The Changing Graduate Labour Market: A Review of the · PDF fileThe Changing Graduate Labour Market: A Review ... graduates are likely to work outside of the Principality at some stage

Technical Report

The Changing GraduateLabour Market: A Reviewof the Evidence

Professor Phillip BrownSchool of Social Sciences, CardiffUniversity

With

Claire SmetherhamSchool of Social Sciences, Cardiff University

ii

TABLE OF CONTENTS

List of Figures iv

List of Tables vi

Executive Summary viiii

Background viiii

Key Findings ixii

Implications for Policy and Future Research xiii

Main Report 1

1. Purpose 1

2. Background 1

3. Findings 4

3.1 What many graduates earn will not match 4their expectations

3.2 Most graduates will earn more than non-graduates 7but the ‘premium’ is often much lower than the50 per cent difference in life-time earnings reportedin the White Paper on The Future ofHigher Education

3.3 There is evidence of significant and growing income 11differences between graduates in the labour market

3.4 There are significant ‘sectoral’ and ‘occuptional’differences in graduate/non-graduate earnings ratios 17

3.5 There are significant regional differences in 21graduate incomes

3.6 Existing research evidence also shows that 23graduate earnings reflect a number of other factors,including the status of the university attended, thedegree subject studied, and degree classification

3.7 Major gender differences remain in areas of 30specialist knowledge, industrial sector, occupationaland income

3.8 Class and ethnic inequalities also continue to be a 35feature of educational and labour market outcomes

3.9 There is evidence of graduate ‘over-qualification’but little agreement about how to measure it

3.10 Claims that half of those entering the workforce 40will require higher education to meet therequirements of the ‘knowledge’ economy have

iii

little empirical support on either side of the Atlantic.There is also little evidence to support the view thatWales is a vibrant knowledge-driven economy

3.11 There is little agreement about the extent or nature 52of skill gaps in Wales

3.12 There is high degree of integration between the 56Welsh and English labour market, with considerablemovement of students and graduate employeesacross national borders

4. Implications for Policy and Future Research 59

Bibliography 64

iv

List of Figures

Figure 1: Job hunter’s expected starting salaries 5

Figure 2: Average returns to education: Private 7Real Rates of Return, University-LevelEducation, 1999-2007

Figure 3: Male Graduates aged 26 to 35 – net weekly 11earnings (full-time employees and

self-employed)

Figure 4: Women Graduates aged 26 to 35 – net 12weekly earnings (full-time employees andself-employed)

Figure 5: U.S. male hourly wages by decile within 13educated groups, 1973-2001

Figure 6: U.S. female hourly wages by decile within 14educated groups, 1973-2001

Figure 7: Estimated net weekly income for full-time 15employees only, by gender, qualificationand decile (1983-2001)

Figure 8: Average earnings of graduates by degree 27subject

Figure 9: The occupational distribution of 32graduates

Figure 10: Distribution of graduates by industry 34sector

Figure 11: Higher education entrants by social class 35groups (1960-2000)

Figure 12: Blue chip graduate vacancies in the UK 421997-2003

v

Figure 13: Changing occupational structure of male 43employees in employment, 1975 to 2000by SOC (HE)

Figure 14: Changing occupational structure of 43female employees in employment, 1975to 2000 by SOC (HE)

Figure 15: Job classifications – all UK jobs 471971-2010

Figure 16: Percentage changes within categories 48between periods

Figure 17: Job classifications – all jobs 1971-2010 49(Wales)

Figure 18: Annual employment increase: 50head count

vi

List of TablesTable 1: Net weekly earnings (full-time employees only) by 9

cohort by gender by qualification level.

Table 2: Graduates and non-graduates: net hourly pay 10(employees full-time and part-time) 30 Year Olds.

Table 3: Earnings ratio between 10th percentile and 90th 16Percentile, Wales 1993, 1997 and 2001*

Table 4: Earnings ratio between 10th percentile and 90th 17percentile, Wales 1993, 1997 and 2001*

Table 5: Net weekly earnings of graduates aged 26-35 18(full-time workers) by industrial sector [prices/year]

Table 6: Median weekly gross pay between degree holders 20and those without in the UK (21 years old or older)1993 and 2001

Table 7: Earnings (£) of graduates in Wales over time 21compared with other regions

Table 8: Average weekly earnings (£) of full-time graduate 22employees by region of place of work, and age

Table 9: 'Real' graduate earnings by region 23

Table 10: Top ten worst graduate starting salaries by 25institution

Table 11: Average annual income of respondents with 26Firsts by area of degree study (1997 cohort)

Table 12: Table of HE subjects and differences over time (UK) 28

Table 13: Average annual income of respondents by degree 29classification (1997 cohort)

Table 14: Average annual income of respondents with 30Firsts by gender (1997 cohort)

Table 15: Average annual earnings of graduates, 1999-2000 31

Table 16: Occupational distribution of graduates by disciplinary 32background

Table 17: Distribution of employment of graduates by major 33economic sector

vii

Table 18: First destinations of full-time degree (home 37domiciled) graduates from English universities,2001/02 (percentages are based on knowndestinations)

Table 19: Whether a degree was a formal entry requirement 39for respondent’s current job by degree class(1997 cohort)

Table 20: Percentage increases and percentage shares of 41economy compared

Table 21: Employment by education and training category, 452000-2010. (Monthly Labor Review)

Table 22: Distribution of full-time graduate employees in the 51private and public sectors in each region by placeof work

Table 23: Qualifications demand and supply, Wales, 2001 53

Table 24: Patterns of qualification mismatch, Wales and 55Britain, 2001

Table 25: Particular skills index By region/country, 2001 56

Table 26: The ‘retention’ rate of graduates entering 57employment by country of domicile and countryof HEI attended

Table 27: Employment destination of UK graduates gaining 57qualifications in UK IHEs by country of domicile,country of IHE and location of first destination

viii

Executive Summary

Background

The Westminster Government has primarily justified the expansion and futurefunding of higher education on economic grounds. It is assumed that there isa growing demand for high skilled workers and that graduates typically earnsignificantly more than non-graduates.

The purpose of this report is to ensure that a comprehensive review of theavailable evidence is conducted to enable the National Assembly of Wales todevelop an effective and fair system of student support.

This review examines evidence from across the UK, given that many Welshgraduates are likely to work outside of the Principality at some stage in theirworking lives. International comparisons are limited to the United States.

The evidence presented in this report is based on survey data to examinetrends in the demand for graduates in Wales and beyond. It also investigatestrends in graduate incomes. It goes beyond an analysis of rates of returnbetween graduates and non-graduates to include differences amongst thosewith graduate qualifications. Gender differences are also examined, asexisting evidence shows that although there are more women in highereducation, men typically earn higher incomes.

One of the major problems in undertaking this research is the lack ofadequate data on graduate employment in Wales. Most large-scale data setsfor the United Kingdom have relatively small samples at the regional level thatmake it difficult to assess their reliability when attempting to generalise aboutgraduate employment in the Principality.

The evidence on graduate employment and income is more complex than theheadline figures suggest. There are also conceptual confusions and disputesas to what constitutes a ‘graduate’ job or ‘over-qualification’. Therefore, theevidence presented in this report should be treated as indicative rather thanconclusive. However, its general findings raise serious doubts about theclaims made about graduate earnings and employment trends in the WhitePaper on the Future of Higher Education that informed new legislation on thefuture funding of higher education in England.

ix

Key Findings

� What many graduates will earn will not match their expectations. Onaverage, students expect to earn around £19,000 in their first job afteruniversity. While this seems to be in line with what many largegraduate employers offer, the Higher Education Statistical Agency(HESA) reported an average of £17,000 per annum in 2003 (includingfull-time and part-time), and the NatWest survey reported a seven percent rise in graduate starting salaries from £12,659 in 2003 to £13,600in 2004.

� Most graduates will earn more than non-graduates but the ‘premium’ isoften much lower than the 50 per cent assumed by the WestminsterGovernment. Based on a new analysis of longitudinal data, the lifetimedifference between the median graduate and non-graduate was£166,000 (net income). In 2000, the median graduate earned 36 percent more, rather than the 50 per cent premium reported in The Futureof Higher Education (2003:59). It also shows that higher earning non-graduates earned 17 per cent more than the median graduate and 80per cent more than low earning graduates. None of these figures takeinto account the financial costs of a university education.

� There is evidence of significant and growing income differencesamongst graduates. Drawing on the same data set as above,differences within the graduate category were found to be twice thatbetween graduates and non-graduates. This amounts to a £374,000lifetime difference between median and high earning graduates [..].Other evidence from the General Household Survey shows that topearning male graduates are racing away from the rest, although thiswas not the case for women. Data from the United States also showthat high earning graduates have pulled further away from the rest, butunlike the British data, it also shows that high earning female graduateshave made the most significant gains.

� There is little evidence of a significant decline in graduate earnings, butthere is also no evidence that the ‘knowledge’ economy has increasedthe value of human capital for graduate employees. On both sides ofthe Atlantic, it is only high earning graduates that have enjoyed asizable increase in real income over the last years.

� Labour Force Survey data points to significant ‘sectoral’ and‘occuptional’ differences in graduate/non-graduate earnings ratios. InWales, this data suggests that these differentials may be narrowing,especially for graduates on lower incomes. However, the small size ofthe sample means that this finding needs to be treated with caution,although there is clearly a need for further research.

x

� Major gender differences remain in areas of specialist knowledge,industrial sector, occupation and income. Research shows that femalegraduates in their early twenties receive on average 15 per cent less,and the gender gap appears to widen with age. Female graduates arealmost twice as likely to work in the public sector and dominate ineducation and other public services, while men dominate inmanufacturing and construction. However, women have madeconsiderable progress in gaining access to the professions.

� Class and ethnic inequalities continue to be a feature of educationaland labour market outcomes. Recent studies show that ethnic minoritystudents are now more likely to participate in some form of highereducation than the white population. However, they are more likely tobe found in low-status institutions and confront disadvantage within thejob market. On first destination, the unemployment rate for whitegraduates was six per cent but just over 11 percent for ethnic minoritygroups. However, there are also significant differences within the ‘non-white’ category.

� Research also shows that those from higher socio-economic groupsare more likely to be in managerial and professional jobs traditionallyassociated with a university degree. It also shows that those whostudied near the family home or returned ‘home’ after graduation tosave money had particular difficulty finding suitable employment. Itshould also be remembered that the impact of class and ethnicinequalities are often hidden because they are ‘institutionalised’. While,for example, leading companies do not directly discriminate againstthose from working class backgrounds or those of Caribbean descent,they are less likely to attend the elite universities from which thesecompanies frequently recruit.

� There are significant regional differences in graduate incomes. Themajor difference is between London and the South East, compared toother regions. The median income for graduates between the ages of20-34 in Wales was £19,292, whereas the figure for London was£25,012 in 2002. The median for all regions was £21,474. Thesedifferences reflect the pattern of employment in various parts of theUnited Kingdom. Half of all full-time graduates in Wales work in thepublic sector, whereas it is less than a quarter in London and a littleunder 40 per cent in the North West of England and Scotland.

A more detailed analysis of regional difference in earnings show thatwhile ‘managers and senior officials’ in Wales are paid less (onaverage) than anywhere else in the UK, those in ‘personal serviceoccupations’ or in ‘administrative and secretarial occupations’ fareconsiderably better when compared to other regions. Moreover, whencost of living is taken into account, graduate earnings in Wales movefrom ninth to third behind the North and Yorkshire and Humberside,‘Greater London’ moves from second to bottom.

xi

� There is also evidence that differences in graduate earnings reflect thestatus of the university attended, degree subject and degreeclassification. The existing evidence shows that those in ‘older’universities have an earnings premium of at least 10 percent overthose from other institutions of higher education. It also shows thatwhat undergraduates study at university can have an impact onoccupational earnings. Both males and females with degrees in thefields of Maths and Computing and Medicine and Related Subjects hadthe highest incomes, while those with Humanities, Arts and Educationdegrees the lowest. However, there were important gender differences,as men are shown to earn more than women in all subject areas.

� Recent research has also compared the earnings of graduates withfirst class degrees against those with degrees classified as lowersecond class. On average, those with first class degrees were not onlyfound to earn more, but also more likely to be in a job that utilised theirknowledge and skills than those with lower second class degrees.However, this research also confirms the importance of gender - as theaverage income for women with first class degrees was similar to menwith lower second class degrees, while men with a first enjoyed asignificant premium over the rest.

� It was difficult to find evidence on the graduate premium for part time ormature learners. This is an area that requires further investigation.

� Claims that half of those entering the workforce will require highereducation to meet the requirements of the knowledge economy findslittle empirical support on either side of the Atlantic. Official data fromthe United States shows that occupations requiring a Bachelor’sdegree or higher is predicted to increase by less that 2 percentagepoints, from 20.7 to 21.8 per cent between 2000-2010. A re-analysis ofthe occupational data in the UK also shows that the proportion of‘knowledge’ workers will increase from 28 per cent in 1999 to 29 percent in 2010. The figures for Wales were expected to remain constantat 24 per cent. Other evidence based on data from the Association ofGraduate Recruiters, which represents many of Britain’s leadingemployers, shows considerable fluctuation in the demand forgraduates. Vacancy levels in 2002/03 had not recovered from a high of25,000 in 1997/98.

� Regional differences in the demand for skills are subject to contrastinginterpretations. Evidence drawn from the British Skills Survey suggeststhat in 1992, jobs in Wales required better qualifications, needed moretraining and took longer to learn than jobs elsewhere in Britain. Butover the last decade, jobs in Britain have generally become moreskilled, yet in Wales the trend has been in the opposite direction.

xii

Again, these conclusions are based on a very small sample thathighlights the urgent need for more research. Other research suggeststhat there is no evidence that Wales is falling behind other regions ofthe United Kingdom.

� A number of studies have found that approximately 30 per cent ofgraduates are ‘over-qualified’ in terms of what is required to fulfil theircurrent employment. However, there are different measures of ‘over-qualification’ that have led researchers to report findings varying from10 to almost 50 per cent. Studies that report over-qualification at thehigher end are usually based on the first jobs that are entered aftergraduation. Virtually all studies show that this mismatch declines asgraduates gain more work experience, but there is little evidence tosupport the claim that employers will ‘grow’ jobs to utilise this untappedsupply of higher level skills. Further research is required to examinewhether over-qualification has increased with the expansion of HigherEducation.

Implications for Policy and Future Research

� This report has focused on the relationship between higher educationand the labour market. While the contribution of higher education toindividual prosperity and economic competitiveness is of obviousimportance, this analysis needs to be considered alongside the widerpersonal and social benefits of learning within higher education.

� Although it is legitimate to talk about higher education in Wales or theWelsh economy, they are inextricably linked to institutionalarrangements in England. Many Welsh students study and work inEngland, although many return to the Principality. Ultimately, changingthe system of higher education in Wales depends on changing thesystem in England. Therefore, the room for manoeuvre in terms ofpolicy reform in Wales is constrained by these broader considerations.

� Much of the debate about ‘graduate’ employment is based on outdatedassumptions. The idea of a ‘graduate job’ had some validity in the1960s as many of the ‘elite’ that entered university found themselveson ‘fast track’ management training programmes or in the establishedprofessions such as law and medicine. It also provided a useful ‘shorthand’ for jobs that offered high salaries, good promotion prospects,high social status and job security. While this picture of graduateemployment may remain dominant in the brochures of leadinggraduate employers, it is far removed from the occupational realitiesconfronting many graduates.

