View in EDS HTML Full Text

A Waste of Resources? Social Rates of Return to Higher Education in the 1990s.

Saved in:
Bibliographic Details
Title: A Waste of Resources? Social Rates of Return to Higher Education in the 1990s.
Language: English
Authors: Ashworth, John
Source: Education Economics. Apr 1998 6(1):27-44.
Peer Reviewed: Y
Page Count: 18
Publication Date: 1998
Document Type: Journal Articles
Reports - Evaluative
Descriptors: Economic Factors, Educational Development, Foreign Countries, Higher Education, Human Capital, Program Effectiveness, Resource Allocation
Geographic Terms: United Kingdom
ISSN: 0964-5292
Abstract: Demonstrates that popular assertions regarding the social benefits of additional higher education are dubious. Reveals four decisive factors: presumed economic growth; changes in graduates' and nongraduates' relative earnings; differences between the marginal and average student; and belief in scale economies. Unless economic growth favors graduates, society's reward is too low to justify further expansion of higher education. (100 references) (MLH)
Entry Date: 1999
Accession Number: EJ567344
Database: ERIC
Full text is not displayed to guests.
FullText Text:
  Availability: 1
  Value: <anid>AN0000779665;EDE01APR.98;1998Jul07.13:14;v2.3</anid> <title id="AN0000779665-1">A WASTE OF RESOURCES? SOCIAL RATES OF RETURN TO HIGHER EDUCATION IN THE 1990S </title> <p>ABSTRACT </p> <p>There has been considerable debate concerning the benefits or otherwise of the 1990s' expansion of higher education in the UK yet there has been no attempt to examine the human capital underpinnings of continued expansion from the perspective of society. This paper demonstrates that the popular assertions regarding the benefits to society of additional higher education are now very dubious. Four factors are shown to be decisive: (presumed) economic growth; changes in the relative earnings of graduates and nongraduates; the difference between the average and the marginal student; the belief in scale economies and therefore the costs of higher education. Of these the most important is the presumption made about growth and the distribution of the rewards from growth. The paper shows that, unless the growth is such that it favours graduates, society's reward is so low that further expansion is difficult to justify on economic growth terms. There would appear to be no substantive rationale for a further increase in student numbers until there is some evidence of a social return on the initial investment. </p> <hd id="AN0000779665-2"> Introduction </hd> <p>There has been considerable debate concerning the benefits or otherwise of the 1990s' expansion of higher education in the UK (see, for example, CBI, 1994; Johnes, 1993b; Keep & Mayhew, 1996; Murphy, 1993, 1994), yet, despite its significance, there has been no attempt to examine the human capital underpinnings of continued expansion from the perspective of society. </p> <p>What is important in the debate is the analytical distinction between the stock of graduates and the future flow. Johnes (1993b) correctly makes the point that it is the future graduates and their impact on the economy which is of importance, not the present stock. However, this aspect seems to have been overlooked in some discussions of the expansion and financing options of higher education. Commentators, for example a leading member of the Confederation of British Industry (CBI) (THES, 1995a), continue to use returns enjoyed from the present stock of graduates from past operations of the higher education system and project these as likely to be enjoyed in the future without any consideration of the assumptions which underlie such an assertion. This paper demonstrates that the popular assertions regarding the benefits to society of additional higher education are now very dubious without a clear statement of the conditions which are required to support the assertions. Four factors are shown to be decisive: (presumed) economic growth; changes in the relative earnings of non-graduates; the difference between the average and the marginal student; and the belief in scale economies and, therefore, the costs of higher education. Further expansion should only be contemplated with these factors completely clear. </p> <p>This paper considers two groups of students, those who become the average graduates and those who are the marginal graduates. This subdivision is crucial to the analysis because by identifying different groups, this paper shows that society's reward may already be so low that further expansion is difficult to justify on economic growth terms. Overall, the results suggest that the only rationale for the expansion is a fervent and widely held belief that the investment will lead to sufficient and greater growth than could be generated by alternative investment and that, even with that belief, the expansion cannot go further unless the growth rates generated are beyond the usual expectations or are such that the returns are unequally spread. Without this, society must believe that the immeasurable social benefits of more higher education are sufficient to outweigh the low rates of return on what is 'measurable'.(<reflink idref="bib1" id="ref1">n1</reflink>) There would appear to be no substantive rationale for a further increase in student numbers until there is some evidence of a social return on the initial investment. </p> <hd id="AN0000779665-3"> The 'Problem' of Rising Numbers in Higher Education </hd> <p>Murphy (1993) suggests that society, due to an over reliance on the assertion that investment in human capital causes growth, has wasted resources by expanding higher education. Keep and Mayhew (1996) also cast doubts, if not necessarily on the expansion which has already occurred, then on the wisdom of continued expansion without a serious consideration of other options. As Keep and Mayhew (1996) note, at least in numerical terms, current graduate output more than satisfies the demand for graduates to fill traditional graduate jobs.(<reflink idref="bib2" id="ref2">n2</reflink>) Indeed, there is some evidence that even prior to the expansion of the 1990s there was over-education in the UK. Using 1986 data, Sloane et al. (1995a,b) reported that 30% of the graduates in their sample see themselves as over-educated for their job. However, as they demonstrate, in terms of productivity, some of this over-education is mitigated by lack of experience and education may have led to more efficient screening. Nonetheless, there must be concern that an expansion of higher education since 1990 has exacerbated the figure and there is a clear potential for society to be receiving a poor return on its investments.(<reflink idref="bib3" id="ref3">n3</reflink>) </p> <p>Of related concern is the question of how to fund the expansion. The governments of 1979-1996 attempted to hold down the cost of higher education in a number of ways. While overall spending on higher education has increased, the expansion has also been financed within the universities themselves and by the students. Firstly, the pay of the sector has been held down. Academic pay in the 'old' universities over the period has risen in real terms by only 3%, compared with over 40% for other non-manual earnings sectors (see Association of University Teachers (AUT), 1995). The effect of this is to increase the social rate of return, thus making society more willing to see the expansion occur.(<reflink idref="bib4" id="ref4">n4</reflink>) Secondly, the maintenance grant to students has been cut in real terms and been replaced by a grantiloan mix. For the purposes of the analysis in this paper, this will be treated as a transfer payment and, thus, there is no effect in the long term on the returns to society, although discounting may have some small effects. Thirdly, the unit of resource to universities, particularly for arts and humanities subjects, has been cut. Thus, although more students are engaged in higher education, more students must be recruited by universities to keep their income constant. Were it the case that economies of scale existed, this will tend to increase the social rate of return on students. </p> <hd id="AN0000779665-4"> Social Rates of Return for Graduates </hd> <p>Typically, social rates of return are viewed as being difficult to calculate and, thus, have been explored less widely than their private counterparts.(<reflink idref="bib5" id="ref5">n5</reflink>) In the general folklore associated with higher education in the UK, the social rate of return has been viewed as being around 7%, although there is variation with subject discipline. Indeed, a figure of 8% was used by the government when it introduced its proposal for loans to replace grants (Department of Education and Science (DES), 1988); and much was made of the large gap between this and the private rate of around 27.5% The figures used were calculated for the cohort of graduates as of 1988 but, crucially, no attempt has been made to investigate the social rate of return since the 1990s' expansion.(<reflink idref="bib6" id="ref6">n6</reflink>) </p> <p>The usual way in which rates of return are calculated is to employ survey data and then to infer alternative earnings from other members of the cohort. Clearly, this is not possible when engaged in a predictive exercise and, anyway, it falls directly into the trap, emphasized by Johnes (1993b), of using the stock data when the flow (or the stock in a later equilibrium) is what matters. The approach adopted here is to simulate the position of representative students.