Technology and the Broken Higher Education Cost Model

Home , Higher education

Technology and the Broken CostHigher Education Model Insights from the Delta Cost Project By Rita Kirshstein and Jane Wellman

lthough U.S. higher education has faced numerous crises and di- lemmas in its history, the situation in which colleges and universities find themselves at the moment is indeed different. Shrink- ing public subsidies coupled with historic rises in tuitions come at the same time that colleges and universities have been tasked to dramatically increase the number of individuals with postsecondary degrees. Additionally, many of these students need financial aid, putting further strains on the higher education system. The stratification between rich and poor institutions in their access to resources Ais also growing. These conditions make the current “cost model” under which higher education has typically operated no longer sustainable and have led to college and university leaders examining alternative ways to deliver both high-quality and af- fordable higher education. These alternatives incorporate technology and include access to distance-delivered education and services, a focus on learners’ outcomes rather than inputs, and technologically sophisticated buildings and classrooms.

The changes are welcome and to peers. Doing so, however, is easier (IPEDS) survey—the project has har- largely overdue in much of higher said than done. The higher education monized and standardized the data as education, but unless the use of tech- “industry” has long overlooked the much as possible in order to account nology, whether in instruction or in spending side of the college cost equa- for revisions in accounting standards the operation of the institution, is tion. Analyses of institutional finances and IPEDS reporting formats. These guided by an understanding of higher have historically been dominated by adjustments allow the analysis of education costs and cost structures, a focus on revenues and institutional changes over time and the benchmark- its use will not fix the problem of a financial conditions (such as year-end ing of costs among different types of broken higher education cost model. financial statements or audits), on stu- institutions. The project also has devel- This problem is not confined to the dent tuition and financial aid, and on oped a number of metrics to focus on way that instruction is funded and the broader concerns around college different types of spending—for exam- delivered; rather, it is much broader, affordability. Such an emphasis con- ple, not only the spending associated including the costs of academic and ad- flates spending with revenues, which with the direct educational mission of ministrative overhead and the largely is a mistake, since many sources of colleges and universities but also the unexamined “fixed costs” that drive revenue are not available for general spending on research, public service,

It is important not only to refocus attention on college spending but also to understand the constituent elements of costs on both a per-student and a per-degree basis, since that is the only way to determine whether alternative investments are cost-effective.

so much of institutional spending. To fund purposes. Moreover, the focus and auxiliaries. Even at the aggregate implement technological innovations on revenues also means that relatively national level, the Delta Cost Project that can improve both efficiency and little attention has been paid to what measures rebut some commonly held effectiveness, leaders must be guided institutions do with the revenue and assumptions about college finances, in their efforts by a strong understand- what proportions they spend on fac- in ways that should be instructive for ing of the impact of the innovations ulty salaries or employee benefits or efforts to improve efficiencies through on both costs and revenues, as well as student services or admissions or any- technology. The project findings yield on learning outcomes. Without this thing else. It is important not only to insights into spending rates, cost areas, understanding, leaders are likely to fol- refocus attention on college spending cost per student, and cost per degree. low the usual model of innovation in but also to understand the constituent 1. Prices are going up faster than spend- higher education: implementing pro- elements of costs on both a per-student ing. The rise in college tuitions is one of gram add-ons, which are sometimes and a per-degree basis, since that is the the most obvious manifestations of the successful and sometimes not but only way to determine whether alterna- broken higher education cost model. which inevitably increase costs rather tive investments are cost-effective. Over the last twenty years, sticker than replacing or reducing them and prices have consistently increased at a ultimately fail to take hold in ways that Insights from the Delta Cost Project far faster pace than general inflation.1 will leverage systemic improvements. To help shift the higher education This is particularly the case among finance discussion to college spend- public four-year institutions (see Fig- A New Focus on Spending ing, since 2007 the Delta Cost Project ure 1), which enroll more than 40 per- Getting a better handle on college costs (http://www.deltacostproject.org) has cent of all full-time undergraduates.2 requires a new focus on institutional developed data and metrics focused Increases in public institution tuitions spending, buttressed by analytics that on spending in public and non-profit have been driven by a number of fac- allow institutions to look at spend- higher education institutions. Working tors, including (1) state appropria- ing in relation to outcomes (not just with existing data reported by institution declines, which have been fairly inputs) and that permit them to bench- tions to the federal government—pri- dramatic in recent years, and (2) the mark costs by examining spending marily data from the annual Integrated increased practice of tuition discount- patterns over time and in comparison Postsecondary Education Data System ing, whereby institutions provide aid