� The evidence presented in this report highlights the problem of usingdata from organisations such as the Association of GraduateRecruiters (AGR), as representative of all graduates. The majority ofgraduates do not work for leading ‘blue chip’ companies but for small

xiii

and medium sized organisation which typically have lower salaries andfewer prospects for internal promotion. But it also highlights the dangerof using ‘average’ incomes which obscure the realities of today’s labourmarket, and that reveal large differences in the incomes of people withthe same level of education.

� The evidence presented in this report shows that it is misleading toassume that a university degree is the route to high skilled, high wagedemployment. While some graduates are able to capitalise on theirhigher education to secure managerial and professional jobs, manyothers are forced to enter jobs for which they are over-qualified andunder-paid (in terms of the expectations of what they thought aBachelor’s degree would deliver). The prospect of significant debtresulting from their ‘investment’ in higher education will only add to afeeling that the ‘system’ has let them down.

� While the abolition of university fees has been ruled out by the currentLabour Government in Westminster, the findings of this report suggestthat a better way needs to be found to share the ‘risks’ involved inpursuing higher education. The issue of shared risk must also take intoaccount periods of economic downturn and rising graduateunemployment.

Given the current limitations on what a viable system of funding forhigher education in Wales could look like, one option is to adopt theprinciple of ‘no win, no fee’. This has the major advantage over theEnglish system of removing the sense of personal debt. There is nopersonal debt attached to an individual who has been through highereducation in Wales (although it may be difficult [..] to apply this toWelsh students who study in England). But if an individual is inemployment that meets the appropriate earnings threshold - if they‘win’ in the sense of earning a significant graduate premium - they willpay a ‘knowledge dividend’ through the tax code while they are earningat that level. They would continue to pay this dividend until both feesand state support with subsistence are covered. Those from deprivedbackgrounds could be offered non-repayable grants to cover part oftheir subsistence costs or fees, and the overall amount payable couldbe adjusted for those working in ‘key’ services.

� This report also highlights the need for a major review of social andeconomic data sources within Wales. It is essential that the LabourForce Survey, British Skills Survey, etc. include a significant Welshsample to better understand trends in employment and the labourmarket within Wales, and how these compare with trends in other partsof Britain and beyond. There is also a need to coordinate all theresearch on education, skills and employment that is currently beingconducted on Wales.

xiv

� Further international comparative research is also required. Virtually allthe developed economies have pursued the same skills agenda in thebelief that advanced knowledge and skills are the key to economiccompetitiveness. However, how the supply of graduates is absorbed(or not) into national labour markets and its consequences for graduateincomes, employment and career prospects is an important area forfuture research.

� The expansion of higher education has been largely welcomed, butthere has been little research on the impact of mass higher educationon the labour market in Wales and England. An unintendedconsequence may be to further disadvantage those without a degree,as employers asked for degree level qualifications for jobs that werepreviously taken by those who had not entered higher education. Thisis also an issue that requires further investigation.

1

Main Report

1. Purpose

The purpose of this report is to conduct a comprehensive review of graduateemployment in both Wales and beyond, to enable the National Assembly ofWales to develop an effective and fair system of student support.

This report also considers implications for policy and identifies areas for futureresearch.

2. Background

The Westminster Government has primarily justified the expansion and futurefunding of higher education on economic grounds. It is assumed that there isa growing demand for high skilled workers and that graduates typically enjoya significant knowledge dividend compared to non-graduates.

Existing research evidence has already called into question governmentassumptions about the relationship between education, jobs and rewards1 IfWales is to develop an effective and fair system of student support, it willneed to be evidence-based. There is a need for a systematic review of theevidence, both within Wales and throughout the UK. This focus is not onlyimportant as a ‘benchmark’ for Wales, but because large numbers of Welshstudents work in other parts of the UK at some stage in their working lives.

This research will investigate a number of large-scale quantitative data sets inorder to examine trend data in the demand for graduates in Wales andbeyond. It will also investigate trends in graduate incomes. This will gobeyond an analysis of rates of return between graduates and non-graduatesto include differences amongst those with graduate qualifications. Genderdifferences will also be examined, as existing evidence shows that althoughthere are more women in higher education, men typically earn higherincomes.

This report is based on evidence derived from seven Briefing Notescommissioned to examine various facets of the changing graduate labourmarket, and different data sets applicable to Wales and the UK. However, thisreport is not limited to the findings presented in the Briefing Notes, and itsfindings do not necessarily reflect the views of their authors. There is noattempt to arrive at a ‘consensus’ as to the nature of graduate employment,future earnings or occupational trends. Indeed, we felt it important to reflectsome of the conflicting evidence and contrasting interpretations of thechanging graduate labour market.

While this report presents an overview of the key issues, it should be read inconjunction with the briefings notes that offer a more detailed account of thecomplexities confronting researchers and policy-makers in this field.

2

The briefing notes are as follows:

Briefing No.1. A re-analysis of the evidence used by the government insupport of changes in student funding in englishuniversitiesDr Anthony Hesketh, School of Management, LancasterUniversity.

This briefing paper is based on a re-analysis of the data used by theGovernment in support of the introduction of tuition fees in England, includingthe idea that 80 per cent of new jobs would require graduate level skills andthe claim of a 50 per cent graduate premium. The existing classifications ofoccupations and future skill demands in Wales and the UK are re-analysed toget a clearer picture of what constitutes ‘knowledge’ or ‘graduate’ work and toassess future trends.

Briefing No.2. Education and the knowledge economy: someobservations on the debate over graduate IncomesProf. Hugh Lauder, School of Education, University of Bath.Dr. Muriel Egerton, Institute for Social and EconomicResearch, University of Essex.Prof. Phillip Brown, School of Social Sciences, CardiffUniversity.

This briefing paper offers a new analysis based on data from the NationalChild Development Study, British Cohort Study and the General HouseholdSurvey. These data are used to examine the relationship betweenqualifications and income. Different data sets are harmonised to examinetrends in graduate incomes over the last decade for both males and females.It also considers variations in the findings from different data sets, and whatthese data can and cannot tell us about changes in graduate incomes overtime. The evidence presented in this briefing serves to question some of thecore assumptions of what is known as the ‘skills bias’ thesis.

Briefing No.3. A review of the literature on graduate employment,underemployment and unemploymentClaire Smetherham, School of Social Sciences, CardiffUniversity.

This paper reviews the existing research literature and findings on graduateemployment, underemployment and unemployment. It includes a discussionof class, gender and racial differences where applicable. In addition, itexamines the conceptual confusion that currently exists around the idea of a‘graduate’ job, ‘graduate glut’ and ‘new’ graduate occupations.

3

Briefing No.4. Graduate Pay in WalesDr Johnny Sung, Centre for Labour Market Studies, Universityof Leicester.Prof. David Ashton, Honorary Professor, School of SocialSciences, Cardiff University.

This briefing is based on a new analysis of the Labour Force Survey,comparing the incomes of graduates and non-graduates in Wales and the UKover time. It examines changes in the relative pay of graduates based ongender, age, industrial sector and occupation. It also investigates within groupdifference in graduate earnings.

Briefing No.5. An Analysis of Graduate Demand, Employment andIncome Wales.Prof. Ralph Fevre, School of Social Sciences, CardiffUniversity.

This briefing investigates the available evidence within Wales on graduatedemand, incomes and employment, including an analysis of higher educationand labour market statistics for the Principality. It also presents a regionalanalysis of Wales and the English regions that challenges some of theexisting accounts of the relative performance of Wales within the UK.

Briefing No.6. The Geography of Work Skills: A Focus on WalesProf. Alan Felstead, Centre for Labour Market Studies,University of Leicester.

This briefing presents new evidence on the skills demanded and used atwork, drawing on the British Skills Survey data for the nine English regions,Wales, and Scotland. The paper also tracks how skill demand indicators havechanged over the last decade. Particular attention is paid to the relativeperformance of Wales.

Briefing No.7. Welsh Graduates and the Welsh Labour MarketMike Tomlinson, School of Social Sciences, Cardiff University.

This briefing reviews the first destination data relating to Welsh graduates,including income and occupational differences related to institution, genderand ethnicity. It also examines the flows of Welsh graduates working in Walesand England.

4

3. Findings

3.1. What many graduates earn will not match their expectations.

On average, students expect to earn around £19,000 in their first job afteruniversity.2 This is expected to rise to £29,500 after five years.3 A recent studyof student expectations found that 89 per cent agreed that the money spenton their education was a good investment in their future, although 62 per centalso agreed that the increasing numbers of graduates entering the job marketwill make it harder to get a job.4

Figure 1: Job hunter’s expected starting salaries:

7%

15%17%

20% 20%

9%6%

4%

0%

5%

10%

15%

20%

25%

less than12,500

12,500-14,999

15,000-17,499

17,500-19,999

20,000-22,499

22,500-24,999

25,500-24,999

30,000 ormore

Salaries £

Perc

enta

ge o

f fin

alis

ts %

Source: The UK Graduate Careers Survey 2004, High Fliers Research/The Times.5

Figure 1 shows considerable variation in what graduates expect to earn intheir first job, with 10 per cent expecting to earn over £25,000 and 22 per centless than £15,000. Therefore, almost a quarter of graduates did not expect toreach the threshold set by the English Government for the repayment ofuniversity fees.

However, when graduate expectations are compared to actual salaries, westill find that many graduates will earn less than they expect. The HigherEducation Statistical Agency (HESA) reported an average of £17,000 perannum in 2003 (including full-time and part-time), but the NatWest surveyreported a 7 per cent rise in graduate starting salaries, from £12,659 in 2003to £13,600 in 20046. However, the Association of Graduate Recruiters thatrepresent leading blue chip organisations reported a median starting salary of£21,000 in 2004.7

5

The reported differences in starting salaries found in these surveys reflect theproblem of collecting accurate information for the full range of graduateswithin the labour market, and also highlight the problem of interpretation. It is,for example, difficult to assess whether a low starting salary reflects therealities of the labour market for graduates or a personal decision to findtemporary employment before looking for a ‘graduate’ job.

What is unequivocal is the fact that graduates will enter the labour market withincreasing personal debt, although there are different estimates concerningthe extent of their indebtedness. According to recent figures from NatWest,which roughly match other estimates (e.g. NUS) but exceed governmentcalculations, the average graduate debt has risen by £4,055 since 2003 to£12,1808. Five years ago, it was £3,700. The proportion of graduatesleaving university with debts of more than £10,000 has almost doubled, from31 per cent in 2003 to 59 per cent in the summer of 2004. A third believe itwill take them more than 10 years to pay off their university debts, an increaseof 13 per cent on those surveyed in 2002. This year’s new students believethat they will have to find £950.32 a month to cover essentials such as rent,food and books, which means that they will require £26,000 to coveruniversity expenses over a three-year degree. According to studentsinterviewed for the NatWest study, the overall cost of a three-year degree nowstands at £23,800, up from £19,400 last year.

The latest survey figures from Barclays Bank show that student debt has risenfive-fold since 1994.9 The average student now leaves HE owing £12,069through credit cards, bank and student loans. Barclays warned that if thecurrent trend continues, average student debt could top £33,708 after a three-year degree by 2010, taking into account plans to charge variable tuition feesof up to £3,000 a year from 2006.10

Finally, the Student Living Report 2004 found that the average debt thatstudents expected to leave university with has risen to £9,341 in 2003, a £525increase from 2002 and a £1,208 increase from 2001.11 Similar findings werereported in the DfES study by Callender and Wilkinson, which showed thatanticipated average debt on graduation increased by two and a half timessince 1998/99 to £8,666.12 Half of all final year students anticipate leavinguniversity with debts averaging over £9,673.

3.2. Most graduates will earn more than non-graduates, but the‘premium’ is often much lower than the 50 per cent difference inlife - time earnings reported in the White Paper on The Future ofHigher Education

It is widely claimed that university graduates enjoy a significant wage premiumover non-graduates that more than compensate for any loss of earnings or thecost of fees and subsistence while at university. The White Paper on theFuture of Higher Education in England states ‘Those who have been throughhigher education in the UK earn, on average, 50 per cent more than thosewho have not, and the rate of return from higher education in the UK is higherthan

6

in any other OECD country’.13 At the time when the Higher Education Bill wasgoing through parliament there was much talk of a £400,000 graduatepremium in lifetime earnings compared to those who had not gone touniversity.14

Figure 2: Average returns to education: private real rates of return,university-level education, 1999-2000

0

2

4

6

8

10

12

14

16

18

20

United

Kingdo

m

United

States

Denmark

France

Netherl

ands

Sweden

German

y

Canad

aJa

pan

Italy

men

w omen

Note: The average rates of return to tertiary education are calculated by comparing the benefits andcosts with those of upper-secondary education.

Source: Blöndal, S., Field, S., and Girouard.15

Figure 2 shows the rates of returns to education for individuals with universitylevel education in a number of countries. Although this figure shows that therate of return is higher for university graduates than non-graduates in all theselected countries, the value of the credential premium is subject to significantvariation between countries. The returns to men in Britain is shown to be overtwice those for Italy, Japan and Canada.

There are a number of problems with this evidence. Firstly, variations in thereturns to education in different countries highlight the fact that therelationship between credentials, jobs and rewards is far from straightforward.It is difficult to believe that knowledge and skills are valued less in Japan,Germany and Sweden than in Britain and the United States. Secondly, thefigures for Britain and the United States may not reflect the higher valueattached to ‘knowledge’ but the inferior quality of job opportunities for non-graduate jobs, compared to those available in other countries.

7

Thirdly, the graduate premium is based on past returns and there are noguarantees that future returns will remain constant or increase.16 Indeed, themore the domestic and global supply of highly qualified people increases, thegreater the chances that graduate incomes may decline over time.17 Fourthly,much of the evidence is based on the ‘average’ graduate, but as we will show,this ignores significant differences in the ability and willingness of graduates tocash in on their investment in their ‘human’ capital.18 Finally, one-off surveysof rates of return offer little indication of trends over time, and especially theimpact of educational expansion on graduate and non-graduate incomes. Forthis we require longitudinal data that is not readily available. To overcome thisproblem a number of new analyses were commissioned. The first is based onthe National Child Development Study (NCDS), the British Cohort Study(BCS70) and the General Household Survey (GHS) (For a detailed discussionsee Lauder, et al. Briefing No.2).

Table 1 shows a clear gradient in earnings by qualification, although bothmales and females with two+ A levels earn more than those with sub-degreessuch as HNDs or Foundation degrees. In 2000, those with a degree or higherearned 47 per cent more than those with no formal qualifications and 19 percent more than those with two+ A levels. The premium for female degreeholders was 58 per cent when compared to those without qualifications,dropping to 19 per cent compared to those with two+ A levels.

Lauder and his colleagues also show that graduate weekly earnings arehigher in 2000 than 1991, with the percentage increase greater for women,although women earn less than men in all categories. While women’searnings have broadly increased, that of non-graduate men has declined,except for the lowest qualified, who may have gained from the introduction ofthe minimum wage. However, this increase in graduate earnings is primarilydue to longer work hours, rather than an increased demand for skill. Over the1990s, there was a four-hour increase in working hours (amounting to anextra half-day’s work per week).