(<reflink idref="bib7" id="ref7">n7</reflink>) The initial analysis relates to an average graduate and then the marginal graduate is considered. The analysis involves a series of assumptions being made, and the aim of this paper is to demonstrate that, provided they are applied consistently, there are clear implications from the analysis which can be brought to bear on the question of the likely social rate of return. Naturally, any results are indicative only, but they will serve to illustrate what makes most difference to the outcomes when using the human capital perspective. </p> <p>A number of different returns are considered: </p> <p>the social rate of return to the average graduate before expansion (the 1990 position); </p> <p>the social rate of return to the average graduate after expansion (the 1995 position); </p> <p>the social rate of return to the marginal graduate after expansion (the 1995 position). </p> <p>All results regarding social rates of return are given in Table 1 and the details of the various calculations made can be found in the appendix. </p> <p>The Social Rate of Return to the Average Graduate Before Expansion </p> <p>The first stage in the analysis is to calculate the social rate of return under the conditions which pertained immediately prior to the large increase in student numbers; the 1990 position. Ibis calculation assumes: (i) that approximately 15% of the cohort of 18-year-olds are students and become graduates; (ii) a 2% growth rate; (iii) a graduate unemployment rate of 4%; and (iv) an unemployment rate for the rest of the population of 10%; the last three following the DES (1988). A critical additional assumption relates to the size of the alpha factor, the proportion of the enhanced income stream from being a graduate assigned to education, rather than inherent ability. For this analysis, the values of 1.0 and 0.6 are used but these are clearly illustrative. As can be seen from column I of Table 1, a social rate of return of 9.7% emerges if the alpha factor is unity and of 6.8% if the alpha factor is 0.6. These are marginally higher figures than those usually found; the alpha factor adjusted return found by the DES (1988) was 5.5%. While they must be treated with caution, it must be remembered that by 1990 university costs per full-time equivalent student had fallen from the levels when previous calculations were made.(<reflink idref="bib8" id="ref8">n8</reflink>) In this paper, for consistency, all assumptions made in the analysis aim to place all other potential biases in the same direction, Thus, the figures presented here aim to provide the maximum possible return and all comparisons must be viewed in that light. </p> <p>The Social Rate of Return to the Average Graduate After Expansion </p> <p>It is now necessary to analyze the effects of increasing the graduate cohort (i.e. changing those who leave education with A-levels into graduates). In the initial stage, following a job-competition model (see Thurow, 1975), employers will substitute graduates for A-level workers, regardless of the job requirements, in what have previously been the best paid A-level jobs. Sloane et al. (1995b) report evidence that, in 1990, 20% of new graduate jobs were suitable for A-level holders and, in two-thirds of the cases, the requirements of the jobs were unchanged. The crucial feature of the shift is that this will lead to changes in the average and, thereby, relative earnings of both the graduate and non-graduate groups. Given that calculations of social rates of return use gross salaries as a measure of marginal productivity (and, thereby, worth to society), these changes are crucial. Commentators, when considering graduate prospects, have identified lower starting salaries of some graduates (for example, Mason, 1995, 1996; IES, 1995); yet, despite evidence from the US (Freeman, 1995), the likely long-term changes in non-graduate income have tended to be ignored. Johnes (1993b) notes that in an adjustment period, graduates may remain unemployed as they search for traditional graduate jobs and pay,(<reflink idref="bib9" id="ref9">n9</reflink>) but in the longer term, a large exogenous increase in graduates must mean that graduates will take the jobs of non-graduates.(<reflink idref="bib10" id="ref10">n10</reflink>) The nature of the jobs may or may not change depending on the production techniques required for specific groups of labour. If workers are hired for jobs for which they are over-educated but their levels of pay and productivity are identical to those of workers in those jobs with the lower required level of education, it is not in society's interest for the investment in higher education to have occurred. If there is merely a job-competition model in place, society is indifferent as to who fills the various jobs and their background before arriving at those jobs. Society will obtain a return only if education improves the productivity of the workforce so that there is economic growth. </p> <p>In this paper, the conventional view of human capital is emphasized, and it is presumed that greater education will lead to greater productivity which will be reflected in growth. Thus, it is assumed that a strict job-competition world, with its associated over-education, is, at most, a short-term phenomenon and that growth is generated. The role of the enhanced productivity and, therefore, the assumed growth rate (usually around 2%) is vital to the calculations in all social rates of return. Importantly, there can be no return to society without growth.(<reflink idref="bib11" id="ref11">n11</reflink>) As Weale (1993) notes, education may identify people of differing ability for employers without adding much. However, even if screening is the only purpose of higher education, then, if there is an efficient screening process, more appropriate people are in the various jobs which will also lead to some economic growth, although how much is attributable to higher education is a matter of debate.(<reflink idref="bib12" id="ref12">n12</reflink>) </p> <p>In order to consider the problem within a rates of return framework, a distribution of income is needed and the details of this are presented in the appendix. The results of the analysis are given in column 2 of Table 1. It is important to note that, aside from distributional matters, society is indifferent as to who pays the agreed bill for the universities to operate-be it the students themselves or the government on society's behalf.(<reflink idref="bib13" id="ref13">n13</reflink>) What is important (for the calculations) is the cost of the higher education; a matter returned to later. Hence, with the same assumptions regarding growth rates and unemployment as above but adjusting the costs of the higher education to reflect the changes in funding of universities, the social rate of return can be seen to remain around the same as before the expansion. What should be noted is the effect of assuming a fall in the alternative earnings stream. If alternative earnings were assumed to remain the same (before applying the growth factor) then the rate of return falls. This will be the social rate of return from the 1995 graduates, who are in the interim period and could have gained the higher alternative wages, but this position does not remain in the long term with the increase in graduates displacing some workers. The figures illustrate that the expansion, with a suitable adjustment in the costs, appears to leave society receiving the same rate of return per graduate. </p> <p>Thus far, a rate of growth of 2% has been presumed with the added assumption that the gains from growth have been equally shared among all members of society.(<reflink idref="bib14" id="ref14">n14</reflink>) Clearly, this assumption has a great bearing on the analysis. During the 1980s, (gross) earnings inequalities rose sharply in the UK, probably due to a rise in the relative demand for skilled labour.(<reflink idref="bib15" id="ref15">n15</reflink>) With a large expansion in graduates, it is a matter of debate as to whether the skilled labour deficiency has been removed. It is well documented, see Gosling et al. (1996), that the ratio of earnings in the 90th/50th percentile has risen (in the same way as the 50th/10th percentile). Moreover, as noted by Gregg and Machin (1994), increases in the supply of educated workers has invariably been allied to rises in returns to schooling (for the US, see Katz & Murphy, 1992; Murphy & Welch, 1992). However, the changes in the number of graduates in the UK are very large and have occurred in a short time period, hence the large increases in graduates might offset the movement in inequality seen in the 1980s. Indications from Wilson (1995) are that the deficiency in skilled labour will be removed in the near future. </p> <p>However, while it may not be unreasonable to suggest that growth will be more equally divided between higher earners than has been the case in the recent past, there is a case for unequal gains to be considered. To this end, a number of different possibilities of differential growth in graduate and non-graduate earnings are considered and given in column 2 of Table 1. It can be seen that unequal growth enhances the returns in all cases. </p> <p>Whatever assumptions are made, the rate of return remains a reasonable proposition, provided it was deemed as such initially. However, one could argue that with many more graduates, the differential in unemployment rates may not continue. This has been the case in Norway which has experienced a similar large rise in graduate numbers in the 1980s and 1990s (see Aamodt & Arnesen, 1995). To see the effect of this an unemployment rate of 7% is considered, retaining the A-level unemployment at 101/0, and it can be seen from Table 1 that this has little effect on the inference that society can view its investment in the average graduate as reasonable. </p> <p>The Social Rate of Return to the Marginal Graduate After Expansion </p> <p>The average graduate would have gone to university under the pre-1990 regime. The previous marginal entrant to higher education in a 15% graduate cohort world is the median graduate in the new regime. Thus, the return calculated in the previous section indicates that all the previous graduates are giving about the same return as before to society. In the pre-expansion world with 'elite higher education', the average and marginal graduates could be treated as being fairly close to one another. With the large expansion, this is no longer the case, which leads to the question of what is the position of the new students in higher education and, crucially, the last in--the marginal students. Is society wise to have agreed to these students taking part in the higher education process?