14 EducausEreview sEptEmbEr/OctObEr 2012 Technology and the Broken Higher Education Cost Model

to students in order to lower the posted 2. Nearly half of spending goes more part-time and adjunct sticker price. Concerns about these for overhead. A common belief is instructors to accommodate rising tuitions are the trigger point for that the biggest driver of college the growing enrollment de- public and political critiques of higher spending is faculty salaries and that mand. The marginal cost to add education: opinion polls show that the changing from classroom to distance- new students in courses taught by public believes tuitions are increas- mediated instruction will therefore part-time faculty is considerably lower ing because institutions are spending save money because it will allow for than average costs, potentially mak- more.3 Yet college costs—spending savings in labor costs. Yet even though ing this an even more cost-effective per student—have actually been flat in spending on instruction—for faculty way to deliver instruction than online many institutions and have declined in salaries and departmental support— delivery. Of course, the quality of the others.4 Although there are price/cost may be the single largest area of spend- online course might be superior to that gaps in almost all types of institu- ing in higher education, somewhere obtained in a large lecture hall, staffed tions, the gap is largest in public between 40 and 50 percent of general with low-cost teaching assistants. But community colleges, where tu- spending (exclusive of sponsored re- again, distance-mediated instruction itions have increased on average search and auxiliaries) goes for some may not be less expensive, and might by nearly 40 percent in the form of what can be called overhead or well be more expensive, than the last ten years, against spend- indirect costs.6 These cost areas range lowest-cost entry-level courses. ing per student, which has from academic and student support 3. Lower costs per student do not trans- decreased over this same (for counseling, computing centers, late into lower costs per degree or outcome. period.5 The simple reason and libraries) to operation and main- Higher education costs—that is that state subsidies have de- tenance of grounds and buildings, is, the amount of money clined, forcing institutions to increase maintenance, and utilities. Delta Cost spent per student—are tuition in order to offset the losses. Project trend analyses show that the considerably higher in Confusion about prices and costs is proportion of spending going to pay research universities not confined to the general public or for the direct cost of instruction has than in compre- even to policy-makers, however; many been largely stable or even declining hensive colleges, administrators and faculty within col- in most institutions over the last few which are more leges and universities also are unaware years, largely because institutions have costly than commu- of their real spending patterns. been saving on faculty costs by hiring nity colleges.7 This is because lower-division education has historically been less expensive FIGURE 1: Changes in College Sticker Price against Other Consumer than upper-division or graduate educa- Areas, 1999–2011 tion due to the smaller class sizes and greater specialization of curriculum 90% I Health at the more senior levels. Many states Insurance have chosen to limit access to higher-

I Public, 4-yr cost institutions, opting instead to 70% enroll the majority of students in less- expensive community colleges as a I Medical Care way to more efficiently get students to their educational goals. But lower-cost 50% I Prescription community colleges often turn out to Drugs have the highest cost when measured I Private, 4-yr against degree performance, simply 30% because of the high rates of attrition in community colleges. Nearly half of instructional spending in community I Public, 2-yr colleges goes to students (and credits) 10% that do not attach to a degree or cer- I New Car J tificate. Unless an institution is effective in transitioning I Median Family Income I Housing -10% students to a degree, lower 99–00’00–’01 ’01–’02 ’02–’03 ’03–’04 ’04–’05 ’05–’06 ’06–’07 ’07–’08 ’08–’09 ’09–’10 '10–’11 costs per student may not