8

Table 1: Net weekly earnings (full-time employees only) by cohort bygender by qualification level

Women Men

Qualification 1991 2000 Difference2000-1991

1991 2000 Difference2000-1991

None 162 199 37 219 256 37

Below CSE grade 1 179 198 19 255 249 -6

O-levels and CSE 1 201 218 17 269 263 -6

1 A-level and NVQ3 208 220 12 286 283 -3

2+ A-levels 254 265 11 339 316 -23

Sub-degree 255 260 5 324 307 -17

Degree and higher 299 314 15 368 377 9

All 225 248 23 294 290 -4

N 1622 2580 3538 3909 Source : NCDS (1991) and BCS70 (2000):earnings adjusted to April 2001 prices. Briefing No.2

Table 2: Graduates and non-graduates: net hourly pay (employeesfull-time and part-time) 30 year olds.

All All Women Men

1991 2000 1991 2000 1991 2000

Low 3.6 3.7 3.4 3.6 4.0 3.9

Median 5.4 5.5 4.7 5.2 6.0 5.7

Below Degree

High 8.8 8.8 7.7 8.2 9.4 9.4

Low 5.1 4.9 5.0 4.6 5.4 5.3

Median 7.8 7.5 7.7 7.0 8.1 7.9

Degree or higher

High 12.0 12.0 11.3 10.8 12.5 13.0Source : 1991 National Child Development Study 1958 and 2000 British Cohort Study 1970Wages indexed to April 2000 prices. Sample does not include the self-employed.Low =10th percentile, High = 90th percentile. Adjusted for age (see Briefing No.2)

If we now analyse this data based on net hourly earnings (and adjust thefigures to take into account age differences between the two samples) we findthat median earnings for both male and female graduates have experienced aslight decline in real earnings.

In terms of the graduate premium, the lifetime difference between the mediangraduate and non-graduate was £166,000 (net income). In 2000, the mediangraduate earned 36 per cent more, rather than the 50 per cent premiumreported in The Future of Higher Education (2003:59). However, thisrepresents a 17 per cent decline, from a premium of £200,000 in 1991 (a dropfrom 44 per cent to 36 per cent) in the differential between the median

9

graduate and non-graduate. Table 2 also shows that high earning non-graduates achieved a 17 per cent advantage over the median graduates (alifetime difference of £108,000) and 80 per cent more than low earninggraduates.19 This analysis clearly demonstrates the problem of basing ourunderstanding of graduate incomes on the ‘typical’ graduate, a problemfurther highlighted when we examine differences in earnings amongst thosewith graduate qualifications.

3.3. There is evidence of significant and growing income differencesbetween graduates in the labour market

Drawing on the same data set, Lauder and his colleagues also found that thedifferences in graduate earning were twice as large as those found betweengraduates and non-graduates. This amounts to a £374,000 lifetime differencebetween median and high earning graduates, which equates to a 60 per centdifference. Since 1991, there has been an increase in graduate incomedifferentials, although this was slightly larger for males when the earnings ofmedian and high earning graduates are compared. A review of the evidenceon graduate earnings in the UK by Green et al. also found that the variance ofgraduate pay has increased since the mid-1980s.20

When broken down according to the 90th percentile (High), the Median and10th Percentile (Low), a new analysis of the General Household Survey ofthose between the ages of 26 and 35 also found that top earning malegraduates were racing away from the rest, although this was not the case forwomen (see Figures 3 and 4).

Figure 3: Male Graduates aged 26 to 35 - net weekly earnings(fulltime employees and self-employed)

0

200

400

600

800

1000

1200

1983-84 1991-92 2001

Lowest decileMedianHighest decile

Source: General Household Survey (Crown Copyright) earnings adjusted to 2001 prices.Briefing No.2

10

It can be seen from the above figure that for low earning male graduates therewas an increase in earnings between 1983/4 and 1991/2, but a slight declinebetween 1992/3 and 2001. For median earners there was an increase acrossthe years, while for high earners there was a substantial increase between1983/4 and 1991/2, after which their incomes take-off. Overall, the differencesin income between high, median and low earners increased.

For women the trends are different. For all income earners there is anincrease between 1983/4 and 1991/2, thereafter incomes remain the same forlow and median earners, while there was only a slight increase for highearners unlike their male counterparts. However, differences between high,median and low earners increase throughout this period for female graduates.

Figure 4: Women graduates aged 26 to 35 - net weekly earnings(full- time employees and self-employed)

0

200

400

600

800

1000

1200

1983-84 1991-92 2001

Lowest decileMedianHighest decile

Source: General Household Survey (Crown Copyright); earnings adjusted to 2001 prices. Briefing No.2

Data from the United States also show that high earning graduates havepulled further away from the rest, but unlike the British data, it also shows thathigh earning female graduates have made the most significant gains.

Based on evidence from the US, Figures 5 and 6 give a very different pictureto the mantra of ‘learning is earning’. Indeed, if the idea of a graduatepremium was calculated on the economic value of a college degree in 1973as opposed to its current market value, the ‘headline’ story would look verydifferent. These figures show that it is only male and female collegegraduates in the ‘higher earner’ category that enjoyed any growth in realincome since 1973. In other words, the vast majority of college graduateshave received no additional ‘premium’ on their investments in their humancapital compared to those in the 1970s, although they continue to earn morethan employees without a college education.

11

Figure 5: US male hourly wages by decile within educated groups,1973-2001

$0$5

$10$15$20$25$30$35$40$45$50

1973 1977 1981 1985 1989 1993 1997 2001

yea

Hourlyw

HS lowHS medianHS HighCollege lowCollege medianCollege high

Note: low, median and high earners refer to, respectively the 10th, 50th and 90th percentile wage.Source: Lawrence Mishel, Jared Bernstein, Heather Boushy, The State of Working America 2002/2003Economic Policy Institute.

Figure 6: US female hourly wages by decile within educated groups,1973-2001

$0

$5

$10

$15

$20

$25

$30

$35

1973 1977 1981 1985 1989 1993 1997 2001

yea

Hourlyw

HS lowHS medianHS HighCollege lowCollege medianCollege high

Note: low, median and high earners refer to, respectively the 10th, 50th and 90th percentile wage.Source: Lawrence Mishel, Jared Bernstein and Heather Boushy. The State of Working America2002/2003 Economic Policy Institute.

12

But even here the nature of the graduate premium is in question, as highearning non-college students were better paid than those college graduateson median incomes (although the differential has narrowed in recent years).As in Britain, female college graduates continue to earn less in eachequivalent category than males, but perhaps the most striking difference is theway male and female top earners have raced away from the rest. In theUnited States, they now earn over twice as much as the median collegegraduate of the same sex. This underlines the need to avoid talking about the‘average’ college or university graduate, for when it comes to rewards withinthe job market, some graduates are far more equal than others.21

Returning to the UK, the complexity of the changing relationship betweencredentials, jobs and reward can again be shown by drawing on data from theGeneral Household Survey. Here we find that while the incomes of lowearning graduates and median non-graduates rose during the 1980s, thepicture changed in the 1990s. While the non-graduate incomes continued torise for both women and men, they remain virtually static for male graduateson low incomes, and actually fell for low income female graduates (see Figure7).

Figure 7: ‘Estimated net weekly income for full-time employees only, bygender, qualification and decile (1983-2001)

142

211 205

150

201

228

179

234 237

203

249

275

125

175

225

275

1983-84 1991-92 2001

Women Lowestdecile graduateWomen Median non-graduateMen Lowest decilegraduateMen Median non-graduate

Source : General Household Survey (Crown Copyright). Briefing No.2

Evidence for Wales outlined in Briefing No. 4 by Sung and Ashton alsoprovides evidence of the contrasting fortunes of graduates within the labourmarket. Based on the Labour Force Survey, they calculate the ratio whichmeasures the difference between the earnings of those at the bottom of thelabour market, at the 10th percentile, and those at the top of the labour

13

market, at the 90th percentile (See Table 3). For example, the gross earningsof graduates at the 90th percentile in 1993 was £538 per week, divided bythose at the 10th percentile of £192 per week, produces a ratio of 2.80. By1997, this had increased to 5.56, falling back to 4.21 in 2001. Why theincomes for low earning graduates (10th percentile) drop dramatically in 1997is not altogether clear, but it does suggest that despite a modest recovery by2001, they were earning less than in 1993. At the same time, high earninggraduates (90 percentile) witnessed an acceleration in income between 1993and 1997 which has remained constant. This evidence suggests a significantwidening in earnings differential between those at the top and bottom of thegraduate labour market since the early 1990s. This adds further support to theargument that high earners were putting more distance between themselvesand other graduates.

Table 3: Earnings ratio between 10th percentile and 90th percentile,Wales 1993, 1997 and 2001*

A B B/A10 percentile 90 percentile

1993 192 538 2.80

1997 121 673 5.56With a degree

2001 160 674 4.21

1993 55 375 6.82

1997 62 408 6.58Without a degree

2001 89 475 5.34*Earnings – gross weekly pay in main jobSource: Labour Force Survey, Sung and Ashton, Briefing No.4

For the non-graduates, there is a different story in Wales. In 1993, thedifferential between those at the top and those at the bottom was muchgreater than that for graduates, namely a ratio of 6.82. But by 1997 it hadfallen to 6.58, and by 2001 to 5.34, an overall decrease of –1.48. Therefore,while the difference between the high earning and low earning graduatesincreased over this period, that between the non-graduates decreased.

14

Table 4: Earnings ratio between 10th percentile and 90th percentile,Wales 1993, 1997 and 2001*

A B B/A10 Percentile 90 percentile

1993 250.00 577.00 2.31

1997 173.00 750.00 4.34

Maledegree

2001 231.00 712.00 3.08

1993 125.00 442.00 3.54

1997 100.00 528.00 5.28

Femaledegree

2001 127.00 577.00 4.54* Earnings = gross weekly pay in main jobSource: Labour Force Survey, Sung and Ashton, Briefing No.4

There are important gender differences in the Welsh context that are revealedin Table 4. This table shows data for male and female graduates that highlightmajor differences in the incomes of top earning graduates at the 90thpercentile. It also shows that the differences in the earnings of high and lowearning female graduates were greater than those for men. Although notshown in the above table, these ratios changed the longer graduates were inthe labour market. The ratio for 30 year old male and female graduatesincreased from 3.52 to 4.82 between 1993 and 1997, but fell back to 3.32 by2001. However, the ratio for graduates in their thirties was 4.74 in 2001 and4.12 for those over 40. (See Sung and Ashton, Briefing No.4).

3.4. There are significant ‘sectoral’ and ‘occuptional’ differences ingraduate/non-graduate earnings ratios.

The income received by graduates depends on a range of factors explored inthis review. One such factor is the industrial sector graduate employees areengaged in. Table 5 shows that within the UK, ‘business services’ offer thehighest incomes for men, followed by ‘distribution’, whereas for women‘transport and communications’ come top of the list, followed by ‘businessservices’. This is based on the General Household Survey where the numbersof graduates is around 600, but it does provide further evidence aboutincomes over time (see Lauder, et al., Briefing No.2). The evidence heresuggests that women not only receive less than men in every industrial sector,but that during the 1990s, female incomes fell in a number of key sectors suchas ‘business services’, ‘transport and communications’ and ‘distribution’.

15

Table 5: Net weekly earnings of graduates aged 26-35 (full-timeworkers) by industrial sector [prices/year]

1983-84 1991-92 2001Industry

£ £ £

Primary, manufacturing, construction 266 364 392

Distribution 236 320 414

Transport and communications 337 454 381

Business services 290 376 516

Other services 257 318 352

Men

All Industries 267 348 417

Primary, manufacturing, construction 219 275 323

Distribution 207 352 275

Transport and communications 119 428 366

Business services 203 356 346

Other services 218 289 303

Women

All Industries 216 306 312Source: General Household Survey, Briefing No.2.

Further evidence on these issues can be gleaned from the Labour ForceSurvey (See Sung and Ashton, Briefing No. 4). This evidence is based on theworkforce over the age of 21 and compares 1993 and 2001. It is also basedon the difference between graduate and non-graduate median gross incomes.This is presented as a ratio: a ratio of 1.79 means that graduate pay is 79 percent more than that of non-graduates.

Table 6 shows a small decline in the graduate premium consistent with ourearlier findings. Again, it also highlights the complexity of the issues and thedangers associated with headline figures. The decline in the graduatepremium is explained by the decline, in the differential for women, as malegraduates experienced a minor increase in the differential over male non-graduates within the UK. It also shows that it is those in their thirties who havewitnessed a significant decline, as the ratio for those under 30 and over 40show little change.

This table also shows considerable variation based on industrial sector andoccupation. The median graduate in distribution, hotels and restaurants, andin ‘public administration, health and education, earn twice as much as non-graduates in the same industry, whereas it is around a half for those intransportation and communications. When we consider differences withinoccupational groups, the differentials are lower: managers and administratorsearned 50 per cent more in 2001, although these figures also show that those

16

in ‘professional’ occupations and ‘associate professional and technical’occupations where 10 and 14 per cent, respectively. Finally, what Table 6also reveals is a change in the ratio between 1993 and 2001, where theminuses indicate a decline in the graduate premium.

Sung and Ashton (Briefing No. 4) extend their analysis to consider sectoraland occupational differences within Wales. However, when the data arebroken down in this way, the small number of cases inevitably raisesquestions of validity and reliability. Nevertheless, they found major differencesin the graduate premium for males and females and those entering differentindustries and occupations in both Wales and the rest of the UK. In terms ofchanges over time, they also observe a general narrowing of the paydifferentials between graduate and non-graduate employees over this period.However, there are some noticeable exceptions between Wales and the widerUK sample.

In Wales, the decline in the differentials between graduates and non-graduates is more pronounced for those at the bottom of the labour market,especially among females and those in the 30-39 age range. It is the lowerpaid graduates among this group that have particularly lost out in relation totheir non-graduate counterparts. Some of this may be a result of femalegraduates returning to the labour market at lower income levels after breaksfor childrearing, but some of this change may also reflect a failure to find‘traditional’ graduate employment.

17

Table 6: Median weekly gross pay between degree holders andthose without in the UK (21 years old or older) 1993 and2001

Ratio1993

Ratio2001

Change in Ratio

Whole sample 1.92 1.79 -0.13

Gender

Male 1.55 1.65 0.10

Female 2.31 2.10 -0.20

Age

Less than 30 yrs old 1.44 1.43 -0.01

30 to 39 yrs old 1.81 1.71 -0.10

40 or older 2.08 2.10 0.02

Industry sector

Agriculture and fishing 2.23 1.96 -0.27

Energy and water 1.38 1.85 0.47

Manufacturing 1.71 1.67 -0.04

Construction 1.78 1.63 -0.15

Distri., hotels and restaurants 2.30 2.01 -0.30

Transport and communication 1.82 1.55 -0.27

Banking, finance and insurance 1.82 1.85 0.02

Pub. Admin., health and education 1.95 2.00 0.05

Other services 2.09 1.83 -0.27

Occupation

Manager and administration 1.57 1.50 -0.07

Professional occupations 1.20 1.10 -0.10

Asso. prof and tech occupations 1.24 1.14 -0.09

Clerical, secretarial occupations 1.18 1.13 -0.04

Craft and related occupations 1.80 1.51 -0.28

Personal, protective occupations 0.68 1.27 0.59

Sales occupations 1.92 1.47 -0.45

Plant and machine operatives 1.04 1.36 0.33

Other occupations 0.98 0.89 -0.09

N 7,832 15,532

Pay = gross weekly pay; * figures have been subject to rounding performed by Excel.The interpretation of the ratio: 1.92 means that graduates pay is 1.92 times more that that of non-graduates' pay.Change in ratio: -0.13 means that the differential between the two groups hasdecreased by 13% during the intervening period.

