HEFCE

HEFCE Research Series M12/97

July 1997

International Comparison of the Cost of Teaching in Higher Education

A bench-marking study conducted for the National Committee of Inquiry and the Higher Education Funding Council for England

The electronic version of this document contains the Contents and Executive Summary only. The complete printed document is available from the HEFCE, price £7.00.

EXECUTIVE SUMMARY
1.	INTRODUCTION
	The study
	Study method
	Outputs of case studies
	Note on data quality
2.	COMPARISON OF THE NATIONAL SYSTEMS OF HIGHER EDUCATION
	Introduction
	Similarities and differences from the English system
		Australia
		Germany
		Netherlands
		USA
	Common issues in HE on management and funding
	Some statistics
3.	THE CASE STUDY INSTITUTIONS
	The case study institutions
	Key characteristics

4.	COSTS OF UNDERGRADUATE TEACHING IN CASE STUDY INSTITUTIONS
	Introduction
	Strategic cost analysis method
	Step 1: Attribution of costs to undergraduate teaching
		importance of UGT in the institution
		balance of effort between T & R
		different costs of Bands
		relative resource burden of years in a course
	Step 2: Identification of relevant student population
		proportion of part-time students in the institution
		attrition rates
	Step 3: Calculation of cost per programme
5.	FACTORS INFLUENCING THE ANNUAL COST OF A STUDENT
	Nature of the cost information used
	Cost structures
	The lowest cost institutions
	The higher cost institutions
	The institutions broadly comparable with the benchmarks
	Cost factors and other differences
		staff
		estates
		central departments
		student services and support
		educational factors
		student population
6.	OTHER CONSIDERATIONS
	Government funding of teaching costs
	Key sensitivities
		attrition
		split of academic time and costs
		other areas of subjectivity
		conclusions on sensitivity
	Conversion to £ sterling
APPENDICES

A	Terms of reference
B	Tables
	1	Institution characteristics
	2	Annual cost of a student
	3	Cost of a graduate
	4	Factors leading to cost differences with benchmark
	5	Activities in academic departments
	6	Institutional cost structures
	7	Income

In a separate volume:

C	National profiles (Australia, Germany, Netherlands, USA)
D	Summary institutional reports

Summary

Overview

1. We were commissioned to undertake an international benchmarking study of the cost of producing a graduate in five countries including England. The aim of the study was to understand and explain differences in the cost of a typical undergraduate programme in these countries. We based our work on case studies at ten institutions: one typical research university and one primarily teaching university in each country. In each, we looked at both classroom-based (Band 1) subjects and laboratory or workshop subjects (Band 2).

2. Our work shows that the average cost of producing these non-medical graduates is lowest in the Australian institutions and is highest in the Dutch and the US institutions. Costs in the English and German institutions generally lie between these.

3. For a Band 1 graduate, the cost varies from about £8,250 in the Australian research university to £29,000 in the private research university in the USA. In all countries except the USA, the costs of a Band 1 graduate are slightly lower in the research institution than in the primarily teaching institution where teaching makes up a larger proportion of total institutional costs.

4. The equivalent comparisons for Band 2 (laboratory-based subjects) are less robust, but follow a similar order of rank, varying from about £14,000 in the Australian research university to about £45,000 in the Dutch counterpart. The relative costs of Bands 1 and 2 varies significantly between countries and types of institution. In most institutions, a Band 2 graduate costs between 1.3 and 1.8 times the cost of a Band 1 graduate. In the USA there is little difference between the Bands. In the German and Dutch research institutions the difference is much greater: the Band 2 cost is about 3 times Band 1.

5. Our work indicates that the factors particularly associated with the relatively low Undergraduate teaching costs in the English institutions are:

the short length of the degree programme
the low attrition rate of students (who fail to complete).

6. It could be argued that these factors are as much associated with national attitudes to higher education as with institutional management or efficiency. We have therefore also looked in some detail at the variations in the base cost per student year across these ten institutions, and the factors which influence this.

7. The results quoted above are calculated using Purchasing Power Parity exchange rates to compare between countries. However, the general thrust of the conclusions not sensitive to the exact exchange rates used.

The study

8. Our method involved case studies at ten higher education institutions (two per country) chosen to be reasonably representative of the two main types of university in England-. a research and teaching institution on the one hand-, and a former polytechnic institution on the other.

9. Although each country's system of higher education is different, it was not difficult to find institutions which broadly corresponded to these two types in Australia, Germany and the Netherlands. In the USA, the system of HE is probably more different from that in the other countries, but we chose one leading research university (which happened to be a private university) and one public institution with a particular focus on teaching.

10. Our method of study was a top-down strategic cost analysis. We used different detailed approaches in different countries, but they all followed a series of standard steps to arrive at the cost of a typical undergraduate programme. By its nature, this method can be expected to indicate, and show the reasons for, significant differences, but is less reliable if the aim is to quote absolute figures for the cost of any particular programme

Comparisons

11. We found that all five national systems were subject to pressure on resources and most of the institutions we visited had experienced the need to reduce unit costs and to generate additional sources of income where they could.

12. Our remit was to compare the cost of a typical graduate in each system. This is of course not comparing like with like: the main differences being in the level of achievement of applicants before they enter HE (which is generally high in England) and the nature of the programmes of HE which vary widely in length and specialisation.

13. Although we sought institutions that were as comparable as possible, there was wide variation amongst institutions (as well as national systems). Their size varied from about 5,000 students to about 25,000 and their income appropriate to their main operations from about £20m to £230m. However, they do fall into two groups for benchmarking.