(<reflink idref="bib16" id="ref16">n16</reflink>) </p> <p>Returning to the assumed income distribution, it is possible to consider the position of anyone within this distribution. Thus, it is possible to apply the job-competition model and re-allocate income between graduates and non-graduates as before and then apply any growth in the economy (or sectors of the economy) to the new positions. Continuing with the `naive' presumption that graduates are the most productive and, thereby, the highest paid workers, it is possible to envisage the marginal students lying at the 70% position in the distribution with 30% of the population earning more than them. The assumptions are outlined in the appendix and the results given in Table 1, column 3, identified as the `best marginal' student. The range of rates of return, using equally distributed 2% growth with 7% graduate unemployment, are 6.1% and 3.9%. The implication is that the return to the marginal graduate falls to, at best, three-fifths of what it was before the expansion. </p> <p>However, it is likely that there will be non-graduates with higher incomes than the lowest paid graduates.(<reflink idref="bib17" id="ref17">n17</reflink>) One such position is to assume that a further 5% of the population earns more than this group. This position is examined in column 4 of Table 1, labelled as `likely marginal' and, as anticipated, the rate of return falls further with a rate of return below half that of the former average student.(<reflink idref="bib18" id="ref18">n18</reflink>) All the above assumes that the marginal graduates have the same growth in earnings as their counterparts, who are non-graduates. In the case of marginal graduates, the assumption of the same growth rates can be deemed to be more likely due to the anticipated similarity in the types of occupations and individuals in them.(<reflink idref="bib19" id="ref19">n19</reflink>) </p> <p>Further, the presumption in the aforementioned is that the students who start the courses actually graduate, unless these are captured in the unemployment rate. There is considerable evidence (University Statistical Records, 1990, 1994) that the `drop-out' rate is rising in higher education and a great deal of this is occurring amongst the marginal students. Thus, the return on the investment as a whole is diminished unless the act of receiving some of the education is beneficial in enhancing growth.(<reflink idref="bib20" id="ref20">n20</reflink>) </p> <p>The interesting factor to note here is that the social rate of return from the marginal student may be greater than the private rate as perceived by this group of students. Ashworth (1996) has shown that the private rate of return with zero perceived growth and a loan as small as £4000 is only 1.81/0 for the marginal graduates. If society as a whole, as reflected by the median voter, is convinced of an education growth correlation but the individual student is not, then it can be seen that society will need to provide some subsidy to induce the students into higher education, though why they should do this for returns as low as those suggested here is not at all clear. Society must follow Murphy (1993) and question such investment. Indeed, the government, in a former green paper (DES, 1985), inferred that social rates of return of 5% were too low and the usual treasury discount rate is 5% (see Mallier & Rodgers, 1995). Clear presumptions about growth and the nature of that growth are needed to meet this requirement. </p> <p>The Fall in Unit Cost per Student </p> <p>The fall in the social rate of return for all the groups considered is not as great as it might have been because the unit of resource paid by the government (and, therefore, society) to universities has fallen and so the costs have fallen. This clearly raises the question of the quality of the output, as it has been presupposed that not only has this not diminished but that the proportion of the social return attributable to higher education has not diminished. </p> <p>As the returns to the provider of higher education have diminished, the standard human capital arguments will imply that the quality of academics will fall.(<reflink idref="bib21" id="ref21">n21</reflink>) If this were to lower the quality of the output and/or there is a fall in quality of output due to there being diseconomies of scale,(<reflink idref="bib22" id="ref22">n22</reflink>) then the social rate of return falls. There are two reasons to suppose a fall. Either higher education becomes more of a screening process than previously, in which case the alpha factor falls lowering the returns, or the productivity upon graduation falls, lowering initial salaries until the effects of further training become effective, also lowering the returns. </p> <p>Alternatively, if society observes a fall in the quality of the output and attributes this to either a fall in labour input quality or a fall in capital input this will raise costs, again lowering the social rate of return from the higher education sector. This ]ends further weight to the view that equilibrium student numbers may have been exceeded as society's returns become even lower once this is accounted for.(<reflink idref="bib23" id="ref23">n23</reflink>) </p> <p>Should Society Support the Expansion? </p> <p>To gain returns as low as 3% on the marginal group of graduates, society needs to be convinced of 2% growth in the economy as a whole, from increased participation. Any increase above this comes from making alternative assumptions about how the growth will manifest itself. Society also needs to be clear that the growth is coming from this source (a point that Levin and Kelly (1994), Murphy (1993), Maglen (1990) and Daly (1982) would dispute) and to be convinced that the fall in the unit of resource has not had a negative effect on the quality of the graduate output. </p> <p>Overall, the results here suggest that one's opinion of the higher education expansion of the 1990s and whether this has been a `waste of money' depends upon what is deemed to be a reasonable rate of return on capital investment. It would appear that, as expansion was undertaken with published social rates of return of 8% or 5.5% with the reasonable alpha factor adjustment, society may be satisfied with a smaller return than this, perhaps as low as 5%. Nonetheless, society would potentially appear to be receiving a very poor deal for their investment on the marginal group, especially if we follow Levin and King (1994) and take the view that education alone cannot generate the growth and "complimentary requirements are not in place". Further, Rumberger (1987) suggests that education which is required for the job is rewarded highly but excess education is rewarded at a lower rate. Additional education does not always raise productivity which would, in turn, intimate the possibility of even lower returns.(<reflink idref="bib24" id="ref24">n24</reflink>) </p> <p>The implications of the above are that while some expansion is defensible, there are clear limits to this. With 30% participation, the social rate of return on expansion graduates may have fallen to a very low level. </p> <hd id="AN0000779665-5"> Conclusions </hd> <p>This paper offers some support through human capital based techniques for the conclusions of Murphy (1993, 1994) that some of the expansion in higher education is a waste of resources. Two sets of rates of return have been calculated, those to the average graduate and those to the marginal graduate. While the returns to the average graduate have only been affected very slightly by the expansion, the social rates of return to higher education for the marginal graduate appear low enough to suggest that investment elsewhere would seem a better proposition. Indeed, the returns are only as high as they are because of the fall in the unit of resource to universities, which raises questions about the quality of the output or the proportion of the returns attributable to higher education. This paper has only touched on the issue of whether the standard of graduates has fallen, to what extent graduates can be considered as a homogeneous group or whether economies of scale with respect to undergraduate provision in universities may have been exhausted. Further, it has also avoided the issue of whether students will require 4-year degrees (or a Master's degree) to gain the `best' jobs. Any increase in training period lowers any social returns to higher education.