

Expenditure detail

Teacher salaries 1,248,000 966,000 666,000 2,880,000Aide salaries 0 64,000 126,000 190,000Total salaries 1,248,000 1,030,000 792,000 3,070,000Benefits 374,400 309,000 237,600 921,000Nonpersonnel costs 157,500 144,000 135,000 436,500

Personnel costs (percentage) 91 90 88 90Total regular programs $1,779,900 $1,483,000 $1,164,600 $4,427,500

Cost per student $3,955 $3,296 $2,588 $3,280

Causes for expenditure differencesDifferences in average salaries

Average salary per teacher $52,000 $46,000 $37,000 $45,714Average salary per aide — $16,000 $14,000 $14,615

Cost per student at district average salaries $3,520 $3,262 $3,057 $3,280

Staffing ratiosStudent/teacher ratio 18.8 21.4 25.0 21.4Teacher/aide ratio — 5.3 2.0 4.8Student/adult ratio 18.8 18.0 16.7 17.8

Mix of personnelTeachers 24.0 21.0 18.0 63.0Aides (weighted by salary ratio)* 0.0 1.4 3.4 4.2Total weighted staff 24.0 22.4 21.4 67.2

— Not applicable.

* Salary ratio is the ratio of average aide salary to average teacher salary. This ratio shows the equivalentteachers that aide salaries could purchase.


8 In most districts, teacher salaries are primarily determined by length of service and educational level with length of service having thelargest impact.


116

ratio at Aster (18.8) translates into higherper student costs, while the higher ratio atCamellia (25.0) results in lower per stu-dent costs. So for Aster both the higheraverage teacher salary and the lower stu-dent/teacher ratio create higher per studentcosts; at Camellia the reverse is true. Thelower teacher costs for Camellia (and Blue-bell) are partially offset by the addition ofthe expenditures for instructional aides.

The mix of personnel is the third cause ofcost variations. More detailed staff utili-zation data from the RCM, such as differ-entiation of staff within programs, permitsanalysts to examine variations in staffingpatterns in schools and instructional pro-grams. Applying the salary ratio (averageaide salary to average teacher ratio) to thenumber of aides gives the number ofequivalent teachers that the monies allo-cated to aide salaries could buy for eachschool. Aster has the highest number ofweighted staff, which raises the cost perstudent, and Camellia has the lowest num-ber of weighted staff, which results in alower relative cost per student. In otherwords, the nine aides employed in Camel-lia cost much less than the six teachers thatthey replaced (in comparison with Aster).Further calculations could quantify theamount of expenditure variation attribut-able to each factor.

Additional analyses linking the standardcost information with student outcomescan lead to more detailed analyses and theidentification of high performing and cost-effective programs. In turn, these resultscan lead to policy decisions and legislationthat directs funding toward these pro-grams. Through dissemination of the re-search results and policy directives, schooland district administrators can focus theirefforts on more effective instructional ap-proaches.

Summary: Stakeholders andRecommended Approaches

To summarize, the synthesis of the account-ing and RCM approaches to the collectionand reporting of school-level data is predi-

cated on a downward extension of the cur-rent district-level accounting system andthe collection of personnel data (allocationof time) utilizing the RCM approach. Thecombination of the two approaches pro-vides both financial and personnel resourceinformation at the school level. Havingschool-level expenditure data for bothmanagement control over operations andcomparisons among schools enhances un-derstanding of the allocation of funds.Availability of personnel time allocationsfor the same programs and functions per-mits improved analyses of program opera-tion and effectiveness. Along with studentdata, the accounting and RCM approachesprovide many more opportunities for per-ceptive and beneficial uses of the data thaneither approach alone.

How well do the recommendations fill theinformation needs of stakeholders forschool-level data and how will the dataprocedures deal with the major issues fac-ing them? Table 17 presents a summary re-sponse to these questions, which is sup-ported by the following discussion.