18

3.5. There are significant regional differences in graduate incomes

Evidence on graduate incomes in Wales, as opposed to other regions of theUK, is provided by the Graduate Market Trends Salary Survey - which againdraws on the Labour Force Survey. This shows that median graduateincomes in Wales (£19,292) fall short of those in Scotland (£19,656) andsignificantly lag behind London (£25,012). The higher salaries paid in Londonand the South East lifts the overall average for the UK, and consequentlyexaggerates the differences in graduate earnings in England and Wales(Fevre, Briefing No.5).

Table 7: Earnings (£) of graduates in Wales over time compared withother regions

Region Age 20-24 Age 25-29 30-34 Mean MedianWales 13,564 19,883 22,493 19,657 19,292Scotland 15,524 20,005 28,646 22,548 19,656London 20,052 28,838 37,710 30,646 25,012All regions 16,170 22,573 29,807 24,235 21,474Source: The Graduate Market Trends Salary Survey (GMTSS) spring 02, Tomlinson, Briefing No.7.

When income data are broken down by region, we find that graduate earningsfor those aged 21-30 in Wales are comparable to some of the English regionsincluding North East, Yorkshire and Humberside, East Midlands; andsignificantly higher than Northern Ireland (See Table 8). However, by the agesof 31-40, those in the comparable English regions had pulled ahead - leavingWales below all the English regions, but still ahead of Northern Ireland.

19

Table 8: Average weekly earnings (£) of full-time graduateemployees by region of place of work, and age

Age: 21-30 Age: 31-40Region

All degrees All degreesNorth East 374 601North West 409 574Yorkshire & Humberside 374 587East Midlands 375 584West Midlands 411 613East of England 451 650London 532 836South East 464 697South West 406 614Wales 379 557Scotland 404 599Northern Ireland 352 511Workplace outside UK 288*All regions 445 670*Weighted sample size less than 30,000; Source: Labour Force Survey 2002-03 Survey (as reported inGraduate Market Trends, Autumn 2003; Fevre, Briefing No.5)

None of these data take into account differences in the cost of living in Wales,as opposed to London or any of the other English regions (Fevre, Briefing No.5). As reported in Graduate Market Trends (autumn, 2002), the ComparativeStudies Unit (CSU) collected regional average salary data from the ProspectsToday salary and vacancy survey and then used a cost of living index basedon house prices to ‘rationalise’ graduate earnings. House prices were usedbecause they represented a substantial slice of household spending and werea good indicator of localised inflation. In Briefing No.5, Fevre suggests thatthis is a rough and ready way to factor in the cost of living, and other earningsdata might have been more reliable. But the attempt to make some allowancefor the cost of living effect on the overall rankings has a dramatic effect. It isalso worth noting that since 2002, house prices in Wales have continued torise while they have slowed in other parts of the UK, especially London andthe South East. Nevertheless, Wales was ranked nine out of 10 regions in thebase data but three out of 10 when the cost of living was taken into account.London dropped to the bottom of the table.

20

Table 9: 'Real' graduate earnings by region

AverageHousePrices (£)

Percentagedifferencefrom the basevalue (North)

Indexfigure

Averageearnings (£)(Ranking)

Real earnings byregionrationalised bycost of livingindex (£) (ranking)

North 69,337 0 1.0 (5) 17,471 (1) 17,471

North West 78,394 13.06 1.021 (10)16,486 (6) 16,147

Yorkshire andHumberside

76,526 10.37 1.017 (3) 17,566 (2) 17,272

Wales 75,341 8.66 1.014 (9) 16,844 (3) 16,611

West Midlands 97,554 40.70 1.065 (6) 17,398 (4) 16,336

East Midlands 88,833 28.11 1.045 (8)16,917 (5) 16,189

East Anglia 113,781 64.10 1.103 (4) 17,500 (7) 15,866

South West 127,010 83.18 1.133 (7) 17,088 (9) 15,082

South East 157,284 126.84 1.203 (1) 18,549 (8) 15,419

Greater London 214,001 208.64 1.334 (2) 18,413 (10) 13,802Source: Graduate Market Trends, Autumn 2002. Fevre, Briefing No.5

3.6. Existing research evidence also shows that graduate earnings

reflect a number of other factors, including the status of theuniversity attended, the degree subject studied, and degreeclassification.

The empirical evidence shows that graduates from higher status or higherranking institutions, those from vocational and/or numeric disciplines (e.g.Engineering or Mathematics) and those with a First - compared to a Lower-Second Class degree classification fare comparatively better in terms of theirearnings, both on initial entry to the labour market and over time.

Research evidence shows that some of the observed variation in graduateearnings reflects the status of the institution attended, whether it is an ‘old’ or‘new’ (post-1992) university, or an institution’s ranking in published leaguetables. Of course, differences in earnings will not only reflect institutionalstatus, but the ‘social mix’ of the students. High ranking universities typicallyhave a larger proportion of students from privileged social backgrounds, withthe personal, cultural and social capital valued by established graduateemployers (see below). Lower ranked universities typically have morestudents from the ethnic minorities, and more mature students or distancelearners, often studying on a part-time basis.

Because students come from a variety of backgrounds and with a wholerange of different experiences of work and the labour market, this willinevitably impact on their occupational and career decisions, along with theirpotential life-time earnings. Evidence continues to show that the large, blue

21

chip employers, who offer ‘fast track’ management training schemes and highstarting salaries, still tend to target graduates from the more elite end of theHE spectrum. In other words, the opportunities that exist for different groupsof graduates in the labour market are often highly conditioned by the status ofthe institution that they have attended.

Chevalier and Conlon (2003) show that the financial returns associated withdegree level qualifications vary substantially, according to the type of HEIattended. They find that even after controlling for academic achievement,subject of degree and family background, graduating from a Russell Groupinstitution (compared with a Modern university) leads to an earnings premiumranging from 0 to 6 per cent for men and 2.5 per cent for women in theyounger cohorts that they examined. It is important to note, however, thatthese researchers do not take into consideration regional or sectoraldifferences in the jobs undertaken by graduates - which are likely to accountfor large differences in their earnings.

The Sunday Times’ ‘Good University Guide’ (2004) also shows that the 10institutions with the worst starting salaries are also (with the exception of oneRussell Group institution – Cardiff University in Wales) lower statusuniversities. In fact, half of the institutions that appear in this list are WelshHEIs. As Table 10 shows, the average starting salary for graduates fromAberystwyth University was just £12,968 in 2004. Even graduates fromCardiff University, the only Russell Group institution in Wales, obtainedaverage starting salaries of £14,219 - less than the average starting salariesof graduates from the North East Wales Institute or Central Lancaster.

Table 10: Top ten worst graduate starting salaries by institution

Institution Starting Salary (£)

Aberystwyth 12,968

Chester 13,531

Bath Spa 13,613

Stirling 13,627

Lincoln 13,809

Swansea Institute 13,965

Bangor 14,103

Cardiff 14,219

Central Lancashire 14,327

North East Wales Institute 14,388Source: HESA/The Sunday Times Good University Guide

The existing research evidence also shows clearly that the subject studied atdegree level has a significant effect on graduate earnings. The generalconsensus is that graduates with Arts/Humanities or Social Sciencequalifications have relatively low earnings premiums relative to graduates

22

from other disciplines. Those with vocational or numeric degrees, on theother hand (e.g. engineering, business studies, mathematics), tend to farecomparatively well in the earnings stakes. Such differences in earnings arealso related to the industrial sectors of the economy in which graduates fromdifferent degree disciplines tend to work. Even amongst those with the samehigh level credentials (Firsts), there are significant differences in the earningsof graduates from different degree disciplines, even after five years in thelabour market, as shown in Table 11.

Table 11: Average annual income of respondents with Firsts by areaof degree study (1997 cohort)

Arts /Humanities

NaturalSciences

SocialSciences Vocational Maths/

Computing Totals

Count % Count % Count % Count % Count % Count %

Under£19,999

19 50 10 28 8 31 2 3 6 17 45 23

£20,000 -£29,999

14 37 14 39 9 35 21 34 9 25 67 34

£30,000 -£39,999

2 5 9 25 4 15 17 28 7 19 39 20

Over£40,000

3 8 3 8 5 19 21 35 14 39 46 23

Totals 38 100 36 100 26 100 61 100 36 100 197 100

Source: Smetherham (2005)

This table shows that even after five years in the labour market, those withArts/Humanities degrees fare comparatively worse in terms of earnings thanthose with ‘Vocational’ or ‘Maths and Computing’ degrees. Half of those with‘Arts or Humanities’ degrees are still earning under £19,999 after five years inthe labour market, compared to just three per cent of those from vocationalcourses. In contrast, 35 per cent of those with vocational qualifications areearning over £40,000 after five years compared to just eight per cent of ‘Artsor Humanities’ graduates. Such evidence points to significant differences inearnings, according to subject or area of degree level study.

Walker and Zhu (2003) also found that individuals with Arts, Education,Languages or Social Science qualifications at HE level tended to haverelatively low earnings premiums – if one exists at all. Those with Law,Economics, Business or Maths degrees, on the other hand, earned up to 25per cent more than their equivalents with two or more A Levels.

What graduates study is closely tied to gender. Young men and womencontinue to opt for degree subjects according to traditionally genderedpatterns, so that women are more heavily concentrated in the Arts orHumanities disciplines than among the more quantitative or vocationalsubjects such as Engineering or Mathematics. This then impacts on earnings.Consistent with other findings presented in this report, Purcell (2002) shows

23

that women earn less on average than their male counterparts, irrespective ofthe subject in which they specialise. Indeed, for all 12 degree subjectsanalysed, male graduates had higher average earnings than their femalecounterparts three years after graduation. The difference between theearnings of comparable samples of men and women was widest (in terms ofactually earnings, as well as in percentage terms) for Law and Mathematicsand Computing graduates and narrowest for Education graduates.

Figure 8: Average Earnings of Graduates by Degree Subject

Source: Purcell, 2002 fig. 2

Before turning our attention to the relationship between degree classificationand earnings, it is worth noting that there is little evidence to support the claimthat students will follow ‘market signals’ as an indicator of what to study.Table 12 is based on the number of students enrolled in different subjectareas. This shows that students have not been deterred from studying theArts, Humanities, Education or Social Sciences, but, surprisingly, given -much of the rhetoric about a knowledge-driven economy, the overall numbersin Engineering and Technology and Physical Sciences has declined since themid-1990s (although part-time enrolments have increased).

24

Table 12: Table of HE subjects and differences over time (UK)

Total HE entrants 1996/97 Total HE entrants 2002/3Subject of study

Total FT PT Total FT PTDifferenceOver time

Medicine andDentistry 40476 27707 117 48915 33345 155 + 8439

Subjects relating tomedicine 153105 82306 53207 266415 130960 10092

5 + 113310

Biological sciences 81750 59853 4477 125860 92340 8215 +44110Veterinary sciences 3301 2487 19 3840 3165 40 +539Agriculture andrelated subjects 15226 11327 779 14980 9485 1795 -246

Physical sciences 74496 51031 4638 71040 47510 6335 -3456Mathematicalsciences 19908 15528 393 26190 19470 1135 +6282

Computer sciences 73612 49417 11188 134035 85535 9970 +60423Engineering andtechnology 134041 90930 14835 131575 76855 18185 -2466

Architecture,building andplanning

46545 27059 8248 45830 21600 11510 -715

Social studies 129328 77666 18017 168920 94310 36950 +39592Law 54767 34281 7396 77680 44435 9430 +22913Business andadministrativestudies

222321 119241 39671 284550 148160 41950 +62229

Mass Communic.anddocumentations

20004 12933 1892 42175 30135 3095 +22171

Languages 91214 61514 16678 130225 77735 36050 +39011Historical andphilosophicalstudies

60656 33594 13590 87630 46935 23725 +26974

Creative arts anddesign 87170 71460 6727 132675 104620 14010 +45505

Education 136888 51296 18359 169800 35020 43570 +32912Combined 311371 118031 174715 212785 9700 18628

0 -98586

Total 1756179 2175115 +418936Source: Higher Education Statistical Agency (HESA)

In terms of degree classification, research has also shown significantdifferences in the earnings of those with different levels of credential. Higherdegree classifications carry a premium over lower degree classifications.Research has shown that after five years in the labour market (looking at1997 graduates surveyed in 2002), the earnings of those with Firsts and 2.2sare significantly different, as shown in Table 13 below.

25

Table 13: Average annual income of respondents by degreeclassification (1997 cohort)

1st 2.2 Totals

Count % Count % Count %

Under £19,999 47 23 43 30 90 26

£20,000 - £29,999 71 35 67 47 138 40

£30,000 - £39,999 41 20 24 17 65 18

Over £40,000 46 22 9 6 55 16

Totals 205 100 143 100 348 100Source: Smetherham, 2005

Research also shows that this premium is greater for men than for women,and that the gap between the earnings of men and women with the samecredentials widens with increasing levels of achievement. In other words, thegender gap between the average earnings of female and male graduates withequivalent degree classifications is largest for those with Firsts; narrowsslightly between men and women with 2.1s; decreases further when looking atmen and women with 2.2s; and is smallest between those with Third Classdegrees.22

These findings are supported by Smetherham’s (2005) findings that alsoreveal significant differences between the earnings of men and women withthe same First Class degrees after five years in the labour market. Table 14shows that over twice as many female graduates were earning under£20,000, while half of the male graduates were earning over £30,000,compared to a third of the women.

Table 14: Average annual income of respondents with firsts bygender (1997 cohort)

Male Female TotalsCount % Count % Count %

Under £19,999 18 15 28 32 46 23

£20,000 - £29,999 40 35 30 35 70 34

£30,000 - £39,999 27 23 14 16 41 20

Over £40,000 31 27 15 17 46 23

Totals 116 100 87 100 203 100Source: Smetherham (2005)

26

Flat rates of tuition fee operate on the assumption that rates of return tograduates will somehow be standardised evenly across the graduatepopulation. But the existing research evidence shows that rates of return aredependent on each of the factors considered above (institution attended,subject studied and degree classification), as well as the complex inter-relationships between them, thus rendering highly problematic anyassumption of a ‘standard’ or ‘average’ graduate premium or rate of return.

3.7. Major gender differences remain in areas of specialist knowledge,industrial sector, occupation and income

Reported differences in the rates of returns for male and female graduateswarrant further investigation, given that debates around tuition fees have alsolargely ignored the importance of gender. All of the recent published studieson the relative earnings of male and female graduates have consistentlyshown a large and significant discrepancy in earnings between men andwomen, despite improvement in women’s achievement at degree level andsome advances in accessing professional jobs within the labour market.Young women today are equally as likely to enter HE and have exceededaverage male performance and achievement at undergraduate level. Womennow comprise over half of the undergraduate student population and for thefirst time, in 2001, more women than men gained First Class degrees (SeeSmetherham, Briefing No.3).

Despite this improvement in female academic performance at degree level,further research by Purcell shows that income differences between men andwomen remain important. This research shows that average male graduateearnings were approximately 15 per cent higher than those for women,graduates under 30 years of age, but these income inequalities rise to over 40per cent from the age of 45.23

Table 15: Average annual earnings of graduates, 1999-2000

Males (£) Females (£) % by which averagemale earningsexceeded those offemales

20-24 16,738 14,592 14.7

24-29 23,302 20,154 15.6

30-34 30,448 24,939 22.1

35-39 36,948 27,774 33.2

40-44 36,949 26,691 38.4

45-49 36,696 26,113 40.5

50-54 38,153 26,549 43.7

All ages under 60 32,555 23,630 37.8Source: Labour Force Survey, September 1999 – August 2000 (pooled data), Purcell (2002)

27

Whether these differentials will narrow as more women enter professional andmanagerial occupations is open to debate, but Purcell’s research found thatfor graduates under 40 years of age, there were more female graduates in theprofessions than men, and only a slightly higher proportion of men employedas Managers and Administrators.