(<reflink idref="bib25" id="ref25">n25</reflink>) Finally, it has avoided the more pessimistic views regarding very high levels of future unemployment for both graduates (see Wilson, 1995) and the rest of the workforce. </p> <p>Additionally, other than the identification of two groups of graduates, there is no consideration of differing returns to type of education. There have been claims that the expansion has led to greater diversity but there is no work (to the author's knowledge) of the type undertaken by Groot (1994) in the Netherlands and by Grubb (1992a, b, 1993, 1995a, b) and Kane and Rouse (1995a, b) regarding returns to types of education. As de Meulemeester and Rochat (1995) discuss, the mix of the higher education may well be more important than the amount. In addition, all the calculations and the whole expansion does really require the increased participation in education to generate associated growth, of which there is no guarantee (see Duncan & Hoffman, 1981; Levin & King, 1994; Lynch, 1993). All these factors reinforce the arguments that the social returns to the marginal students appear very low. </p> <p>While no claims are made that the values presented in this paper are perfect, they do give an indication of the relative returns under different circumstances. In particular, they show the enormous role played by assumed growth rates in the analysis. What is clear is that all groups in society need to be convinced of the effects of the increase in human capital on economic growth and, at least in the shorter term, acquiesce in a low return for the present expansion to have been reasonable. In order to gain returns in line with what are usually deemed to be acceptable, growth rates will have to rise considerably above those experienced before or be distributed in a very definite manner. While it is one of the presumptions that increased education will lead to greater growth rates, the evidence of the size of the growth rates is very patchy. Given that all the figures presented in this paper are liable to an upward bias, they provide support for the considerable doubts of Keep and Mayhew (1996) about the desire for even more expansion as espoused by the CBI (1994), certainly before the effects of the present expansion have worked their way through the system. </p> <hd id="AN0000779665-6"> Acknowledgements </hd> <p>I would like to thank Jim Coleman, Lynne Evans, Geraint Johnes, Jim Murphy, Denis O'Brien and an anonymous referee for commenting on earlier drafts and considerably improving the exposition. The paper also benefited from suggestions by participants at the DfEE Education and Economics Group meeting, November 1996. The usual disclaimer naturally applies. </p> <hd id="AN0000779665-7"> Notes </hd> <p>(n1.) If the evidence presented by Chatterji and McKaig (1995) regarding tertiary education being the vital ingredient for growth is followed then this might lead to further expansion. However, the results of McMahon (1987) and Pscharopoulos (1994) and most micro-economic rates of return work which indicate a less optimistic view must temper this argument. Further, the results of de Meulemeester and Rochat (199 5) suggest that there may be short-term causality between higher education and economic development but there is no real evidence of a long-term link between these variables alone; see Levin and Kelly (1994) for further details. </p> <p>(<reflink idref="bib2" id="ref26">n2</reflink>) Work by Wilson (1995) would suggest that there will be an over-supply of graduates by the year 2000. </p> <p>(n3.) This does presume that graduates are alike and degrees are homogeneous. An alternative view is that of Robst (1995), who suggests that attending a `lower quality university' results in an individual possessing less human capital. Thus, an apparently over-educated worker may not be over-qualified for the position. The additional education may be necessary for the person to possess a sufficient amount of human capital for the job. For the purpose of this paper, although two groups of graduates are identified, the view of graduates will be more in the traditional mould as opposed to the Robst mould. </p> <p>(n4.) For the present, this ignores any marginal productivity argument applying to the academics so lowering the value of the output, i.e. graduates are less good. However, as most staff are in post, and transfer and re-training is costly to them, even in the medium term, lowering academic salaries probably has little effect. Support for this view can be found in Dolton and Mavromaras (1994) and Dolton and van de Klaauw (1995), assuming that teachers' decisions can be mapped on to those of university lecturers. </p> <p>(n5.) Examples of social rates of return for a number of countries are presented by Psacharopoulos (1981, 1985, 1994). A recent example of work on private and social rates of return is given by Vaillancourt (1995) for Canada. A representative list of work on private rates is given by A Ashworth (1996). </p> <p>(n6.) Lissenburgh and Bryson (1996), using the Youth Cohort Survey, examine starting salaries and occupations in the early 1990s. However, as they point out, their work makes use of graduates from the higher education system before a full expansion. </p> <p>(n7.) A similar, more sophisticated and much more data intensive, micro-simulation approach for rates of return is reported by Chapman and Harding (1993) for Australia, following work by Chia (1990). </p> <p>(n8.) Cultural benefits are ignored in this analysis. Such externalities, although difficult to place a figure on, may increase the calculated rates of return. The author is unaware of any calculations of cultural components of educational benefit. However, it is unlikely that the externalities will be greater than the other inherent biases so the author believes that the figures are still a maximum. Indeed, if sound investments are required to yield 10%, then the DES (1988) return of 8% would suggest that over-education was already occurring in 1990 and that investment should have been made in physical, not human, capital (see Tsang and Levin (1985) for a discussion of this phenomenon). The returns to higher education may also have risen markedly during the 1980s due to the increase in income inequality owing to skill deficiencies (see Moll, 1992; Gregg & Machin, 1994; Gosling et al., 1996). Whether this is a long-term phenomenon with the increase in graduates is discussed later. </p> <p>(n9.) This follows from Jovanovic (1979), Rosen (1972) and Sicherman and Gabor (1990) job matching arid occupational mobility ideas. </p> <p>(n10.) Over time, various professions have adjusted their requirements. For example, the UK accountancy profession moved to graduate entry over approximately 20 years. Whether the desire for graduates is due to changes in the jobs requiring a more qualified workforce or merely screening (see Arrow, 1973; Layard & Psacharopoulos, 1974; Spence, 1973) is irrelevant for the purposes of this analysis. Further discussion of the changing nature of graduate jobs and the role of the push effects of higher education, as well as demand operations, can be found in Teichler and Kehm (1995). </p> <p>(n11.) It is important to see that any growth occurs after the initial redistribution of graduate/nongraduate jobs and to stress that without growth there can be no social return as there are merely transfers among members of society. </p> <p>(n12.) It should be noted that this provides the clear link between the micro-economic based human capital approach to returns to education and the macro-economic based work of Dennison (1974, 1979). Once it is clear that there is no return to society from investment in education without economic growth, then the link from education to growth becomes clear. If society obtains no or very low growth from its marginal investment in higher education, it should withdraw that investment from education to something else such as physical capital until it restores, an optimal mix for growth. However, as Levin and Kelly (1994) illustrate, education is unlikely to achieve growth on its own and so it is likely that to receive the returns "complimentary conditions must be in place". For the purpose of the calculations, it is presumed that the conditions apply to give the growth. </p> <p>(n13.) This circumvents the optimal size of the public grant and how that is determined. See Creedy (1994), Johnes and Johnes (1994) and Skidelsky (1995) for discussions of this matter. </p> <p>(n14.) The assumption does not completely imply this-only higher earning workers in the nongraduate group need to have the growth in earnings for the analysis presented to hold. </p> <p>(n15.) See Moll (1992) and Gregg and Machin (1994) for more on the reasons for this change in inequality in the 1980s. </p> <p>(n16.) The two groups in this paper may fit with the views of Soskice (1993) and Smithers and Robinson (1995) that more graduates can solve failings in craft and technician training. Traditional graduates are one group, with the marginal students filling technical skills gaps forming another group. </p> <p>(n17.) All the evidence on over-education and earnings (see, for example, Freeman, 1976, 1980; Hartog & Oosterbeek, 1988; Rumberger, 1987; Tsang & Levin, 1985; Verdugo & Verdugo, 1989) incidates that some individuals with lower qualifications than graduates earn more than them. </p> <p>(n18.) Clearly, this position is identical to having 35% graduates and assuming a `best marginal' position. Thus, the same ideas can be applied to further expansion so that 40% or more of the population in equilibrium as graduates can be examined; whatever the proportion, there will always be a marginal group. </p> <p>(n19.) The effects of differential growth rates are also shown in columns 3 and 4 of Table 1, with the expected increases in the returns, </p> <p>(n20.) It must be noted that there is no discussion here of the appropriateness of university relative to training per se or whether universities should train directly (see THES, 1995b). Further, the possibility of a huge rise in graduate unemployment, as forecast by Wilson (1995), is reported in the Financial Times, 12 February 1996. However, what happens to non-graduate unemployment is equally important for rates of return, Generally, as unemployment rises, even if the relative unemployment rates stay the same, rates of return fall. Recent evidence (from the US) is mixed on whether graduation is necessary to gain rewards (see Grubb, 1993, 1995a; Jaeger & Page, 1996; Kane & Rouse, 1995b). </p> <p>(n21.) There is sizeable literature