16 EducausEreview sEptEmbEr/OctObEr 2012 Technology and the Broken Higher Education Cost Model

translate into lower costs per degree. the types of information needed so that Metrics on Costs by Discipline, Level of 4. If higher education is to be more costhigher education can reduce unneces- Instruction, and Enrollment Status effective and efficient, the unit of analysis sary spending and put increasingly lim- Although the IPEDS data from which the needs to shift from cost per student to cost per ited resources into areas that pay off in Delta Cost Project database is built does degree. To know what types of instruc- quality. More granular data about costs not provide information on spending tional delivery are most cost-effective, will help educators and policy-makers by individual discipline, level of instruc- we should be looking at the trade-offs learn more about efficiencies and de- tion, or enrollment status (full-time or between different types of invest- termine how technology might best part-time), individual institutions and ments and at the translation of be leveraged to increase both op- state policy-makers need cost informa- credits to degrees. For some portunities and effectiveness. No tion at these levels to determine both types of instruction, including one has the time or the need to where their money is coming from and distance learning, spending immerse themselves in elegant where it is being spent. Educating stu- per student may go up, but if bet- analytics that ultimately do dents in certain fields (e.g., some of the ter retention and graduation rates re- not produce actionable data. sciences, which require extensive lab sult, cost per outcome will be lower. Future efforts to refine cost work and equipment) costs more than And some of the biggest savings may analysis in higher education educating students in other fields (e.g., well be to students and families, in the need to be guided by a highly the humanities). In addition, graduate form of shorter time to degree and bet- focused agenda that will get at the and professional courses are typically ter course scheduling and sequencing. most critical cost drivers to help both smaller in size than many undergradu- institutional and policy leaders make ate courses and generally require more Critical Cost Drivers better decisions about spending and senior faculty to teach, thus The Delta Cost Project data and analyt- ultimately to connect spending to making them more expen- ics are only scratching the surface of performance, including quality. sive to offer. And although

18 EducausEreview sEptEmbEr/OctObEr 2012 Technology and the Broken Higher Education Cost Model

assumptions are usually made about the level work—has resulted in the need for but there are a num- costs of educating part-time students, a range of student services. Some ber of other factors to who now account for 38 percent research has shown that the use consider. The recent of all enrollments in higher of various student services can recession and deep cuts education,8 these students are improve persistence, gradua- in state appropriations to most likely to enroll in com- tion rates, and transfers from public higher education munity colleges and typi- two-year to four-year colleges, have resulted in signifi- cally arrive with a num- but not enough is known cant deferred main- ber of educational about what the individual tenance to existing and social needs that services actually cost to pro- buildings on college must be addressed vide and, more important, campuses. Sightlines if they are to succeed. about which services are most has indicated that nearly Understanding all of these effective. 25 percent of public university costs at a more granular level is necessary space will be over fifty years for leaders to make decisions that can Capital and Maintenance Costs old by 2020.9 Buildings of this both cut costs and help students obtain A common assumption in higher educa- age that have not been renovated degrees. tion is that some of the biggest savings are at a high risk of failures in mechanical from technology will occur in facilities systems (e.g., HVAC and elec- Metrics on Costs per Unit of Student Service costs. Yet the relationship between tech- trical) and of general dete- The current push to produce more col- nology and facilities costs may not be as rioration of usable space. lege graduates—coupled with the fact straightforward as it initially appears. In addition, although that many students entering higher Online courses obviously mean less use many in higher ed- education are not prepared for college- of and less wear-and-tear on buildings, ucation assume