Schools and School Districts

The DAE fits the needs of school and dis-trict administrators and school boards. Itprovides school-level financial data that arecompatible in content and format with theexisting district-level accounting system.Consequently, while the implementation ofthe DAE requires additional effort, it doesnot represent a new or unfamiliar approachto the collection and reporting of financialdata. In fact, many districts already use theoperational unit dimension of the account-ing system that is recommended to be mademandatory. The expenditures to include forthe school-level data are the direct expen-ditures for the school along with the allo-cation of joint and multiple expendituresidentifiable by school. District office ex-penditures are not allocated to the school,but are accounted for separately at the cen-tral level. The resultant financial data arefocused on expenditures at or related tothe school; these are expenditures that are

Along with student

data, the

accounting and

RCM approaches

provide many

more opportunities

for perceptive and

beneficial uses of

the data than

either approach

alone.


117

Table 17.—Stakeholders and synthesis of school-level dataStakeholders Accounting approach Resource cost model (RCM) approach

Easily understood extension ofpresent district-level accountingsystem

Can collect expenditure data without Staff data obtained from district staffsubstantial change to present assignment reports or budget, notoperation from personnel time reporting

system

School administrators Can use information by program to Can compare staff use by programDistrict administrators monitor and compare school and compare different mixes ofSchool boards expenditure performance instructional staff

Can use expenditure comparisons to Can use staff use comparisons toidentify areas for improvement and identify areas for improvement andtake action take action

Centralized expenditures kept at level Have access to research reports towith authority for them identify most effective and cost-

efficient practices

Have data to maintain fiduciaryresponsibilities

Will have staff resource use data forcomparative analyses of staffingpatterns

Can apply standard costs to staff dataResearchers Have expenditure data for to develop comparative analyses ofPolicy analysts comparative analyses program and school operation

Can use staff data, standard costs, andstudent outcome measures todevelop analyses and comparisons ofcost-effective programs

Uniform system of collecting and Uniform system of collecting andreporting school expenditure data reporting staff utilization data

Extension of present accounting Collection of staff time data can beState policymakers system to school level is a done from budget plans and notNational policymakers straightforward step require an extensive time reporting

system

Have statewide and national Analyses of cost-effective programscomparative data for monitoring use provide information for resourceof public funds and policy decisions allocation decisions and educational

policy


reasonably under the authority and respon-sibility of the school principal.

The school-level expenditure reporting byinstructional program and support func-tions provide school and district adminis-trators with information to develop accu-

rate school-level budgets, to monitorschool-level spending, to maintain fi-duciary responsibility for public funds,and to compare levels of spendingacross schools within a district, withina region, or with similar schools in thestate or nation. The comparisons should


118

lead to identification of spending levelswithin a given school that are higher andlower than comparative schools. Armedwith this information, administrators andschool staff can target appropriate areas forimprovement and take action to rectifyproblem areas. When combined with stu-dent outcome measures, the school-level ex-penditure data will facilitate accountabil-ity analyses at the school level. The separa-tion of district-level expenditures will per-mit more relevant comparative analyses ofdistrict-level operations.

Staffing data will also be produced at theschool level that will specify the number ofstaff (FTE) in the same instructional pro-gram and support function categories as ex-penditures. This consistency will permitcombination of the data for comparativeanalyses by school district personnel as wellas researchers and policy analysts. Compari-sons of staffing patterns can be made acrossschools and districts to identify efficient andinefficient practices.

Researchers and Policy Analysts

The expenditure and staff allocation dataavailable at the school level will allow re-searchers and policy analysts to fulfill theirrole as producers of analyses and reportsfor use by the other two groups. The ex-penditure data will permit comparativeanalyses of variations in actual spendingacross schools, programs, functions, andobjects. These expenditure comparisons willlead to further investigations of equity andadequacy of the school spending.