on the economics of teaching supply (for example, Bee & Dolton, 1995; Dolton, 1990; Dolton & Mavromaras, 1994; Dolton & van der Klaauw, 1995; Murnane & Olsen, 1989; Zabalza, 1979; Zabalza et al., 1979) but much less consideration has been given to higher education, although a number of the arguments are similar. At present, the income of an average academic is always lower than that of an average graduate, hence there is a negative return to postgraduate students who become university academics. Although there is some recent evidence of vocational calling among scientists (Daily Telegraph survey, March 1996), if marginal productivity theory is strictly applied only below average students will now become university lecturers. Rudd (1990) and Tarsh (1992) have considered the issue of further degrees and concluded that, for the period of their studies, a pay disadvantage did come from further degrees relative to similar well qualified candidates. Tarsh, in particular, is clear that this disadvantage did not occur relative to average graduates, hence there has been a change over the 1980s and 1990s. While there is some evidence (see Johnes, 1993a) that, internal to the academic labour market, the pay for academics is not strictly determined by marginal productivity, this is riot the issue here as it involves the decision to enter the profession Naturally, matters may not be quite as severe as portrayed here due to the international nature of the academic market. `Brain drains' are commonplace and the costs much discussed (see, for example, Grubel, 1987). Thus, UK universities can gain from lower paying nations and, over the period 1980-1989, were a net importer of labour. Nonetheless, once the rigidities of the labour market (see note 2) have been removed, the long term would indicate a decline in the quality of one of the inputs to higher education. </p> <p>(n22.) While there is evidence (Glass et al., 1995a,b; Johnes, 1996) that prior to the expansion there were increasing economies of scale present in undergraduate teaching, Glass et al. (1995a,b) would suggest that doubling student numbers from those in place in 1989 and/or 1992 would remove all such economies of scale and replace them with diminishing returns. However, the concerns of Keep and Mayhew (1996) about the scope and sustainability of additional reductions in. tile unit of resources must be noted. </p> <p>(<reflink idref="bib23" id="ref27">n23</reflink>) If it is assumed. that there were no economies of scale to be had and the costs had remained as in 1990, the return for the average student would have fallen from 5.9% to 4.8%, with similar falls in the other figures. Crucially, for the likely marginal student, the return falls below 2% and, even with some of the advantageous assumptions about growth rates, close to 5%. </p> <p>(n24.) Verdugo and Verdugo (1989) suggest that the returns to over-education could be negative, although Cohn (1992) and Gill and Solberg (1992) give compelling reasons why this may be an overstatement. </p> <p>(n25.) Using the same assumptions, rates of return for 4-year degrees tend to fall by 1% and 2% depending on the assumptions made about growth. </p> <p>(n26.) A reasonable approximation for this should be the overseas fees charged by universities as these are supposed to represent the average cost of providing degree education. At present, this is £5775 for arts students and £7655 for science students. If this were true, a figure of 6600 would therefore not be an unreasonable figure to use for the cost of the education for each student to society. However, it should be noted that there is no imputed rent in this figure and so the costs may be low. Changing the costs has the effect of changing the rates of return by around 0.3% per thousand pound change in unit cost. </p> <p>(n27.) The use of future earning can only give indications and the cautionary notes of Eckhaus (1973) about such earnings, emphasized by Levin and Kelly (1994), must be acknowledged. Additionally, comparison across sexes is clearly very difficult. The cycle of earning for women has tended to be less than the figures above, although there is no reason this should continue to be the case. For the calculation of rates of return, it must be remembered that aspects such as child birth which affect the cycle of earnings are not just undertaken by graduate women and so there may be cancelling effects within sexes, if not across sexes. The purpose of this paper is to demonstrate the role of various factors and this, for simplicity, is outside the scope of reference. The general implications from, for example, Bennett et al. (1992) is that the private rates of return to women from higher education have tended to be larger than those for men, though this may have changed with increased participation. Overall, the issue of gender on rates of return may not be as important as may initially be envisaged. </p> <p>(n28.) It is possible for a log-normal distribution to remain with differential growth rates. In this case, the parameters of the distribution would change. </p> <hd id="AN0000779665-8">Table 1. Social rates of return</hd> <ct id="AN0000779665-9"> Average Average graduate traditional after expansion graduate (%) 2% Growth, 2% growth 6.8% 6.7 4% graduate unemployment (9.7%) (9.6) 2% Growth, 2% growth X 5.9 7% graduate unemployment (8.7) 2% Growth, 1% growth X 9.2 4% graduate unemployment (12.4) 2% Growth, 1% growth X 8.6 7% graduate unemployment (11.7) 2% Growth 0% growth X 10.6 4% graduate unemployment (14.1) 2% Growth 0% growth X 10.2 7% graduate unemployment (13.5) 2.5% Growt, 1.5% growth X 9.7 4% graduate unemployment (13.0) 2.5% Growth, 1.5% growth X 9.2 7% graduate unemployment (12.3) 3% Growth, 1% growth X 11.7 4% graduate unemployment (15.3) 3% Growth, 1% growth X 11.2 7% graduate unemployment (14.7) Best marginal Likely marginal graduate after graduate after expansion (%) expansion (%) 2% Growth, 2% growth 4.5 3.1 4% graduate unemployment (6.7) (5.2) 2% Growth, 2% growth 3.9 2.5 7% graduate unemployment (6.1) (4.5) 2% Growth, 1% growth 7.6 6.9 4% graduate unemployment (10.2) (9.4) 2% Growth, 1% growth 7.1 6.4 7% graduate unemployment (13.0) (8.7) 2% Growth 0% growth 9.2 8.6 4% graduate unemployment (12.0) (11.3) 2% Growth 0% growth 8.7 8.2 7% graduate unemployment (11.4) (10.8) 2.5% Growth, 1.5% growth 8.1 7.4 4% graduate unemployment (10.7) (9.9) 2.5% Growth, 1.5% growth 7.6 6.9 7% graduate unemployment (10.0) (9.2) 3% Growth, 1% growth 10.2 9.6 4% graduate unemployment (13.0) (12.4) 3% Growth, 1% growth 9.7 9.2 7% graduate unemployment (12.5) (11.8) Note: Non-graduate unemployment is 10% in all cases. Figures in parentheses are using an alpha factor of unity. The other figures use an alpha factor of 0.6. With regard to growth, the first figure given is the growth of graduate incomes and the second figure given is that of non-graduates.</ct> <hd id="AN0000779665-10">Table 2. Summary of earnings' profile before growth</hd> <ct id="AN0000779665-11"> Marginal traditional/ Median median traditional expansion Starting salary at graduation £12500 £11000 (£14375) (£12650) Salary at age 40 £27500 £23000 (£31625) (£26500) Alternative starting wage at 18 £10650 £9250 (£12500) (£10650) Alternative wage at 40 £21950 £18700 (£25225) (£21500) `Best' marginal Marginal expansion expansion Starting salary at graduation £8000 £8000 (£9200) (£9200) Salary at age 40 £19000 £18000 (£21850) (£20700) Alternative starting wage at 18 £8000 £8000 (£8840) (£8840) Alternative wage at 40 £16125 £16125 (£18550) (£18550) Notes: Adjustments must then be made for expected employment rates and for any assumed growth in order to gain rates of return.</ct> <hd1 id="AN0000779665-12"> References </hd1> <p>Aamodt, P. O. & Arnesen, C. A. (1995) The relationship between expansion in higher education and the labour market in Norway, European Journal of Education, 30, pp. 65-76. </p> <p>Arrow, K. (1973) Higher education as a filter, Journal of Public Economics, 2, pp. 193-216. </p> <p>Ashworth, J. S. (1996) A waste of time? (Private rates of return to higher education in the 1990s), University (if Durham Department of Economics, Working Paper No. 170 (Durham, University of Durham). </p> <p>AUT (1995) Higher Education: Preparing for the 21st Century (London, AUT). </p> <p>Bee, M. & Dolton, P. J. (1995) The remuneration of school teachers: time-series and cross-section evidence, Manchester School, 63, pp. 1-22. </p> <p>Bennett, R. J., Glennerster, H. & Nevison, D. (1992) Investing in skill: to stay on or not to stay on, Oxford Review of Economic Policy, 8, pp. 130-145. </p> <p>CBI (1994) Thinking Ahead: Ensuring the Expansion of Higher Education into the 21st Century (London, CBI). </p> <p>Chapman, B. J. & Harding, A. (1993) Australian student loans, Australian National University Centre for Economic Policy Research, Discussion Paper 287 (Australia, Australia National University). </p> <p>Chatterji, M. & McKaig, A. J. (1995) Convergence, growth and education, University of Dundee, Discussion Paper in Economics No. 65 (Dundee, University of Dundee). </p> <p>Chia, T. T. (1990) Returns to Higher Education in Australia, PhD thesis, Australian National University, Canberra. </p> <p>Cohn, E. (1992) The impact of surplus schooling on earnings: comment, Journal of Human Resources, 27, pp. 679-682. </p> <p>Creedy, J. (1994) Financing higher education: public choice and social welfare, Fiscal Studies, 15, pp. 87-108 </p> <p>Daly, A. (1982) The contribution of education to economic growth in Britain, National Institute of Economic and Social Research, 101, pp. 48- 56. </p> <p>de Meulemeester, J.