Figure 9: The occupational distribution of graduates

Source: Purcell, 2002 fig. 7

This does not tell us anything about the levels of seniority held by women andmen in these occupations, which led Purcell to examine the relationshipbetween occupation and subject studied at university. The occupationaldistribution of graduates by their disciplinary background show someimportant differences between men and women graduates with degrees inLanguages, Engineering and Business Studies (See Table 16).

Table 16: Occupational distribution of graduates by disciplinarybackground

Subject of degree

Languages Engineering Business Studies

Major occupationalcategory of currentemployment

Males Females Males Females Males Females

Managers andadministrators

31.6 29.8 23.3 17.9 38.7 48.3

Professional 34.5 27.0 43.0 37.4 18.7 13.5

Associate Professional 15.3 13.8 17.8 18.5 18.7 11.8

Non-manual / clerical 16.1 20.6 3.8 7.9 14.6 16.7

Manual / routineoperative

2.5 8.8 12.1 18.3 9.3 9.7

All (100%) 478 1,446 4,715 519 2,913 2,261Sample: Full-time employed 1995 graduates aged under 40 in winter 1998-99Source: Moving On dataset, Purcell, 2002 Table 3

28

In all but the Business Studies sub-sample, male graduates were more likelythan their female counterparts to be managers and administrators or inprofessional occupations (i.e. in the highest paid and highest statusoccupational categories). What is more, in all of the categories, over aquarter of female graduates were in non-manual/clerical or manual/routineoperative jobs (i.e. ‘non-graduate’ jobs).

The research evidence also shows that female graduates are alsoconcentrated in a small number of industrial sectors and more likely to work inthe public sector, rather than private sector.24

Table 17: Distribution of employment of graduates by majoreconomic sector.

Sector of Employment Males Females

Public 23.5 43.8

Private 74.6 51.9

Voluntary 1.9 4.3

Total weighted sample (n =100%)

25,557 24,084

Sample: Full-time employed 1995 graduates aged under 40 in 1998-99Source: Purcell, 2002, Moving On dataset. Smetherham, Briefing No.3

This table shows that nearly half of the female under 40s in the 1995 cohortworked in either the public or voluntary sectors, whereas men weresubstantially more likely than women to be employed in the private sector.Three quarters of men were employed in the private sector, compared toapproximately half of the female graduates. But close to twice the number offemales worked in the public sector (23.5% and 43.8% respectively). Purcellalso notes that women in the public and voluntary sectors were more likelythan men to have been employed on permanent or open-ended -as opposedto fixed-term - contracts.

29

Figure 10: Distribution of graduates by industry sector

[Source: Purcell, 2002 fig. 5]

The existing research evidence also highlights very different distributions interms of gender by industry sector. These data show that the employment offemale graduates is far more concentrated than that of their male peers, withmen more likely to be employed in the primary and secondary sectors of theeconomy and women in the service sector (as in the labour market as awhole). Female graduates are most likely to be employed in Public Services,Education and Business Services, which together account for roughly 70 percent of the female graduates surveyed.

The existing research data all show clearly that gender continues to constitutea key variable in terms not only of income, but also in terms of industrialsector and occupation. Furthermore, this is the case despite overallequalisation in female and male academic performance at degree level.

This analysis also shows that the labour market that graduates enter is notcharacterised by a hierarchy of jobs that correspond to a hierarchy ofacademic achievement. It points to enduring questions of ‘positional’advantage based on gender as well as academic performance. This picture isfurther complicated by issues of social class, race and ethnicity.

3.8 Class and ethnic inequalities also continue to be a feature ofeducational and labour market outcomes

The DfES Future of Higher Education White Paper states that:

‘The social class gap among those entering higher education is unacceptablywide. Those from the top three social classes are almost three times as likelyto enter higher education as those from the bottom three.’25

30

Figure 11 below, which is cited in the ‘Future of Higher Education’ WhitePaper, shows that the working classes have not benefited from the expansionof HE to the same extent as the middle classes. The higher education entrygap between the top and bottom three social classes has - in fact - widenedover time, and the differences between young people from professional andunskilled backgrounds is even more pronounced. Young people fromprofessional backgrounds are over five times more likely to enter highereducation than their counterparts from unskilled backgrounds.

Figure 11: Higher education entrants by social class groups (1960–2000)

Moreover, class difference in access to higher education continues to matterwithin the job market. A recent report by the Institute of Employment Studies(Pollard, Pearson et al. 2004) suggests that students from families with noexperience of university fare least well after graduating. These researchersdrew on the experiences of 1,500 young people who applied to enter HE in1998, the first year of tuition fees, and who were surveyed for a third time in2003 two years after graduating.

According to Pollard et al’s (2004) report, many graduates were surprised athow difficult it was to find work, and a substantial minority went into temporaryor low-quality jobs. Although for some this was for ‘breathing space’, or tostart to pay off debt, for others it was due to ‘a lack of viable alternatives’.Those earning the higher salaries - in what they perceived to be good qualityjobs - tended to be male, from higher socio-economic groups, and fromfamilies with higher incomes. Those in poor quality jobs tended to be from‘non-traditional’ backgrounds – poorer homes or areas – and to have doneless well at university. The report states that “those who moved to study andwork were the most likely to find ‘good’ jobs; those who returned to their home

31

region to look for work had a particularly difficult time”. They argue that this inturn reinforces the deprivation trap, because increasingly youngsters aretending to study at their ‘local’ university and live at or near the family home tosave money: ‘Costs were a significant issue for many’. They found that theaverage graduate debt of the graduates they surveyed was almost £10,000 –with those on low family incomes and who studied away from home havingthe highest debt.26

The participation of ethnic minority groups in HE varies considerably, withsome having a higher participation rates than for white students. But despitethese high participation rates, ethnic minority students do not ‘on average’perform as well at degree level as white students (Connor, 2004). Ethnicminority groups also fare less well in the labour market.

In terms of their initial destinations, First Destination Survey (FDS) data showthat initial employment outcomes are poorer overall for ethnic minority groups(in aggregate) than for white graduates. However, this disguises differencesbetween ethnic minority groups. This applies to differences within each ethnicminority group, but there is also variation between ethnic minority groups. Inaddition, outcomes are further influenced and complicated by characteristicssuch as social class and gender. Ethnic minority graduates were both morelikely to have enrolled for further study or training and to be unemployed sixmonths after graduation when compared to their white counterparts. Table 18shows the higher initial unemployment rates for full-time degree students fromall minority ethnic groups. The ethnic minority average was just over 11 percent compared to nearly half that, just over six per cent, for white graduates.27

The unemployment rate for the ‘mixed ethnic’ category was 8.3 per centcompared to 13.9 per cent for Pakistanis.

32

Table 18: First destinations of full-time degree (home domiciled)graduates from English universities, 2001/02 (percentagesare based on known destinations)

UK

pai

d w

ork

Ove

rsea

s w

ork

UK

sel

f-em

ploy

ed

Stud

y o

r tra

inin

g

Une

mpl

oyed

Oth

er

Not

ava

ilabl

e fo

rem

ploy

men

t or

trai

ning

Tota

l num

ber

White 63.4 2.2 1.6 17.4 6.1 2.8 6.5 118,069

All minority ethnic groups 55.3 1.0 1.1 22.4 11.1 4.2 5.1 19,851

Black Caribbean 63.8 0.4 1.3 14.3 9.8 4.8 5.7 1,200

Black African 54.3 0.4 1.2 21.6 12.2 5.9 4.4 1,950

Black Other 60.1 1.5 1.5 17.4 10.8 3.3 5.4 666

Indian 58.9 0.5 0.7 22.5 10.2 3.3 3.9 6,762

Pakistani 51.5 0.4 1.4 24.0 13.9 3.8 5.1 2,722

Bangladeshi 53.0 0.2 0.6 23.2 12.4 4.4 6.1 862

Chinese 47.7 2.8 1.1 26.7 11.6 5.1 5.1 1,485

Asian Other 52.7 1.2 1.5 23.9 9.7 4.8 6.1 1,631

Mixed ethnic 46.7 2.9 2.9 26.4 8.3 5.4 7.4 242

Other 52.6 2.3 1.7 21.6 10.3 4.4 7.1 2,331

All (known ethnicity) 62.2 2.0 1.5 18.1 6.9 3.0 6.3 137,920

Source: HESA (FDS), taken from Connor et al. (2004) p.89, Table 7.1

3.9. There is evidence of graduate ‘over-qualification’ but littleagreement about how to measure it

A number of studies have found that approximately 30 per cent of graduatesare ‘over-qualified’ in terms of what is required to fulfil their currentemployment. Several different studies have found that a substantial minorityof graduates are over-educated and/or their skills are under-utilised in theirjobs, particularly during the first few years after graduation (Belfield, Bullock etal., 1997; Battu, Belfield et al., 2000; Dolton and Vignoles, 2000;Smetherham, 2005).

The likelihood of being over-educated or under-utilised does appear to declineover the first few years after graduation, as some graduates manage tosecure graduate-level employment after first accepting a period of lower-levelemployment (Connor, La Valle et al. 1997; Elias and McKnight 1999).Nonetheless, the initial period of over-education or under-utilisation may havelingering negative effects on future salary and career prospects (Green,McIntosh et al. 2002).

33

Indeed, Green et al. (2000) argue that an estimated 30 per cent of UKgraduates have more education than their job requires six years aftergraduating. Sloane et al. (1999) also found that 40 per cent of graduateswere over-educated six years after graduating, using survey data from theUniversity of Birmingham. Purcell and Elias are also reported as stating thatone third of the 1995 cohort of graduates that they surveyed in 2002 felt that adegree was not required for their job.28

Smetherham (2005) shows that after five years in the labour market, lookingat a 1997 cohort of graduates - those with Firsts were significantly more likelyto be in a job for which a degree was a formal entry requirement, compared tothose with 2.2s, although even a large minority of those with Firsts remainedin jobs for which a degree was not a formal entry requirement. Nearly half ofall respondents with 2.2s, after five years in the labour market, were in a jobfor which a degree was not a formal entry requirement (although the cell sizesare relatively small).

Table 19: Whether a degree was a formal entry requirement forrespondent’s current job by degree class (1997 cohort)

1st 2:2 TotalCount Col % Count Col % Count Col %

A degree was a formal entryrequirement 150 75.4 72 51.4 222 65.5

A degree was not a formalentry requirement 47 23.6 60 42.9 107 31.6

Don't know / not sure 2 1.0 8 5.7 10 2.9

Total 199 100 140 100 339 100Source: Smetherham, Briefing No.3

The evidence on whether over-qualification for graduates has increased overtime is open to different and often conflicting interpretations. However, mostagree that there may have been a modest increase over the last few years onthe broad definition of over-qualification, namely whether graduates areworking in jobs for which a degree is a formal entry requirement.

Research by Mason (2002) shows that in retailing, the proportion of graduatesin clerical, sales and other occupations below associate professional levelrose from 22 per cent to 37 per cent between 1988 and 1998, according toestimates based on LFS data. In transport and communications, theproportion of graduates classified to occupations below associate professionallevel in 1998 was one in three, up from 26 per cent 10 years earlier.

In Briefing No.2 Lauder and his colleagues discuss some of the differentdefinitions of what constitutes graduate work. For example, one way ofdefining graduate work is to ask graduates whether they feel that they are

34

utilising the knowledge and skills commensurate with their degree studies.Most of the research in this area has adopted such an approach. Another iswhen employers are asked whether they have recruited graduates toundertake graduate level work, or whether they have employed them for otherreasons such as their social confidence, reliability or punctuality. Smetherhamalso discusses some of the different methodologies used for examininggraduate over-qualification (see Briefing No.3).

3.10. Claims that half of those entering the workforce will require highereducation to meet the requirements of the ‘knowledge’ economyhave little empirical support on either side of the Atlantic. There isalso little evidence to support the view that Wales is a vibrantknowledge-driven economy.

In the White Paper, The Future of Higher Education, it is asserted that the‘demand for graduates is very strong’. This is supported by research evidencepurported to show that 80 per cent of the 1.7 million new jobs which areexpected to be created by the end of the decade will be in occupations whichnormally recruit those with higher education qualifications.29

These figures are based on UK occupational data analysed by Wilson andcolleagues.30 However, as Hesketh shows in Briefing No.1, it is not at all clearthat the Government’s claim of 80 per cent of the 1.7 million jobs createdbetween 2000-2010 is even supported by Wilson’s analysis. Far fromsupporting the Government’s interpretation of their data, Wilson and his teampresented a note of caution:

‘[I]t seems likely that many more highly qualified individuals will have to findjobs outside the ‘traditional’ areas of graduate employment, many of which arenow approaching saturation point in terms of graduate penetration rates. Thismay often be at the expense of less well qualified people, who may bedisplaced into less attractive jobs.31 On closer examination, of the 2 million ofso extra new jobs expected to be created, only 900,000 - or just over four inevery 10 - are expected to be located in professional occupations, thetraditional destinations of graduates.

It is important to note that projecting occupational demand over a five or tenyear period is as much of an art as a science. Given these uncertainties,Wilson and his colleagues present different scenarios in relation to the futuredemand for higher-level qualification (NVQ 4 and 5) (See Hesketh, BriefingNo. 1). The Department of Education and Skills assumes that the demand forgraduates will remain in balance with expected supply, whereas what is calledthe fixed-penetration scenario alternatively assumes that the proportionqualified in each occupation will remain fixed at 1999 levels.32 If the fixed-penetration scenario became a reality, we would expect to experienceincreasing problems of over-education as the labour market fails to absorbgrowing number of university graduates.

35

Based on Wilson’s analysis, Table 20 shows the projected increase indemand for employees by level of education. The predicted increase indemand for those requiring NVQ4 (32.4%) and NVQ5 (30.8%) looksuperficially impressive. But when these increases are presented as a shareof all UK jobs in 1999, NVQ4 and 5 combined represented around a quarter ofthe workforce, rising to little under a third (31.8%) in 2010.

Table 20: Percentage increases and percentage shares of economycompared

1999

(000s)

2010

(000s)

Change inJobs 1999-2010 (000s)

% Change inJobs1999-2010

% ShareUK Jobs1999

% ShareUK Jobs2010

Higher degrees(NVQ5)

1237 1618 381 30.8 4.5 5.5

First degree andequivalent(NVQ4)

5862 7809 1900 32.4 21.2 26.3

A level andequivalent(NVQ3)

5101 5377 276 5.4 18.5 18.1

GCSE GradesA-C (NVQ2)

6173 6120 -53 -0.9 22.4 20.6

GCSE BelowGrade C (NVQ1)

5780 6254 474 8.2 21 21

No Qualification 3389 2494 -895 -26.4 12.4 8.4

Total 27546 29673 2127 7.7 100 100Source: Extrapolated from Wilson, 2001, Table 5.11: 76. Hesketh, Briefing No. 1

Other evidence also points to the danger of assuming that there is a linearincrease in the demand for graduates. Figure 12 presents evidence from theAssociation of Graduate Recruiters (AGR), which represents around 400 ofBritain’s leading graduate employers. The figures show considerablefluctuations in the demand for graduates. It also shows that these ‘elite’employers recruited approximately 25,000 graduates in 1997/8 but failed tomatch this in any subsequent year.