20 EducausEreview sEptEmbEr/OctObEr 2012 Technology and the Broken Higher Education Cost Model

student services, but we know very National Center for Education Statistics, little about their effectiveness and Institute of Education Sciences, U.S. Department if they indeed save money. As of Education, 2012), http://nces.ed.gov/ programs/digest/d11/. institutional leaders think about 3. John Immerwahr and Jean Johnson, Squeeze Play refining metrics on costs by dis- 2010: Continued Public Anxiety on Costs, Harsher cipline, level of instruction, Judgments on How Colleges Are Run, a report prepared by Public Agenda for the National enrollment status, and student Center for Public Policy and Higher Education, services, they need to under- 2010, http://www.highereducation.org/reports/ stand where technology fits in, squeeze_play_10/index.shtml. where technology works and 4. Donna M. Desrochers and Jane V. Wellman, Trends in College Spending, 1999–2009: Where Does does not work, and where tech- the Money Come From? Where Does It Go? What Does nology can save money. It Buy? (Washington, D.C.: Delta Cost Project, We have suggested areas in which additional data is needed, particularly at the institutional and state levels, but the most important point is that budget and spending decisions need to be based on data, not on rumor or public opinion or perceived impact.

Summary 2011), Figure 7, http://www.deltacostproject.org/ that the newer, high-tech buildings and The Delta Cost Project has drawn at- resources/pdf/Trends2011_Final_090711.pdf. classrooms need less attention than tention to the need for a new focus on 5. Ibid., Figure 16. older buildings, the opposite is more institutional spending and has provided 6. Ibid., Figure 7. 7. Ibid. often the case. The technical complex- a number of useful metrics for thinking 8. Calculated from Snyder and Dillow, Digest of ity of these newer buildings necessitates about how different types of institutions Education Statistics 2011, table 202, http://nces.ed regular maintenance and upkeep by spend money, about the relationships .gov/programs/digest/d11/tables/dt11_202.asp. 9. Author communication with Jim Kadamus, vice skilled technicians; if this upkeep is between costs and revenues, and about president of Sightlines. ignored, costly repairs can be required. the declining public subsidies in public 10. I. Elaine Allen and Jeff Seaman, Going the higher education. However, the Delta Distance: Online Education in the United States, 2011 (Babson Park, Mass.: Babson Survey Research Technology Costs Cost Project data is just a starting point Group and Quahog Research Group, 2011), Another common assumption is that the for policy-makers and decision-makers. http://www.onlinelearningsurvey.com/reports/ wise use of technology will save money We have suggested areas in which ad- goingthedistance.pdf. 11. Kenneth C. Green, with Scott Jaschik and Doug in higher education by providing a ditional data is needed, particularly at Lederman, The 2011 Inside Higher Ed Survey of range of services, including instruction, the institutional and state levels, but the College and University Business Officers (Washington, more efficiently. Almost one-third of all most important point is that budget and D.C.: Inside Higher Ed, 2011), table 19. students in higher education have taken spending decisions need to be based on data, not © 2012 Rita Kirshstein and Jane Wellman. The at least one online course, and about on rumor or public opinion or perceived impact. text of this article is licensed under the Creative two-thirds of academic leaders believe In an Inside Higher Ed survey of college Commons Attribution 3.0 Unported License (http:// that students’ learning in online courses and university business officers in 2011, creativecommons.org/licenses/by/3.0/). is similar to or even better than their only 39 percent of the respondents indi- learning in face-to-face classes.10 We cated that their institution was effective Rita Kirshstein (rkirshstein@ know less, however, about what types of in “using financial data to aid and inform air.org) is the Director of students are most likely to benefit and campus decision-making.”11 Better data the Delta Cost Project and what types of courses are most success- is clearly needed. But even more impor- a Managing Director at the fully offered as online courses. And with tant, better use of data is mandatory if American Institutes for the widespread use of adjunct faculty higher education is going to fix its cur- Research. and graduate assistants to teach lower- rent, broken cost model. n Jane Wellman (jane@nash-dc level courses, per-student instructional .org) is the Founding Director costs for online courses will likely not Notes of the Delta Cost Project and be that much lower than the already 1. The College Board, “Trends in College Pricing the Executive Director of low unit costs for these entry-level 2011,” http://trends.collegeboard.org/college_ the National Association of pricing/indicator/index. general-education courses. Technol- 2. Thomas D. Snyder and Sally A. Dillow, Digest System Heads. ogy is also being used to provide some of Education Statistics 2011 (Washington, D.C.:

22 EducausEreview September/October 2012