-L. & Rochat, D. (1995) A causality analysis of the link between higher education and economic development, Economics of Education Review, 14, pp. 351-361. </p> <p>Dennision, F. F. (1974) Accounting for United States Economic Growth 1929-1969 (Washington, DC, Brookings Institution). </p> <p>Dennison, E. F. (1979) Accounting for Slower Economic Growth: The United States in the 1970s (Washington, DC, Brookings Institution). </p> <p>Department of Employment (1990) Labour costs in 1988, Employment Gazette, 98, pp. 431-437. </p> <p>DES (1985) The Development of Higher Education into the 1990s (Cmnd 9524) (London, HMSO). </p> <p>DES (1988) Top-Up Loans for Students (Cmnd Paper 520) (London, HMSO). </p> <p>Dolton, P. J. (1990) The economics of UK teacher supply, Economic Journal, 100, pp. S91-S104. </p> <p>Dolton, P. J. & Mavromaras, K. G. (1994) Inter-generational occupational choice comparisons: the case of teachers in the UK, Economic Journal, 104, pp. 841-863. </p> <p>Dolton, P. J. & van der Klaauw, W. (1995) Leaving teaching in the UK: a duration analysis, Economic Journal, 105, pp. 431-444. </p> <p>Duncan, G. & Hoffman, D. (1981) The incidence and the wage effects of over-education, Economics of Education Review, 1, pp. 75-86. </p> <p>Eckhaus, R, S. (1973) Estimating the Returns to Education: A Disaggregated Approach (Berkeley, CA, Carnegie Commission on Higher Education). </p> <p>Economist (1995) Britain's University Challenge and Degrees of Freedom, April 22. </p> <p>Freeman, R. (1976) The Overeducated American (New York, Academic Press). </p> <p>Freeman, R. (1980) The facts about the declining value of college, Journal of Human Resources, 15, pp. 124-142. </p> <p>Freeman, R. (1995) The limits of wage flexibility to curing unemployment, Oxford Review of Economic Policy, 11, pp. 63-72. </p> <p>Gill, A. M. & Solberg, E. J. (1992) Surplus schooling and earnings: a critique, Journal of Human Resources, 27, pp. 683-689. </p> <p>Glass, J. C., McKillop, D. G. & Hyndman, N. (1995a) Efficiency of provision of university teaching and research: an empirical analysis of UK universities, Journal of Applied Econometrics, 10, pp. 61-72. </p> <p>Glass, J. C., McKillop, D. G. & Hyndman, N. (1995b) The achievement of scale efficiency in UK universities: a multiple-input multiple-output analysis, Education Economics, 3, pp. 249-264. </p> <p>Gosling, A., Machin, S. & Meghir, C. (1996) What has happened to the wages of men since 1966?, in: Hills, J. (Ed.) New Inequalities: The Changing Distribution of Income and Wealth in the UK (Cambridge, Cambridge University Press). </p> <p>Gregg, P. & Machin, S. (1994) Is the UK rise in inequality different, in: Barrell, R. (Ed.) The UK Labour Market (Cambridge, Cambridge University Press). </p> <p>Groot, W. (1994) Differences in rates of return by type of education, Education Economics, 2, pp. 209-214. </p> <p>Grubb, W. N. (1992a) Post-secondary education and the sub-baccalaureate labor market: new evidence on economic returns, Economics of Education Review, 11, pp. 225-248. </p> <p>Grubb, W. N. (1992b) The economic returns to baccalaureate degrees: new evidence from the class of 1972, Review of Higher Education, 15, pp. 213-230. </p> <p>Grubb, W. N. (1993) The varied economic returns to post-secondary education, Journal of Human Resources, 28, pp. 365-382. </p> <p>Grubb, W. N. (1995a) Post-secondary education and the sub-baccalaureate labor market: corrections and extensions, Economics of Education Review, 14, pp. 285-299. </p> <p>Grubb, W. N. (1995b) Response to comment, Journal of Human Resources, 30, pp. 222-228. </p> <p>Grubel, H. (1987) The economics of the brain drain, in: Psacharopoulos, G. (Ed.) Economics of Education (Oxford, Pergamon), </p> <p>Hartog, J. & Oosterbeek, H. (1988) Education, allocation and earnings in the Netherlands: overschooling?, Economics of Education Review, 7, pp. 185-194. </p> <p>IES (Institute of Employment Studies) (1995) Graduate Review 1995-1996, IES Report 296 (Poole, BEBC). </p> <p>Jaeger, D. A. & Page, M. E. (1996) Degrees matter: new evidence on sheepskin effects and returns to education, Review of Economics and Statistics, 78, pp. 733-740. </p> <p>Johnes, G. (1993a) The Economics of Education (London, Macmillan). </p> <p>Johnes, G. (1993b) A degree of waste: a dissenting view, Oxford Review of Education, 19, pp. 459-464. </p> <p>Johnes, G. (1996) Multi-product cost functions and the funding of tuition in UK universities, Applied Economics Letters, 3, pp. 557-561. </p> <p>Johnes, G. & Johnes, J. (1994) Policy reforms and the theory of education finance, Journal of Economic Studies, 21, pp, 3-15. </p> <p>Javonic, B. (11979) Job matching and the theory of turnover, Journal of Political Economy, 87, pp. 972-990. </p> <p>Kane, T. J. & Rouse, C. E. (1995a) Labor-market returns to two- and four-year college, American Economic Review, 85, pp. 600-614. </p> <p>Kane, T. J. & Rouse, C. E. (1995b) Comment on the varied economic returns to post-secondary education, Journal of Human Resources, 30, pp. 205-221. </p> <p>Katz, L. & Murphy, K. (1992) Changes in the structure of wages, 1963-1987: supply and demand factors, Quarterly Journal of Economics, 107, pp. 35-78. </p> <p>Keep, E. & Mayhew, K. (1996) Economic demand for higher education--a sound foundation for further expansion?, Higher Education Quarterly, 50, pp. 89-109. </p> <p>Layard, P. R. G. & Psacharopoulos, G. (1974) The screening hypothesis and the returns to education, Journal of Political Economy, 82, pp. 985-998. </p> <p>Levin, H. M. & Kelly, C. (1994) Can education do it alone?, Economics of Education Review, 13, pp. 97-108. </p> <p>Lissenburgh, S. & Bryson, A. (1996) The Returns to Graduation (London, HMSO). </p> <p>Lynch, L. (1993) Youth training in the United States, Economic Journal, 103, pp, 1292-1302. </p> <p>Maglen, L. R. (1990) Challenging the human capital orthodoxy, Economic Record, 66, pp. 281-291. </p> <p>Mallier, T. & Rodgers, T. (1995) Measuring value added in higher education: a proposal, Education Economics, 3, pp. 119-132. </p> <p>Mason, G. (1995) The New Graduate Supply Shock: Recruitment and Utilisation of Graduates in British Industry, National Institute Research Report No. 9 (London, NIESR). </p> <p>Mason, G. (1996) Graduate utilisation in British industry: the impact of mass higher education, National Institute Economic Review, 156, pp. 93-103. </p> <p>McMahon, W. W. (1987) The relation of education and R&D to productivity growth in the developing countries of Africa, Economics of Education Review, 6, pp. 183-194. </p> <p>Moll, T. (1992) Rising earnings inequality and returns to skills in the United Kingdom ]about market, Labour, 6, pp. 45-69. </p> <p>Murnane, R. J. & Olsen, R. J. (1989) Will there be enough teachers?, American Economic Review (Papers and Proceedings), 79, pp. 242-246. </p> <p>Murphy, J. (1993) A degree of waste: the economic benefits of educational expansion, Oxford Review of Education, 19, pp. 9-31. </p> <p>Murphy, J. (1994) A degree of waste: reply to Johnes, Oxford Review of Education, 20, pp. 81-92. </p> <p>Murphy, K. & Welch, F. (1992) The structure of wages, Quarterly Journal of Economics, 107, pp. 258-326. </p> <p>Psacharopoulos, G. (1981) Return to education: an updated international comparison, Comparative Education, 17, pp. 321-341. </p> <p>Psacharopoulos, G. (1985) Return to education: a further international update and implications, journal of Human Resources, 20, pp. 583-604. </p> <p>Psacharopoulos, G. (1994) Returns to investment in education, World Development, 22, pp. 1325-1343. </p> <p>Robst, J. (1995) College quality and overeducation, Economics of Education Review, 14, pp. 221-228. </p> <p>Rosen, S. (1972) Learning and experience in the labor market, Journal of Human Resources, 7, pp. 326-1342. </p> <p>Rudd, E. (1990) The early careers of social science graduates and the value of a PhD, Journal of the Royal Statistical Society, Series A, 153, pp. 203-232. </p> <p>Rumberger, R. (1987) The impact of surplus schooling on productivity and earnings, Journal of Human Resources, 22, pp. 24-50. </p> <p>Sicherman, N. & Galor, O. (1990) A theory of career mobility, Journal of Political Economy, 98, pp. 169-192. </p> <p>Skidelsky, R. (1995) Public finance and education, Oxford Review of Economic Policy, 11, pp. 59-66. </p> <p>Sloane, P. J., Battu, H. & Seaman, P. T. (1995a) Over-education and the formal education/experience and training trade-off, University of Aberdeen Department of Economics Discussion paper 95-08. (Aberdeen, University of Aberdeen). </p> <p>Sloane, P. J., Battu, H. & Seaman, P. T. (1995b) Over-education, under-education and the British labour market, University of Aberdeen Department of Economics, Discussion Raper 95-09 (Aberdeen, University of Aberdeen), </p> <p>Smithers, A. & Robinson, P. (1995) Post-18 Education--Growth, Chance, Prospect (London, Council for Industry and Higher Education). </p> <p>Social Trends (1992, 1993, 1994, 1995) (London, HMSO). </p> <p>Soskice, D. (199 3) Social skills from mass higher education: re-thinking the company based initial training paradigm, Oxford Review of Economic Policy, 9, pp. 101-113. </p> <p>Spence, M. (1973) Job market signalling, Quarterly Journal of Economics, 87, pp. 353-374. </p> <p>Tarsh, J. (1992) The labour market for newly qualified higher degree graduates, From Higher Education to Employment (Paris, OECD). </p> <p>Teichler, U. & Kehm, B. M. (1995) Towards a new understanding of the relationship between higher education and employment, European Journal of Education, 30, pp. 15-132. </p> <p>THES (1995a) Times Higher Education Supplement, 6 May. </p> <p>THES (1995b) Times Higher Education Supplement, 21 July. </p> <p>THES (1995c) Times Higher Education Supplement, 29 September. </p> <p>Thomas, G. (1994) Students' Money Matters (Richmond, Trotman). </p> <p>Thurow, L. (1975) Generating Inequality: Mechanisms of Distribution in the US Economy (London, Macmillan). </p> <p>Tsang, M. C. & Levin, H. M. (1985) The economics of over education, Economics of Education Review, 4, pp. 93-104. </p> <p>University Statistical Records (1990) University