36

Figure 12: Blue chip graduate vacancies in the UK 1997-2003

0

5,000

10,000

15,000

20,000

25,000

30,000

1997/98 1998/99 1999/2000 2000/01 2001/02 2002/03

No o

f vac

anci

es

-40

-30

-20

-10

0

10

20

% c

hang

e on

pre

viou

s ye

ar

Total Vacancies % increase/decrease

Data extrapolated from AGR Graduate Recruitment Surveys 1997-2003.

The relatively small proportion of graduates recruited to large blue chipemployers at a time when over 300,000 graduates enter the labour marketeach year, raises key questions about the demand for professional,managerial and higher-level technical skills.

In order to obtain a longer term view of changes in the occupational structureof men and women in employment, Elias and Purcell use information from theNew Earnings Survey, a survey that has remained relatively unchanged since1975.33 The results of their analysis are shown below, first for maleemployees and second for female employees. Elias and Purcell utilise fivedifferent categories of graduate job: ‘non-graduate’ jobs, ‘niche’ graduatejobs, ‘new’ graduate jobs, ‘modern’ graduate jobs and ‘traditional’ graduatejobs. A more detailed discussion of these categories and their constructioncan be found in Smetherham, Briefing No.3.

37

Figure 13: Changing occupational structure of male employees inemployment, 1975 to 2000 by SOC (HE)

0

5

10

15

20

25

30

35

40

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

%

Non-graduate job

Niche graduate job

New graduate job

Modern graduate job

Traditional graduate job

Source: new Earnings Survey Panel Dataset, 1975 - 2000

Figure 14: Changing occupational structure of female employees inemployment, 1975 to 2000 by SOC (HE)

0

5

10

15

20

25

30

35

40

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

%

Non-graduate job

Niche graduate job

New graduate job

Modern graduate job

Traditional graduate job

Source: New Earnings Survey Panel Dataset, 1975 – 200; Elias and Purcell, 2004: 34

38

These figures show how the occupational structure of employment(specifically, employees in employment) has changed over this 25-yearperiod. Both the ‘new’ and ‘modern’ graduate categories have shownsignificant growth, as indicated by the rising proportion of both male andfemale employment classified to these categories. Although using suchgeneral definitions as ‘modern’ and ‘new’ graduate jobs is problematicbecause it is not always clear whether these are really ‘graduate’ jobs, thesefigures show that the proportion of employment in the traditional graduate jobcategory has remained virtually constant throughout the period considered.

Despite such issues, much of the policy debate has focussed on issues ofsupply, including concerns about whether graduates have the employabilityskills that employers are looking for. If the demand for highly skilled workers isconsistent with the supply then there is little reason for concern, as theexpansion of education could be seen to mirror the technological upgrading ofthe workforce. The problem for the proponents of the knowledge economy isthat the evidence fails to support the argument that today’s economy requiresa significant, let alone exponential, increase in demand for high skilledworkers.

Table 21: Employment by education and training category, 2000-2010.(Monthly Labour Review)35

Most significant source of education or training Number Percentdistribution

2000 2010 2000 2010Total, all occupations (thousands) 145,954 167,754 100,0 100,0

Bachelor's or higher 30,072 36,556 20,7 21,8

First professional degree 2,034 2,404 1,4 1,4

Doctoral degree 1,492 1,845 1,0 1,1

Master's degree 1,426 1,759 1,0 1,0

Bachelor's or higher degree, plus work experience 7,319 8,741 5,0 5,2

Bachelor's degree 17,801 21,807 12,2 13,0

Associate degree or postsecondary vocational award 11,761 14,600 8,1 8,7

Associate degree 5,083 6,710 3,5 4,0

Postsecondary vocational award 6,678 7,891 4,6 4,7

Work related training 103,760 116,597 71,3 69,5

Work experience in a related occupation 10,456 11,559 7,2 6,9

Long term on-the-job training 12,435 13,373 8,5 8,0

Moderate term on-the-job training 27,671 30,794 19,0 18,4

Short term on-the-job training 53,198 60,871 36,5 36,3Source: Daniel E. Hecker, Employment Outlook: 2000-10 Occupational employment projections to 2010Monthly labour Review November 2001 p.83

39

Evidence from the United States - seen by many commentators as thearchetypal ‘knowledge’ economy - shows that although many of the jobs withthe fastest rate of growth in the American economy require a college oruniversity degree, most Americans will continue to work in occupations thatrequire low levels of education and training. Table 21 shows that theproportion of the workforce requiring a bachelor’s degree or higher isexpected to rise by just over one percentage point between 2000 and 2010.This amounts to around one in five of all jobs, whereas the majority of theworkforce will remain engaged in activities that require only work-relatedtraining. Indeed, there will continue to be a buoyant demand for large numberof workers in jobs requiring little training. These jobs will continue to accountfor over half the American workforce.

In Briefing No.1, Hesketh presents a new analysis of the demand for‘knowledge’ workers, based on what employees actually do rather than thequalifications held by employees or what is required to get a particular job.36

This analysis is based on categories originally formulated by Robert Reich(1991). The category of ‘knowledge workers’ (Reich called ‘symbolicanalysts’), include corporate managers and administrators;managers/proprietors in agriculture and services; science and engineeringprofessionals; teaching professionals; health professionals; other professionaloccupations; science and engineering associate professionals; healthassociate professionals; and other associate professional occupations.37

The category of knowledge workers is distinct from in-person service workersthat include secretarial occupations; personal service occupations; buyers,brokers and sales representatives; other sales occupations; and also fromroutine production workers that include clerical occupations; skilledconstruction trades; skilled engineering trades; protective service occupations;industrial plant and machine operators and assemblers; drivers and mobilemachine operators; other occupations in agriculture, forestry and fishing; andother elementary occupations.

This analysis shows a significant increase in knowledge-intensive jobs in the1980s and 1990s. From approximately 20 per cent of the workforce in 1971to 28 per cent by the beginning of the 21st century, but there is little evidenceof a further increase to 2010.

40

Figure 15: Job Classifications – all UK Jobs 1971-2010

0

10

20

30

40

50\60\

70\

1971 1981 1991 1999 2005 2010

%AllUKJobs

KW IP RPSource: Hesketh, Briefing No.1

The question of the future demand for knowledge workers led Hesketh to considerthe internal rates of increase/decrease for each of the three categories of workersand for the UK economy as a whole. Figure 16 shows that during the 1980s, theinternal rates of increase for knowledge workers where four times that of the growthin the total number of jobs in the UK economy. However, this rate of change is notmaintained in the 1990s and there is little expectation of a further acceleration in thedemand for knowledge workers in the near future.

Figure 16: Percentage changes within categories between periods

2.2

27

18

76

16

25

19

12.511

-3.6-5

-4-1.6 -1

0.2

6 64 4

-10

-5

0

5

10

15

20

25

30

1971-81 1981-1991 1991-99 1999-2005 2005-2010

KW IP RP UK Economy

Source: Hesketh, Briefing No.1

41

Turning to the labour market in Wales, Hesketh’s analysis shows that theexpansion in the total number of jobs has failed to match the UK as a whole(seven per cent in Wales, compared to 13 per cent for the UK) between 1971-1999. Based on the same threefold classification of knowledge workers, in-person services and routine producers, Figure 17 shows the composition ofthe labour force in Wales. This shows that the increase in knowledge workersas a proportion of the Welsh economy has expanded, but not at the rateexperienced in the UK. There are now some 288,387 knowledge workers inWales, as opposed to below 200,000 in 1971. This compares with a shift inthe total number of jobs during the same period from 1 million to 1.2 million.This analysis suggests that knowledge workers will account for approximatelyone in four of the Welsh workforce at least until 2010.

Figure 17: Job classifications – all jobs 1971-2010 (Wales)

0

10

20

30

40

50

60

70

80

1971 1981 1991 1999 2005 2010

% a

ll jo

bs in

Wal

es

KW IP RP

Source: Hesketh, Briefing No.1

Such findings inevitably raise questions about the rhetoric of the knowledgeeconomy both in Wales and beyond. But what is often absent in discussionsof employment demand is the role of the public, rather than the private sector.There is an assumption that it is the demand for knowledge and skills in theprivate sector that is driving the demand for university graduates. The realityis that the public sector has been a crucial source of job creation within Britainsince the turn of the century. Stephen Hicks and his colleagues (2005) haveundertaken a detailed analysis of public sector employment that shows -within the UK, it accounts for one in five of all jobs. However, between 2001-04 there was an overall increase of 368,000 additional jobs created in thepublic sector in the UK, compared to just 226,000 in the private sector. Theimportance of recent investment in public sector employment is shown inFigure 18.

42

Figure 18: Annual employment increase: head count

0

0.5

1

1.5

2

2.5

3

2000 2001 2002 2003 2004

Per c

ent

Public State

Source: Annual Employment Increase: Head Count, from S.Hicks et al. Public Sector Employment(2005). www.statistics.gov.uk/cci/article.asp?id=1095

The public sector has not only been important in terms of the volume of jobsthat have been created, but also in terms of the quality of employment. This isobviously important in the context of graduate employment, as many of thesenew public sector jobs have been in the health service and education, whichtypically demand higher levels of skills and better access to training thanoffered by private companies within the service sector.

Indeed, one of the most striking features of the Welsh labour market forgraduates is the extensive role of the public, rather than the private sector,when compared to other regions of the UK. Wales is the only region wherethere are a higher percentage of graduate employees in the public sector (51per cent) than in the private sector (49 per cent). In Wales, the percentage ofgraduates in the private sector is approximately 18 percentage points belowthe regional average (67.1%), and almost 28 percentage points below London(See Tomlinson: Briefing No. 7).

43

Table 22: Distribution of full-time graduate employees in the private andpublic sectors in each region by place of work

Region Private % Public %Wales 49 51

North West 61.5 38.5

Scotland 62.8 37.2

London 76.8 23.2

All regions 67.1 32.9Source: Labour Force Survey (spring 03), Tomlinson Briefing No.7

In London, less than a quarter of graduates were in the public sector in 2003.This suggests that there is a huge gap between the rhetoric and the realitiesof the knowledge economy in Wales, as the demand for higher skills is beingdriven as much by employment in the public services as the private sector,which is suppose to be driving business enterprise and innovation. Indeed,the number of self-employed in Wales fell by 28,000 in the year ending June2004, but at the same time the number of graduate business start-ups haveincreased.38 Therefore, as Fevre points out in Briefing No.5, the top graduatejobs in Wales (and the jobs in which graduates are most in demand), are lesslikely to be in blue chip companies and more likely to be in publicadministration, health and education.

Because the private sector is smaller in Wales, graduate managers andsenior officials - along with graduates in associate professional and technicaloccupations - have less of a presence. Forecasters at the Institute forEmployment Research, University of Warwick, think it unlikely that Wales willcatch up to the UK average for these occupational groups by 2012. One of themajor reasons is the lack of a financial centre and company headquarters.

A survey by Coombes, Page and Wilson (2003) for HEFCW, also reported byFevre, illustrates the importance of the public sector to graduates who want towork in Wales. More than one in three of this group were teachers, asopposed to 18 per cent for those in graduate jobs outside Wales. According tothe Institute for Employment Research, this pattern will continue for at leastanother decade. The largest gains in new graduate jobs are expected forteaching and research professionals, with jobs for health professionals alsoincreasing (rates of growth are forecast to be above the UK average for bothgroups). Slower rates of employment growth than the UK average areexpected in private sector graduate jobs for science and technologyprofessionals and business professionals. Amongst associate professionals,health associate professionals are forecast to have a growth rate similar to theUK average and, of course, many of these will be employed in the publicsector.

44

3.11. There is little agreement about the extent or nature of skill gaps inWales

The Future Skills Wales (FSW, 2003) is based on information gathered frommore than 6,000 individuals and 6,000 organisations across Wales betweenFebruary and July 2003. 39 FSW includes data on skills gaps - the differencebetween the skills employees have and those employers require them to haveto undertake their jobs - and compares skill gaps and shortages in Wales withother parts of the UK. According to the findings of the FSW, 19 per cent ofemployers in Wales reported skill gaps. This figure is higher than in Scotland(16%) and Northern Ireland (13%) but below that in England (23%). Of thoseemployers reporting skill gaps, the most common problem cited was lack of ITskills, followed by communication skills.

Nearly half of all Welsh employers recruiting school and college leavers reporta significant skills gap between the skills school and college leavers have, andthose that employers need. The skills perceived to be lacking includecommunication, initiative, numeracy, literacy and understanding customerneeds. The existence of skills gaps amongst graduates is less evident thanthat for school and college leavers, with a third of employers recruitinggraduates reporting gaps, particularly in communication, initiative, workexperience and business awareness. The proportion of employers reportingskill gaps rises as organisation size increases, and the highest proportion ofemployers reporting skill gaps are found in manufacturing, followed by publicadministration, education and health, together with banking, finance and otherservices.

Interestingly, when employees are surveyed rather than employers, someemployees report having higher levels of skills than employers require. Someof these are in the same sectors where employers perceive deficiencies inworkplace skills. Moreover, for skills including communication, problemsolving and showing initiative, individuals give themselves a rating above thelevel employers report currently requiring. Employers meanwhile report gapsin the skills of the current workforce in these very skills! This may indicate adegree of overestimation of abilities and/or under-utilisation of available skills.It also highlights methodological problems associated with the definition ofskill gaps and shortages.

While Future Skills Wales presents a buoyant picture of graduate demand inWales,40 Felstead presents a pessimistic picture of skill demand in thePrincipality (Briefing No.6). Table 23 shows the difference between thenumber of jobs requiring qualifications at various levels and the number ofeconomically active individuals holding qualifications at each of these levels.41

45

Table 23: Qualifications demand and supply, Wales, 2001

Demand SupplyHighest Qualification Required1 Highest Qualification Held2

‘000s of jobs (% of existingjobs)

‘000s of people (% ofeconomicallyactiveindividuals)

NVQ Level 4 or above 273 24.7 326 28.0Degree 107 9.7 185 15.9Non-degree 166 15.0 141 12.1NVQ Level 3 228 20.6 261 22.4NVQ Level 2 146 13.2 262 22.5NVQ Level 1 86 7.8 144 12.4No qualifications 374 33.8 170 14.6Source: 2001 Skills Survey, own calculations, Felstead, Briefing No.6.

This shows that there are more people qualified at NVQ Level 4 (28%) thanthere are jobs demanding that level of certification (24.7%). According toFelstead’s calculations, this amounts to a surplus of approximately 53,000students. This mismatch is mainly due to the difference between supply(15.9%) and demand (9.7%) at the level of bachelor’s degree, rather that atthe sub-degree level (that is also included at level 4). Even more disturbing isthe ‘skills gap’ at the bottom end of the labour market, which suggests thatover a third of jobs in Wales require no formal qualification, at the same timethat the vast majority of the workforce are qualified to at least NVQ Level 1.

There are at least two methodological issues raised by Felstead’s analysis.Firstly, the demand for skills is not based on employer accounts but on thehighest qualification that employees perceived to be required to get therespondent’s job at the time of interview. Secondly, it highlights the problem offinding a representative sample at the region level, rather than for Britain as awhole. The 2001 Skills Survey contains data on 4,470 individuals, but theWelsh data is based on only 224 respondents, which is why Felstead makes itclear that the data must be ‘treated with some caution’ (Briefing No.6).