14. The degree of public funding of institutions for the cost of Undergraduate teaching varies widely amongst our sample. As a proportion of the total cost of Undergraduate teaching in these institutions, public funding was zero in the private US university; about one third in the public US university; and over two thirds in all the others.

Annual costs of a student (see Figure 1)

15. The annual cost per full time equivalent (FTE) student is a useful figure for benchmarking, although it does not represent the true cost of teaching which should also take account of factors such as programme length and attrition (these are included in the cost of a graduate discussed below). Our calculations of the annual costs per FTE student show the following.

a. The average combined (Bands 1 and 2) annual cost per student varied from about £2,500 in the German research institution to about £6,300 in the US private research university.

b. The cost in Band 1 follows a similar pattern from £1,600 in the German research institution to about £6,300 in the US private research university.

c. The average cost per year of a Band 2 student varies from just over £4,200 in both Australian institutions and the German and Dutch primarily teaching institutions, to £9,500 in the Dutch research university. It is higher than the Band 1 cost in all institutions except in the USA where there is effectively no difference (associated with the nature of the US degree programme).

d. As a percentage of the Band 1 cost at the same institution, Band 2 varied from effectively the same cost (100%) to about three times as expensive (300%). The differences between the Bands were smallest in the USA where the concept was hardest to apply (most students take a mix of arts and science subjects) and were strongest in the Dutch and German research institutions.

Costs of a graduate (ie costs of a typical programme: see Figure 2)

16. The cost of a graduate has to take account of the actual (as opposed to standard or nominal) length of a course and of attrition (students who fall to complete the course). The length of programmes and attrition rates varied widely. Programme lengths varied from 3 years nominal and actual in England-, to four years nominal with actual completion of 5.5 in one Dutch institution, and 4.7 years nominal with actual completion of 6 years in one German institution.

17. An important part of our approach was to adjust the FTE student numbers in the different institutions to a standard basis which we called true FTES. In the Netherlands and Germany students which institutions described as full-time nevertheless take longer than the standard length to complete a degree programme. In resource-demand terms, we would therefore describe these students as part-time and this is shown in the table of institution characteristics in Appendix B. Our calculations allow for this by converting all student numbers to true FTES.

18. Attrition (net loss of students before completion) varied from 9% to 45%. The English attrition was at the lower end of this range. In most other countries the loss of first year entrants was partly off-set by movement into the institutions part-way through a programme as part of generally more flexible systems.

19. When the costs of completion times and attrition are taken into account, our calculations of the cost of a graduate show the following.

a. The average combined (Bands 1 and 2) cost of a graduate varies between about £11,000 in the Australian research university to about £29,000 in the US research university.

b. For Band 1 graduates, the range is from about £8,250 in the Australian research university to £29,250 in the US private research university.

c. For Band 2, the pattern changes significantly. The lowest cost institution is still the Australian research university, but the highest costs are seen in the Dutch and German research institutions. The range is from £13,600 (Al) to £45,400 (NI).

d. For Band 1 alone, the research institutions have lower costs than their primarily teaching counterparts in all countries except the USA where the comparison is complicated by the private/public split. For Band 2 the position is largely reversed with the research institutions having higher costs in all countries except the USA. A notable feature is the very high costs of Band 2 graduates in the Dutch and German research universities.

Main factors influencing costs

20. We review the main factors that account for these differences in chapter 5. The most significant are:

programme lengths (nominal and actual);
attrition (students who incur costs but do not complete);
staff costs;
staff:student ratios;
(related to SSRs) teaching methods - eg large classes vs seminars;
non-staff costs in academic departments;
central costs.

21. There are also non-standard costs which affect some institutions. Examples include unemployment provisions in the Netherlands and tuition fee remission paid to students in the USA.

22. It is also notable that the research institutions have much more flexibility to manage their costs of undergraduate teaching - chiefly through the allocation of academic staff time. In all countries except the USA they have significantly lower costs in Band 1 and significantly higher ones in Band 2.

23. The student experience and the philosophy of higher education differs widely amongst countries. We believe that our results are indicative of the most likely cost of a typical graduate in the ten institutions reviewed. However, we have not looked at the outputs of these programmes and the nature of their graduates differs according to factors which are more national than institutional in origin.

General observations on the method

24. The discussion of sensitivity and exchange rates in chapter 6 shows that the results we have quoted above are robust, at the level at which we have made comparisons. However, there are a number of aspects of our method, or of the data available which might be noted as imposing limitations which could be studied in more detail in any further work of this type. The main points are as follows.

a. The split between Band 1 and Band 2 was fairly arbitrary in some institutions and was difficult to make in the USA and Germany.

b. In all Institutions except the German, it was possible to obtain complete and reliable data on institutional costs to the level of individual academic departments.

c. The split of academic staff time between teaching and research is an important cost driver and yet is not easy to measure or estimate on a reliable and consistent basis. Giving any information on the split proved to be a sensitive issue for some institutions and many institutions did not have reliable statistics.

d. Most institutions have good statistics on student numbers but these do not typically show attrition in the way required for calculating costs.

25. A final point is prompted by the question "how typical are these institutions?". The institutions regarded themselves as not untypical in terms of the bottom-line cost of their undergraduate teaching. Where possible we confirmed this by reference to published national bench-marking data relevant to the costs of student teaching. However, in a more detailed study it would be desirable to do more work on the range of costs within these different national systems of higher education, as well as between them.