Statistics, 1989-1990 (London, USR). </p> <p>University Statistical Records (1994) University Statistics, 1993-1994 (London, USR). </p> <p>Vaillancourt, F. (1995) The private and total returns to education in Canada, 1989, Canadian Journal of Economics, 28, pp. 532-554. </p> <p>Verdugo, R. R. & Verdugo, N. T. (1989) The impact of surplus schooling on earnings: some additional findings, Journal of Human Resources, 24, pp. 629-643. </p> <p>von Weizsacher, R. K. (1993) A Theory of Income Distribution (Cambridge, CUP). </p> <p>Weale, M. (1993) A critical evaluation of rate of return analysis, Economic Journal, 103, pp. 729-737. </p> <p>Wilson, R. A. (1986) The social returns of producing teachers, Higher Education Review, 18, pp. 5969. </p> <p>Wilson, R. A. (1995) Review of the Economy and Employment: Future Employment Prospects for the Highly Qualified (University of Warwick, Institute of Employment Research). </p> <p>Zabalza, A. (1979) The determinants of teacher supply, Review of Economic Studies, 46, pp. 131-147. </p> <p>Zabalza, A., Turnbull, P. P. Williams, G. (1979) The Economics of Teacher Supply (Cambridge, CUP). </p> <hd id="AN0000779665-13">Appendix</hd> <p>The social rate of return is defined as r which solves </p> <p>[Multiple line equation(s) cannot be represented in ASCII text] </p> <p>where Y<subs>x</subs> denotes the earnings of someone with the xth level of education and Z is the cost of education. The benefits may be adjusted by the alpha factor to account for basic ability as opposed to gains from education. </p> <p>All figures given here, unless indicated, as 1994/5 prices and, therefore, all increases are in real terms As indicated in the main text, assumptions are necessary to generate the rates of return and these are dealt with next. The main premise is to identify a basic set of reasonable assumptions that generate a social rate of return in line with those calculated in previous work for the traditional pre-1990 graduate and making minimum further assumptions to consider the position of representative graduates after the expansion in students numbers. The salaries are summarized in Table 2 with the rationale for the values given in the following sections. </p> <p>Basic Assumptions for the Social Rates of Return </p> <p>The use of gross salaries. The calculation of social rates of return requires the use of gross salaries (including employers' national insurance contributions and any non-contributory pensions) (see Wilson, 1986; DES, 1988). In this paper, the increase above gross salary is assumed to increase the labour costs by 15%. If only employer's national insurance was used, this would increase gross wages only by 10.5%. Examination of the proportion of wages and salaries in labour costs (see Department of Employment, 1990) might suggest that the figure used may be low for the type of white-collar industries in which graduates are typically employed. The proportion of wages and salaries in banking, finance and insurance labour costs has been as low as 70%) in 1981 and has never been above 77%. However, under expansion, a greater range of jobs will be taken up by graduates and, hence, a lower figure (for all calculations) was deemed appropriate for comparative purposes. </p> <p>The costs of the education and the role of grants and loans. Any grants or loans to students are omitted from the calculations as these are transfer payments as far as society is concerned. The social costs are those incurred providing the education for the students, no matter who pays for it. The relevant costs of the university are the costs of teaching (but not research) plus the costs of using the university premises for university purposes, i.e. the imputed rent. The value for this can be approximated from the money made available to universities for teaching divided by the number of full-time equivalent students,(<reflink idref="bib26" id="ref28">n26</reflink>) with a rounding up to accommodate the land usage. For the calculations in the text, the costs were £8000 per annum before expansion and £6600 per annum after the expansion (see THES, 1995c). Lowering these figures, in line with any future government policy, will increase the returns. </p> <p>Unemployment rates. The expected earnings for the representative student can be adjusted to accommodate the probability of being unemployed. In this paper, a naive multiplier of the wage multiplied by the probability of being employed, using the presumed unemployment rate, is used. As can be seen the results are not over-sensitive to the choice, given that large changes are not made. </p> <p>Assumptions for the Social Rate of Return for Average Pre-expansion Graduates </p> <p>The earnings of graduates. The average graduate under the pre-expansion conditions is assumed to have left university to enter a job paying £12 500 per annum which gives a total gross cost (henceforth referred to as gross income) of around £14 375. In this part of the exercise, where no economic growth is assumed, this wage is assumed to rise in equal annual steps to £27 500 (gross ground £31 625) at the age of 40 and is held constant thereafter. The equal steps are reasonable because of the use of a representative student and reflect the changes in income due to experience and training. The figures are in line with those generally presented by commentators on this subject. In 1992, the average male graduate salary was approximately £25 000 per annum which, when inflated to 1995 figures, rises to around £27 500 (see The Economist, 1995; Social Trends, 1992, 1993, 1994, 1995). As the increase in student numbers was only just beginning to filter through in 1993, the figures can be taken as representative of the old regime. The assumption that all earnings rises occur by the age of 40 is generous, although this can be tempered by not having any increases after that age.(<reflink idref="bib27" id="ref29">n27</reflink>) </p> <p>The earnings of non-graduates. In order to calculate the rate of return, an alternative earnings stream is required. The Economist (1995) suggested that over the 3-year period this foregone earning would total £25 000 after tax. This does not seem to be an unreasonable proposition for the cohort being treated here. The sum can be equated to an employers' gross starting salary of £12 250. While this may appear high, it should be remembered that these are perceived to be the most able 15% of the age group population. Without growth in the economy, the salary is assumed to rise over the 22-year period at 3.4% per annum in real terms to reflect experience and promotion, to a salary of around £21 950 per annum. (gross around £25 225). The growth rate was chosen such that it gave reasonable alternative earnings for both the traditional average and marginal graduates. The earnings are comparable with those of well qualified and highest quartile A-level educated people in the GHS sample. </p> <p>The role of the growth. The relevant growth rates are applied to the earnings streams which have been created above. For example, the earnings in the 10th year of an income stream is multiplied by (1 + g)<sups>10</sups> where g is the growth rate. </p> <p>Making all the assumptions above gives the figures present in Table 1, column 1. </p> <p>The difference between the text figure and those of other studies. Explanations for the slightly higher base figure of the text when compared with the DES (1988) figures and the general `myths' are that they reflect some of the increased returns gained by skilled workers in the 1980s or that the true labour costs have not previously been taken into account. If one follows the notes of the DES report, then there is only explicit notice made of national insurance payments and no mention of the pension contributions. Using a mark up on gross salary of only 10.5%, the social rate of return falls by approximately 1.0%. Even this figure is larger than those previously reported by the DES (1988). However, the earnings of students (see Thomas, 1994) during their time in higher education should also be included in the benefits as this is productive labour in the given time period and it is not clear whether this was the case in the DES figures. This will also serve to increase the rates of return. </p> <p>The approach adopted in this paper creates initial base social rates of return in line with other researchers (see Johnes, 1993b; Psacharopoulos, 1981, 1985, 1994). Naturally, discrepancies will arise due to using a representative student and it must be emphasized that the figures should be taken as relative for discussion. </p> <p>Assumptions for the Social Rate of Return for Average Post-expansion graduates </p> <p>The distribution of income. Initially, a log-normal distribution of income with average earnings of £15 500 with a standard deviation of the log-normal of 0.4 is assumed as this gives 7.5% of the population earning an income of more than £27 500, the salary used for the same representative student in the pre-expansion era. The log-normal is widely recognized to be a good approximation of the distribution of income. Any problems with using this distribution tend to be in the tails of the distribution (see von Weizsacker, 1993). For this exercise, the extreme tails are not being used arid, so, for illustrative purposes the log-normal is a reasonable assumption and is used later. </p> <p>The earnings of graduates. The calculations are made presuming that the long-term position gives air equilibrium position in the income distribution for the representative student at the age of 40. Assuming, for the purposes of illustration, that all the high income earners are graduates, £27 500 is the salary of the median traditional graduate. Using (again) the median income for graduates, which is probably the best approximation for the average graduate, and (again) assuming