When the evidence for Wales is compared to Britain as a whole, there aresome striking similarities as well as differences. Table 24 shows the nature ofqualification mismatch by taking the percentage of jobs where a specific levelof qualification is required, minus the percentage of the workforce qualified ateach level. Therefore, a positive figure indicates excess demand, while anegative figure indicates over-supply. The most surprising finding in bothWales and Britain is the mismatch at the bottom end of the labour market. Theproblem of ‘over-qualification is not restricted to the market for graduates, asthere appear to be many jobs that require no formal qualifications (+19.2percentage points in Wales and +15.2 in Britain). The problem of over-qualification appears to be less of a problem at Level 4 in Britain, although it isa problem within the Principality, as noted above. Again, it is important to

46

remember that these figures are based on that employees perceived to be thequalification required to enter a specific job, rather than whether people usetheir knowledge and skills when they are at work. The other surprising findingis the problem of ‘over-supply’ at Level 3 in Britain, as it has long been arguedby some that the major skills gap in the UK is the shortage of workers withintermediate skills (Level 3). This appears to be less of a problem in Wales.

Table 24: Patterns of qualification mismatch, Wales and Britain, 2001

Qualification Mismatch1

(% of jobs requiring qualifications at each level minus % of workforcequalified at each level)

Wales Britain

Level 4 or above -3.3 +0.3

Degree -6.2 -1.4

Non-degree +2.9 +1.7

Level 3 -1.8 -8.7

Level 2 -9.3 -4.9

Level 1 -4.6 -1.8

No qualifications +19.2 +15.21: A positive figure indicates excess demand, while a negative figure indicates over-supply.Source: 2001 Skills Survey, Felstead, Briefing No.6

Felstead also constructs a skills index based on a range of dimensionsincluding knowledge, manual skills, teamworking, communications and socialskills, computing and problem-solving (see Briefing No.6). This enables him toexamine changes in the skill profiles of different regions of Britain between1997-2001. Table 25 shows that Wales has moved from top spot in 1997 toalmost the bottom of the pile only Scotland had a lower aggregate score onthis index in 2001. Again, it needs to be remembered that each regionconsists of a relatively small number of cases, but Felstead concludes hisBriefing by suggesting that:

‘The results for Wales suggest a dramatic change in job content – withinthe space of a decade, jobs in Wales have become less skilled and theeconomy has shifted from high skilled to low skilled production. What isremarkable is that this pattern of results is robust to a variety of skillmeasures. Nevertheless, the data presented here are based on relativelysmall sample sizes. Larger surveys are urgently required in order tocorroborate and dissect (e.g., by sex) the pattern of results suggestedhere.’ (Briefing No.6).

47

Table 25: Particular skills index by region/country, 2001

Particular Skill Index1Region/Country

1997 2001

Change,1997-2001

Britain 2.410 2.518 +0.108

North East 2.257 2.509 +0.252

West Midlands 2.355 2.562 +0.207

South West 2.360 2.543 +0.183

Yorkshire andHumber

2.356 2.519 +0.163

East Midlands 2.348 2.506 +0.158

South East 2.448 2.551 +0.103

London 2.444 2.542 +0.094

East of England 2.448 2.506 +0.058

North West 2.451 2.506 +0.055

Scotland 2.428 2.449 +0.021

Wales 2.501 2.477 -0.0241: The Particular Skill Index is averaged across 36 common activities between 1997 and 2001.Source: 1997 Skills Survey and 2001 Skills Survey, Felstead, Briefing No.6.

3.12 There is a high degree of integration between the Welsh and Englishlabour market, with considerable movement of students and graduateemployees across national borders.

The overall proportion of students enrolling in 2002/03 in a university withintheir country of domicile is lower for Welsh students than any other BritishUniversities. The enrolment was 62 per cent for Welsh domiciled students,compared to 95 per cent for England and 94 per cent for Scotland. Thirty-seven per cent of Welsh students enrolled in English HEIs, although Waleshas 6,900 enrolments from students domiciled elsewhere in the UK (SeeTomlinson, Briefing No.7).

Table 26 is based on the Higher Education Statistical Agency’s (HESA) firstdestination survey. It shows that the proportion of Welsh domiciled graduatesentering employment in the Principality (66%) is lower than Northern Ireland(76%), Scotland (85%) or England (95%). Likewise, in terms of graduatesfrom Welsh HEIs a similar trend emerges as a little over half from Welsh HEIsenter employment in the Principality, compared to around three quarters ofthose from Northern Ireland and Scotland. The proportion for England wasover 90 per cent, given the size of labour market.

48

Table 26: The ‘retention’ rate of graduates entering employment bycountry of domicile and country of HEI attended

Country of destination Percentage of domicilegraduates

Percentage ofgraduates from country

of HEI

Wales 66 53

England 95 92

Scotland 85 79

Northern Ireland 76 74Source: HESA First destination Supplement 2001/02, Tomlinson, Briefing No.7

Further investigation of retention rates within employment shows that it isWelsh students that decide to study in England who are far more likely towork in another part of the UK. Table 27 shows that almost nine out of 10Welsh graduates who studied within Wales stay in the Principality, against 35per cent of those who studied in England. Moreover, almost a quarter of non-Welsh students who went to university in Wales found employment in thecountry.

Table 27: Employment destination of UK graduates gainingqualifications in UK IHEs by country of domicile, country ofIHE and location of first destination.

DestinationCountry of IHE Country of Domicile

Wales UK

Wales Wales 88 13

Wales UK 23 74

England Wales 35 61

England UK 0.4 93Source: HESA First Destination Supplement 2002/03, Briefing No.7

A study by Coombes, Page and Wilson (2003) based on a sample 14 monthsafter graduation (rather than the six month survey point used in the HESAFirst Destination Survey) found that threequarters of originally Welshdomiciled graduates were working in Wales. This represents an increasefrom 70 per cent in 1999. Coombes, Page and Wilson (2003) alsoinvestigated perceptions of graduate jobs in Wales. They found somepessimism about graduate wage levels, the availability of jobs graduates wantto do, career development prospects and the quality of jobs in their chosenfield. Here there was very little difference in perceptions according to countryof original domicile, with the exception of more negative perceptions of Welsh

49

pay levels amongst Welsh domiciled. However in light of increasing studentdebt, the perception and realities of lower wage rates in Wales may have animpact on the future destinations of Welsh graduate, as this study found that42 per cent of all those surveyed said that student debt had made wage ratesa more important factor in job choice.

For those in graduate jobs outside Wales, graduates originally domiciled inWales were much more likely to say they had searched unsuccessfully foremployment in Wales. Only one in four had not done so. For those who werenot originally domiciled in Wales, over 70% had not done so. Of those whosearched unsuccessfully for employment in Wales, a half cited factors such aslack of jobs in their chosen fields, unable to obtain jobs in chosen location,poor quality of jobs in chosen field and relatively low wage levels. Non-Welshdomiciles were more likely to cite lack of Welsh language skills as an obstacleto finding employment

A recent ‘milkround’ online survey of 1,500 graduates from across the wholeof the UK asked respondents their views of the Welsh graduate labourmarket. The results suggest that Wales is actually a desirable employmentlocation for many graduates, both of Welsh and UK domicile. However, thisfavourable view about working in Wales was offset by a view that it was moredifficult to find suitable employment in the Principality. Again, most of therespondents viewed the Welsh economy as offering less in earnings andcareer scope, and limited mainly to public sector employment.

Research by Drinkwater and Blackaby (2004) show that one in three Welshborn living in other parts of the UK have a degree, compared to less than onein four immigrants in Wales. They conclude that ‘Wales is losing adisproportionate share of it younger and more educated people, even aftercontrolling for other personal characteristics’ (2004:21). However, Fevresuggests that ‘fears of a brain drain of graduate talent from Wales may havebeen exaggerated. While recent graduates are disproportionately likely tostart their careers in major conurbations, and particularly in London, slightlyolder graduates seem to find Wales an attractive place to settle.’ Thisconclusion finds some support in the study by Coombes et al. (2003), notedabove, who found that only 8.4 per cent of those who originally lived in Wales,but who were in graduate jobs outside Wales had no intention of everreturning to the Principality. Similarly, only 9.1 per cent of the non-Welshoriginal domicile who were in graduate jobs in Wales had no intention of everleaving Wales (Briefing No.5). What seems clear from this analysis is thatmany more UK. graduates would be attracted to work and live in Wales, whichcould transform it into a genuine knowledge-driven economy.

50

4. Implications for policy and future research

This report has focused on the relationship between higher education and thelabour market. While the contribution of higher education to individualprosperity and economic competitiveness is of obvious importance, thisanalysis needs to be considered alongside the wider personal and socialbenefits of learning within higher education.

Much of the debate about ‘graduate’ employment is based on outdatedassumptions. The idea of a ‘graduate job’ and the ‘graduate labour market’had some validity in the 1960s, when the majority who entered universityfound themselves on ‘fast track’ management training programmes or enteredthe established professions such as law and medicine. It also provided auseful shorthand for jobs that offered high salaries, good promotion prospects,high social status and job security. While this picture of graduate employmentmay remain dominant in the brochures of leading graduate employers, it is farremoved from the occupational realities confronting many graduates.

We need to avoid the tendency to assume that graduate employment isdefined by what graduates do. People are paid for their employment, notcredentials, although employers may use credentials as a way of screeningcandidates from specific jobs. While there are an increasing number of jobsthat ask for degree level qualification, many of these jobs will not offer highsalaries or good promotion prospects. Both the diversity of the graduatepopulation and the diversity of graduate employment need to be at the heartof future theoretical and empirical research in this field.

The evidence presented in this report highlights the problem of using datafrom organisations such as the Association of Graduate Recruiters (AGR), asrepresentative of all graduates. The majority of graduates do not work forleading blue chip companies, but for small and medium sized organisations,which typically have lower salaries and fewer prospects for internal promotion.The data on graduate incomes also shows that a significant minority ofgraduates will struggle to repay their fees and in some cases, will never earnenough to rid themselves of debt amassed during their time in highereducation. This may prove to be an even bigger problem in Wales, asgraduate incomes are typically lower.

The report highlights the danger of using ‘average’ incomes that obscure therealities of today’s labour market which reveals large differences in theincomes of people with the same level of education. Therefore, whateverfunding and fee regime is introduced in Wales (or reformed in England), itneeds to be sensitive to differences within and across the student andgraduate population. In particular, the research evidence on graduatepremiums and the returns to different graduates has practical implications forthe appropriate level at which student loan repayment thresholds are set,

51

given these vast ‘inequalities’ in graduate earnings. It is also important to beextremely careful about ‘average’ graduate earnings because they disguiseconsiderable variation in outcome according to, for example, the complexinter-relationships between gender, social class, subject choice etc., notedabove.

The evidence presented in this report shows that it is misleading to assumethat a university degree is the route to high skilled, high waged employment.While some graduates are able to capitalise on their higher education tosecure tough entry jobs, many others are forced to enter jobs for which theyare over-qualified and under-paid (in terms of the expectations of what theythought a Bachelor’s degree would deliver). The prospect of significant debtresulting from their ‘investment’ in higher education will only add to a feelingthat the ‘system’ has let them down. This is an issue that not only requiresfurther empirical investigation, but conceptual and theoretical discussionbased on an interdisciplinary framework. This is necessary to overcome theexisting confusion over what constitutes ‘over-qualification’ or ‘over-education’.

While the abolition of university fees is off the current political agenda, thefindings of this report suggest that a better way needs to be found to share the‘risks’ involved in pursuing higher education. The issue of shared risk mustalso take into account periods of economic downturn and rising graduateunemployment. It must also be remembered that past rates of return offerclues, rather than a guarantee, of returns to a university education in thefuture.

Much of the evidence presented in this report challenges the assumptions onwhich the Westminster government based its policy for student support.However, the high degree of integration of both higher education and thelabour market between England and Wales will make it extremely difficult forthe Welsh Assembly to depart in any significant way from the fundingarrangements in England. To fundamentally change the Welsh system, onehas to transform the English system. Therefore, although it is legitimate to talkabout Higher Education in Wales or the Welsh economy, they are inextricablylinked to institutional arrangements in England. Many Welsh students studyand work in England, although many return to the Principally.

Given these limitations on what a viable system of funding for HigherEducation in Wales could look like, one option is to adopt the principle of ‘nowin, no fee’. This has the major advantage over the English system ofremoving the sense of personal debt. There is no personal debt attached toan individual who has been through Higher Education in Wales (although itmay be difficult to apply this to Welsh students who study in England). But ifan individual is in employment that meets the appropriate earnings threshold,they will pay a ‘knowledge dividend’ through the tax code while they areearning at that level. They would continue to pay this dividend until both feesand state support with subsistence are covered. Those from deprivedbackgrounds could be offered non-repayable grants to cover part of their

52

subsistence costs or fees, and the overall amount payable could be adjustedfor those working in ‘key’ services.

This report also highlights the need for a major review of social and economicdata sources within Wales. It is essential that the Labour Force Survey, BritishSkills Survey, etc., include a significant Welsh sample to better understandtrends in employment and the labour market within Wales, and how thesecompare with trends in other parts of Britain and beyond. There is also a needto coordinate all the research on education, skills and employment that iscurrently being conducted on Wales.

Further international comparative research is also required. Virtually all thedeveloped economies have pursued the same skills agenda in the belief thatadvanced knowledge and skills are the key to economic competitiveness.However, how the supply of graduates is absorbed (or not) into nationallabour markets and its consequences for graduate incomes, employment andcareer prospects is an important area for future research.

The expansion of higher education has been largely welcomed, but there hasbeen little research on the impact of mass higher education on the labourmarket in Wales and England. An unintended consequence may be to furtherdisadvantage those without a degree, as employers asked for degree levelqualifications for jobs that were previously taken by those who had notentered higher education (Keep and Mayhew, 2004). This is also an issuethat requires further investigation.

Finally, whatever arrangements are introduced in Wales, a major review of thenew funding arrangements should be simultaneously conducted in both Walesand England at the earliest opportunity. This review must be based on thenew realities of employment, income and careers for all graduates, rather thana privileged minority.

53

Acknowledgements

I would like to thank all those who have contributed to the briefing papers.This has offered a unique opportunity to present a wide range of evidence ongraduates in the labour market that has been largely absent in policy debates.I would also like to thank Michael Tomlinson for the background notes heprepared for this report and Sandra Bonney who ‘works the magic’ in terms ofpresentation. Finally, I’d also like to thank Professor Ewart Keep, DeputyDirector, SKOPE, University of Warwick, for acting as a ‘critical friend’.

54

BibliographyAUT (2002) ‘Does Education Matter?’ AUT LOOK, Issue No 22, October

2002.

Barclay’s Annual Graduate Survey (2004).

Battu, H. and Sloane, P. (2000) ‘Over-education and Crowding Out in Britain’in L. Borghans and A. De Grip (eds.) The Overeducated Worker? TheEconomics of Skill Utilisation, Cheltenham: Edward Elgar.

Battu, H., Belfield, C. R. and Sloane, P. J. (2000). "How well can wemeasure graduate over-education and its effects?" National InstituteEconomic Review. 171: 82-93.

Beaver, W.R. (2000) ‘Cheapening College: sending more high schoolgraduates to college will improve neither their prospects nor theeconomy’. American Outlook Magazine, Summer 2000.(http://www.hudson.org/american_outlook/articles_sm00/beaver.htm)

Bekhradnia, B. (2003) Demand for Graduates: A Review of the EconomicEvidence, Higher Education Policy Institute,www.hepi.ac.uk/articles/docs/graduatesmain.pdf.