School-level staff resource utilization datawill provide researchers with significant andsubstantial new information to conduct avariety of important analyses. Examinationand comparison of staffing patterns willyield insights into the effectiveness of dif-ferent instructional approaches and, in turn,lead to recommendations of more effectiveresource allocation patterns and policies.

Application of standard prices to the school-level staff data permits analysts (not schoolor district personnel) to calculate and com-pare standard costs for various programs

and school operations. Without the con-founding factor of varying local prices forthe same resource, it will be possible todetermine the efficiency of different mix-tures of personnel resources in instructionalprograms. Combining staff use and stan-dard costs of programs with student out-come measures can lead to productivityanalyses that identity cost-effective pro-grams for replication by other districts andfor funding by legislatures.

State and National Policymakers

The recommended policies and proceduresfor school-level financial data establish auniform system for collecting and report-ing expenditure and personnel allocationinformation. While the implementation ofa new and expanded system of data col-lection at the school level is not withoutcosts, the synthesized approach representsan efficient means without placing undueburden on schools or school districts. TheDAE uses a familiar system and data pro-cedures that are extended to the schoollevel. This approach is reported as beingrequired in 19 states and many schools anddistricts in other states already utilizeschool-level expenditure data as well. Sincedistrict data systems need to keep track oftheir most valuable and expensive resource,people, collecting and reporting of school-level staff data will build on existing dataprocedures for maintaining control overpersonnel resources and not represent anoverburdening task.

The synthesized approach will provide aconsiderably improved data system todocument and compare spending levels foreducational programs. Analyses and iden-tification of effective programs and instruc-tional approaches would become possiblewith detailed staff information. When com-bined with student outcome measures, sig-nificantly enhanced analyses will becomepossible to identify high performing pro-grams and the underlying causes for theirresults. This will give state and nationalpolicymakers valuable information tomonitor the use of public funds for educa-tion and to make future decisions aboutthe magnitude and focus of new funding.

School-level staff

resource

utilization data

will provide

researchers with

significant and

substantial new

information to

conduct a variety

of important

analyses.


119

References

Chambers, J.G. 1999. Resources in Education: From Accounting to the Resource Cost Model Approach. Wash-ington, DC: U.S. Department of Education, National Center for Education Statistics (Working Paper 1999–16).

Fowler, W.J. 1997. Financial Accounting for Local and State School Systems, 1990. Washington, DC: U.S.Department of Education, National Center for Education Statistics (NCES 97–096R).

Hartman, W.T. 1999. School District Budgeting. Reston, VA: Association of School Business Administrators.11–14.

Issacs, J.B., Best, C.M., Cullen, A., Garet, M.S., and Sherman, J.D. 1998. Collection of Public School Expendi-ture Data: Development of a Questionnaire. Washington, DC: U.S. Department of Education, National Centerfor Education Statistics (Working Paper 1998–01).

Issacs, J.B., Garet, M.S., Sherman, J.D., Cullen, A., and Phelps, R. 1999. Collection of Resource and Expendi-ture Data on the Schools and Staffing Survey. Washington, DC: U.S. Department of Education, NationalCenter for Education Statistics (Working Paper 1999–07).

Malitz, G. 1995. Staff Data Handbook: Elementary, Secondary, and Early Childhood Education. Washington,DC: U.S. Department of Education, National Center for Education Statistics (NCES 1995–327).

Sherman, J.D., Best, C. and Luskin, L. 1996. Assessment and Analysis of School Level Expenditures. Washing-ton, DC: U.S. Department of Education, National Center for Education Statistics (Working Paper 96–19).

Young, B. 1999. Characteristics of the 100 Largest Public Elementary and Secondary School Districts in theUnited States: 1997–98. Washington, DC: U.S. Department of Education, National Center for EducationStatistics (NCES 1999–318).


120

how has teacher compensation changed?nces.ed.gov/pubs2001/2001378_2.pdf · how has teacher...

Documents