that the top earners are all graduates, a change to a 30% cohort requires that we find the wage in the distribution where there is 15% of the distribution above that point. This lowers the average graduate wage to £23 000, approximately £26 500 gross with the employers' costs, in a no growth position. </p> <p>The earnings of non-graduates. The alternative earnings will also fall in equilibrium, due to the displacement effect. There is a fall in initial earnings, assumed to be to around £10 650 including all costs in the first year, rising again at 3.4% per annum to an income at 40 years of age of approximately £21 500 gross with national insurance and pensions (a salary of around £18 700), again in the no growth position. Clearly, some further assumptions are required, notably in the initial starting salaries. Lowering the starting salary further would raise the returns but this seems unreasonable given the position of the marginal students described. It should be noted that the change in the average income of graduates is greater than that of non-graduates which is entirely in line with the shape of the income distribution. If alternative earnings had stayed at the level before. expansion, the rate of return falls further. This may be the case in the transitional period as less able people do the traditional A-level jobs but is not the case in the longer term as displacement occurs. </p> <p>The role of the growth. Any growth in the economy (or to the real wages in a sector of the economy) can again be assigned to income streams as before. Only if growth is the same for all sectors does the distribution remain unchanged.(<reflink idref="bib28" id="ref30">n28</reflink>) </p> <p>Assumptions for the Social Rate of Return for Marginal Post-expansion Graduates </p> <p>Given the out-turn of the analysis, these students are the crucial ones for consideration. Society is indifferent to who is in given jobs and, for the purposes of these calculations, the marginal student, on entry, can be treated as the marginal graduate. If this were not the case, the returns would fall for this group as alternative earnings would rise (although there would be increases in the returns to graduates elsewhere in the cohort). The use of a representative student precludes a consideration of how a change in the rankings will affect the overall analysis. </p> <p>The earnings of graduates. For these graduates, information on starting salaries suggests that they are as low as the starting salary of `less able' present A-level holders. Figures from the National Institute (see Mason, 1995, 1996) suggest this figure is as low as £8 000 (grossed to £9 200; the allocation of the full pension arrangements immediately is generous-without this and using a gross starting salary of £8 840, the rates of return fall). However, their salary can be assumed to increase to their point in the distribution, thus giving a gain in salary in equilibrium at age 40. For the `best marginal' person this means that 30% of the distribution earn a greater wage than them, which is £19 000 (gross £21 850) and, for the marginal, 35% earn more than £18 000 (gross £ 20 700). </p> <p>The earnings of non-graduates. The alternative earnings are assumed to grow at 3.4% per annum to the age of 40 from the same initial point of a salary of £8 000 (but with the additional costs of pensions only applied after age 21). The assumption is that the gains from education reveal themselves to employers over time after graduation but there is no initial gain from the possession of a degree per se. Clearly, if the higher alternative salary (of the adjustment period) had been used, the rates of return would have fallen but this is an unreasonable proposition in the longer term. In the interim period, for the likely marginal group, the higher alternative could drive returns to below 2%. The alternative earnings increase to a gross salary of just above £18 500 and so it can be seen that the returns arc very slow to accrue if there is equal growth in the sectors. </p> <p>The role of the growth for marginal students. The use of differential growth rates is less clear in the case of the marginal students than average students for the reasons given in the text relating to occupations of marginal students. The higher the alpha factor, the greater the grounds for presupposing unequal rates of growth. As can be seen from the calculations, assumptions about growth are crucial to the returns and, hence, any policy decisions. As the text argues, the uncertainties raised by the necessary assumption, coupled with the projected returns, militate against further expansion in the short term. </p> <aug> <p>By JOHN ASHWORTH </p> <p></p> <p>J. Ashworth, Department of Economics, University of Durham, 23-26 Old Elvet, Durham DH1 3HY, UK. Tel: 0191 374 7281. </p> </aug> <nolink nlid="nl1" bibid="bib1" firstref="ref1"></nolink> <nolink nlid="nl2" bibid="bib2" firstref="ref2"></nolink> <nolink nlid="nl3" bibid="bib3" firstref="ref3"></nolink> <nolink nlid="nl4" bibid="bib4" firstref="ref4"></nolink> <nolink nlid="nl5" bibid="bib5" firstref="ref5"></nolink> <nolink nlid="nl6" bibid="bib6" firstref="ref6"></nolink> <nolink nlid="nl7" bibid="bib7" firstref="ref7"></nolink> <nolink nlid="nl8" bibid="bib8" firstref="ref8"></nolink> <nolink nlid="nl9" bibid="bib9" firstref="ref9"></nolink> <nolink nlid="nl10" bibid="bib10" firstref="ref10"></nolink> <nolink nlid="nl11" bibid="bib11" firstref="ref11"></nolink> <nolink nlid="nl12" bibid="bib12" firstref="ref12"></nolink> <nolink nlid="nl13" bibid="bib13" firstref="ref13"></nolink> <nolink nlid="nl14" bibid="bib14" firstref="ref14"></nolink> <nolink nlid="nl15" bibid="bib15" firstref="ref15"></nolink> <nolink nlid="nl16" bibid="bib16" firstref="ref16"></nolink> <nolink nlid="nl17" bibid="bib17" firstref="ref17"></nolink> <nolink nlid="nl18" bibid="bib18" firstref="ref18"></nolink> <nolink nlid="nl19" bibid="bib19" firstref="ref19"></nolink> <nolink nlid="nl20" bibid="bib20" firstref="ref20"></nolink> <nolink nlid="nl21" bibid="bib21" firstref="ref21"></nolink> <nolink nlid="nl22" bibid="bib22" firstref="ref22"></nolink> <nolink nlid="nl23" bibid="bib23" firstref="ref23"></nolink> <nolink nlid="nl24" bibid="bib24" firstref="ref24"></nolink> <nolink nlid="nl25" bibid="bib25" firstref="ref25"></nolink> <nolink nlid="nl26" bibid="bib26" firstref="ref28"></nolink> <nolink nlid="nl27" bibid="bib27" firstref="ref29"></nolink> <nolink nlid="nl28" bibid="bib28" firstref="ref30"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ567344
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: A Waste of Resources? Social Rates of Return to Higher Education in the 1990s.
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Ashworth%2C+John%22">Ashworth, John</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22Education+Economics%22"><i>Education Economics</i></searchLink>. Apr 1998 6(1):27-44.
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 18
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 1998
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Reports - Evaluative
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Economic+Factors%22">Economic Factors</searchLink><br /><searchLink fieldCode="DE" term="%22Educational+Development%22">Educational Development</searchLink><br /><searchLink fieldCode="DE" term="%22Foreign+Countries%22">Foreign Countries</searchLink><br /><searchLink fieldCode="DE" term="%22Higher+Education%22">Higher Education</searchLink><br /><searchLink fieldCode="DE" term="%22Human+Capital%22">Human Capital</searchLink><br /><searchLink fieldCode="DE" term="%22Program+Effectiveness%22">Program Effectiveness</searchLink><br /><searchLink fieldCode="DE" term="%22Resource+Allocation%22">Resource Allocation</searchLink>
– Name: Subject
  Label: Geographic Terms
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22United+Kingdom%22">United Kingdom</searchLink>
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 0964-5292
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Demonstrates that popular assertions regarding the social benefits of additional higher education are dubious. Reveals four decisive factors: presumed economic growth; changes in graduates' and nongraduates' relative earnings; differences between the marginal and average student; and belief in scale economies. Unless economic growth favors graduates, society's reward is too low to justify further expansion of higher education. (100 references) (MLH)
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 1999
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ567344
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ567344
RecordInfo BibRecord:
  BibEntity:
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 18
        StartPage: 27
    Subjects:
      – SubjectFull: Economic Factors
        Type: general
      – SubjectFull: Educational Development
        Type: general
      – SubjectFull: Foreign Countries
        Type: general
      – SubjectFull: Higher Education
        Type: general
      – SubjectFull: Human Capital
        Type: general
      – SubjectFull: Program Effectiveness
        Type: general
      – SubjectFull: Resource Allocation
        Type: general
      – SubjectFull: United Kingdom
        Type: general
    Titles:
      – TitleFull: A Waste of Resources? Social Rates of Return to Higher Education in the 1990s.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Ashworth, John
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 01
              Type: published
              Y: 1998
          Identifiers:
            – Type: issn-print
              Value: 0964-5292
          Numbering:
            – Type: volume
              Value: 6
            – Type: issue
              Value: 1
          Titles:
            – TitleFull: Education Economics
              Type: main
ResultId 1