Belfield, C. R., Bullock, A., Chevalier, A. N., Fielding, A., Siebert, W. S. andThomas, H. R. (1997). Mapping the Careers of Highly QualifiedWorkers. Bristol, Higher Education Funding Council for England(HEFCE).

Brynin, M. (2000) ‘Graduate density, gender and employment’. British Journalof Sociology, Vol. 53, Issue No. 3, pp. 363-381.

Brynin, M. (2002) ‘Overqualification in employment’. Work, Employment andSociety, Vol. 16, No.4, 637-654.

Cappelli, P. (1995) ‘Is the “Skills Gap” Really About Attitudes?’ CaliforniaManagement Review, Vol. 37, No. 4, pp. 108-124.

Chevalier, A. N. and Conlon. G. (2003). Does it Pay to Attend a PrestigiousUniversity? London, Centre for the Economics of Education, LSE.

Connor, H., Tyers, C., Modood, T. and Hillage, J. (2004) Why the Difference?A Closer Look at Higher Education Minority Ethnic Students andGraduates. Department for Education and Skills.

Connor, H., La Valle, I., Pollard, E. and Millmore, B. (1997) What doGraduates do Next? Brighton. Institute for Employment Studies. 343.

DfES (2002) Key Messages from Skills in England. London: HMSO.

55

DfES (2003) Education and Skills: The Economic Benefit. London HMSO.

DfES (2003) The Future of Higher Education, White Paper. London; HMSO.

Dolton, P. J. and Vignoles, A. (2000). "The Incidence and Effects of Over-Education in the UK Graduate Labour Market." Economics ofEducation Review. 19 (2): 179-198.

Elias, P. and Purcell, K. (2004) A classification for studying the graduatelabour market, Research Paper 6. Bristol: University of the WestEngland & Warwick: Warwick Institute for Employment Research.(www.warwick.ac.uk/go/glmf)

Elias, P. and Purcell, K. (2004a) The earnings of graduates in their earlycareers, Research paper 5. Bristol: University of the West England &Warwick : Warwick Institute for Employment Research.(www.warwick.ac.uk/go/glmf)

Elias, P. and A. McKnight (1999). "Setting the scene: defining andcharacterising the career paths of higher education leavers". MovingOn: Graduate Careers Three Years After Graduation. Manchester,Careers Services Unit.

Employer of Choice. Net (2002) ‘Employer of Choice: A Consulting andLearning Process to Support Employers in the Drive to BecomeRecognized as Employers of Choice.’(http://www.employerofchoice.net)

EOC (2001) Men and Women in Britain: Professional Occupations.Manchester: EOC.

EOC (2002) Qualifications and Careers: Equal opportunities and earningsamong graduates by Kate Purcell. Working Paper Series No.1Manchester: EOC.

EOC (2004) ‘Plugging Britain’s skills gap: challenging gender segregation intraining and work: Report of phase one of the EOC’s investigation intogender segregation and Modern Apprenticeships’. London: EOC.

ESRC (2004) Graduate Oversupply Is a Media Myth, Study Shows. Economicand Social Research Council, Issue 16, June 2004, 31-32.

Extra References for Welsh Review

Forth, J and Mason, G (2004) Information and Communication Technology(ICT) Skill Gaps and Company-Level Performance: Evidence from theICT Professionals Survey 2000-01, National Institute of Economic andSocial Research, London, NIESR Discussion Paper No 236, March2004.

56

Future Skills Wales (2003) (http://www.sqilliaudyfodolcymru.com/index.php)

Graduate Prospects (2002) The graduate labour market. Manchester:Prospects (www.prospects.ac.uk)

Graduate Prospects (2004) Finalist’s career plans and expectations (Summer2004). Manchester:Prospects.(http://www.prospects.ac.uk/cms/showpage/home_page/labour_market_information/)

Green, F., Ashton, D., Burchell, B., Davies, B. and Felstead, A. (2000). "AreBritish Workers Becoming More Skilled?" The Overeducated Worker?In L. Borghans and A. De Grip The Economics of Skill Utilisation.Cheltenham, Edward Elgar: 77-109.

Green, F., McIntosh, S. and Vignoles, A. (2002). "The Utilisation of Educationand Skills: Evidence from Britain." The Manchester School. 70 (6):792-811.

Guardian (2003) ‘Crying out for graduates?’ The Guardian (Education),September 16, 2003.

Guardian (2003) ‘You might be richer than you think The Guardian,September 29, 2003.

Guardian (2003). ‘No problem’. The Guardian (Education) December 16,2003.

Guardian (2004) ‘Students face £2,000 extra debt’. The Guardian, May 12,2004.

Guardian (2004) ‘UK degrees boost pay by 59%’. The Guardian, September15 2004.

Handel, M.J. (2000) ‘Is there a Skills Crisis? Trends in Job SkillRequirements, Technology and Wage Inequality in the United States’.The Jerome Levy Institute of Bard College Public Policy Brief No 62,2000 (http:www.levy.org/docs/summary/sumspr01.html)

HEPI (2003) ‘Demand for Graduates: a review of the economic evidence’ byLibby Aston and Bahram Bekhrandnia.http://www.hepi.ac.uk/articles/docs/graduatesES.pdf

HEPI (2004) ‘Higher education supply and demand to 2010’, by Libby Aston.(http://www.hepi.ac.uk)

Hicks, S., Lindsay, C., Livesey, D., Barford, N., and Williams, R. (2005) PublicSector Employment, London: Office of National Statistics,www.statistics.gov.uk/cci/article.asp?id=1095

57

Howell, D. (2000) Skills and Wage Collapse. The American Prospect, Vol. 11,Issue No. 15., June 19, 2000-July 3, 2000.(http://www.prospect.org/print-friendly/print/v11/15/howell-d.html)

IER (2001) Projections of Occupations and Qualification. IER Bulletin: IER:University of Warwick.

Keep, E. (2004) ‘After Access: Researching Labour Market Issues’, in J.Gallacher (ed.) Researching Access to Higher Education, London:Routledge.

Keep, E. and Mayhew, K. (2004) ‘The Economic and DistributionalImplications of Current Policies on Higher Education’, Oxford Review ofEconomic Policy, 20, 298-314.

Livingstone, D.W. (1999) The Education-Jobs Gap: Under employment orEconomic Democracy. Toronto: Garamond Press.

Mason, G. (2001) ‘The Mix of Graduate and Intermediate-level Skills inBritain: What should the balance be?’ Journal of Education and Work,Vol.14, No. 1, 5-27.

Mason, G. (2002). ‘High Skills Utilisation Under Mass Higher Education:Graduate Employment in Service Industries in Britain’, Journal ofEducation and Work. 15 (4): 427-456.

McIntosh, S. (2002) Further Analysis of the Returns to Academic andVocational Qualifications. DfES Research Report No 370. London:HMSO.

Observer (2003) ‘Degrees of Difference in an education battle’. The Observer,September 21, 2003.

Peck, J. and Theodore, N. (2000) ‘Beyond ‘employability’’. Cambridge Journalof Economics, Vol. 24, pp. 729-749.

Pollard, E., Pearson, R. and Willison, R. (2004) Next Choices: CareerChoices Beyond University. Brighton. Institute for Employment Studies(IES). 405.

Pryor, F.L. and Scaffer, D.L. (1999) Who’s Not Working and Why:Employment, Cognitive Skills, Wages and the Changing U.S. LaborMarket. Cambridge: Cambridge University Press.

Purcell, K. and Elias, P. (2004) Seven Years On: Graduate Careers in aChanging Labour Market. HECSU: London.

Purcell, K. (2002). Qualifications and Careers: Equal Opportunities andEarnings Among Graduates. Manchester, Equal OpportunitiesCommission (EOC).

58

Purcell, K., Elias, P. and Wilton, N. (2004) Higher Education, Skills andEmployment: careers and jobs in the graduate labour market.Research Paper 3. University of the West England & Warwick:Warwick Institute for Employment Research.(www.warwick.ac.uk/go/glmf)

Reich, R. (1991) The Work of Nations, New York: Simon and Schuster.

Sloane, P. J., Battu, H. and Seaman, P.T. (1999). "Overeducation,Undereducation and the British Labour Force." Applied Economics. 31(11): 1437-1454.

Smetherham, C. (2005). The Employability of First Class Graduates. Schoolof Social Sciences. Cardiff, Cardiff University. Unpublished Ph.D.thesis.

Statistical Bulletin SB 50/2004 First Destinations of Students GainingQualifications from Higher Education Institutions: Wales 2001/02,National Statistics/National Assembly for Wales.

Times (2004) ‘Graduate pay falls as student rolls soar’, The Times, August 20,2004 p. 1.

Walker, I. and Zhu. Y. (2003). "Education, Earnings and Productivity." LabourMarket Trends (March 2003): 145-152.

Welsh Assembly Government (2002a) Reaching Higher: Higher Educationand the Learning Country: A strategy for Higher Education in Wales,Cardiff, WAG.

Wolf, A. (2002) Does Education Matter? London: Penguin.

1 See, for example, Keep, E. (2004) ‘After Access: Researching Labour Market Issues’, in J. Gallacher(ed.) Researching Access to Higher Education, London: Routledge.

2 See ‘2004 Finalists’ Career Plans and Expectations (summer 2004),http://www.prospects.ac.uk/cms/ShowPage/Home_page/Labour_market_information/

3 The UNITE Student Experience Report 2005, http://www.mori.com/polls/2004/unite-jan.shtml

4 The UNITE Student Experience Report 2005, http://www.mori.com/polls/2004/unite-jan.shtml

5 See ‘2004 Finalists’ Career Plans and Expectations (summer 2004),http://www.prospects.ac.uk/cms/ShowPage/Home_page/Labour_market_information/

6 NatWest – Press Release, “New Students Expect to Pay £26,000 For University Education”, 10 August,2004, http://www.natwest.com/pressroom/index.asp?navid=PRESS_ROOM&pid=64

7 See http://www.agr.org.uk.

8 The NatWest 2004 Student Money Matters Survey, Natwest August 20049 See BBC (2003) ‘Student debt continues to climb’ BBC News On-line http://news.bbc.co.uk/go/em/fr/-

/1/hi/business/3683421/stm

59

10 The survey also shows that students at university in the South West finish their degrees with the highest

debts, owing £14,802 on average. Students in the Midlands in contrast had the lowest debts, owing£11,484 on average. More than 60% of graduates had bank debts of £4,650 and 53% owed an averageof £2,195 on credit cards, while 86% had student loans averaging £10,161. About 30% of students saidthat they also owed money to their parents, with debts averaging £3,192.

11 Student Living Report 2004, MORI/UNITE 2004.

12 Callender, C. and Wilkinson, D. (2003) 2002/03 Student Income and Expenditure Survey: Students’Income, Expenditure and Debt in 2002/03 and changes since 1998/99, The Policy Studies Institute,DfES, Research Report No. 487.

13 Department of Education and Skills (2003) The Future of Higher Education, London: HMSO, p.11.

14 http://www.publications.parliament.uk/pa/cm200304/cmbills/035/2004035.htm

15 Blöndal, S., Field, S. and Girouard (2002), ‘Investment in human capital through post-compulsoryeducation and training: Selected efficiency and equity aspects.’ OECD Economic Department WorkingPapers 333.

http://www.olis.oecd.org/olis/2002doc.nsf/linkto/eco-wkp(2002)1916 See Levin, H.M and Kelly, C. 1997 ‘Can education do it alone?’ in A.H.Halsey, et. al. (Eds.) Education:

Culture, Economy and Society, Oxford: Oxford University Press.

17 See Brown, P. and Lauder, H. (forthcoming) Globalisation, Knowledge and the Myth of the MagnetEconomy, in Lauder, H. et al. (Eds.) Globalisation, Education and Social Change, Oxford: OxfordUniversity Press; Brown, P. Lauder, H., Ashton, D. and Tholen, G. (forthcoming) The Rise of the HighSkill, Low Wage Economy?

18 Brown, P. and Hesketh, A. (2004) The Mismanagement of Talent: Employability and Jobs in theKnowledge Economy, Oxford; Oxford University Press.

19 All lifetime earnings are calculated on a 40 hour week, 52 week year, over 40 years.

20 Green, F, McIntosh, S. and Vignoles, A (1999) ‘Over-Education and Skills - Clarifying the Concepts’,paper given at the EEEG group annual conference, University of Wales, Swansea, July, 1999(http://www.essex.ac.uk/ilr/eeeg/vignoles.pdf)

21 The within-group inequality in earnings for college graduates is not only far greater than that of the lowerskilled, it has also been growing more rapidly, whereas non-graduates experienced a slow or no growthin within-group variance in earnings. See: Lemieux, T. (2003) ‘Residual Wage Inequality: A re-examination’, University of British Columbia, CLEER working paper nr. 2,http://www.econ.ubc.ca/cleer/papers/cleer002.pdf

22 Purcell, K. (2002). Qualifications and Careers: Equal Opportunities and Earnings Among Graduates.Manchester, Equal Opportunities Commission (EOC).

23 Purcell, K. (2002). Qualifications and Careers: Equal Opportunities and Earnings Among Graduates.Manchester, Equal Opportunities Commission (EOC).

24 Purcell, K. (2002). Qualifications and Careers: Equal Opportunities and Earnings Among Graduates.Manchester, Equal Opportunities Commission (EOC).

25 DfES (2003) The Future of Higher Education, White Paper, para. 1.28.

26 Pollard, E., R. Pearson and R. Willison (2004) Next Choices: Career Choices Beyond University.Brighton. Institute for Employment Studies (IES). 405.

27 Connor, H., C. Tyers, T. Modood and J. Hillage (2004) Why the Difference? A Closer Look at HigherEducation Minority Ethnic Students and Graduates. Department for Education and Skills.

28 The Times Higher Education Supplement, June 18, 2004.

29 The Future of Higher Education, DfES, 2003: para 1.22: 16.

30 Wilson, R.A. (2001) Projections of Occupations and Qualifications 2000/2001 (London: DfEE).

60

31 Institute for Economic Research (IER)/Department for Education & Employment (DfEE) (2001)

Projections of Occupations and Qualifications 2000/2001, DfEE Research Brief No: RBX08-01, April2001 (London: DfEE).

32 For a discussion of these issues, also see Brown, P. and Hesketh, A. (2004) The Mismanagement ofTalent: Employability and Jobs in the Knowledge Economy, Oxford University Press.

33 Elias, P. and K. Purcell (2004) SOC (HE:) A Classification of Occupations for Studying the GraduateLabour Market. Research Paper No. 6. Researching Graduate Careers Seven Years On. ESRU/IER.<www.warwick.ac.uk/go/glmf>.

34 Elias, P. and K. Purcell (2004) SOC (HE): A Classification of Occupations for Studying the GraduateLabour Market. Research Paper No. 6. Researching Graduate Careers Seven Years On. ESRU/IER.<www.warwick.ac.uk/go/glmf>.

35 Post-secondary Awards

� First professional degree. Completion of the degree usually requires at least three years of full-time academic study beyond a bachelor’s degree.

� Doctoral degree. Completion of a Ph.D or other doctoral degree usually requires at least threeyears of full-time study beyond a bachelor’s degree.

� Master’s degree. Completion of the degree usually requires one or two years of full-timeacademic study beyond a bachelor’s degree.

� Bachelor’s or higher degree, plus work experience. Most occupations in this category aremanagement occupations. All require experience in a related non-management position forwhich a bachelor’s or higher degree is usually required

� Bachelor’s degree. Completion of the degree generally requires at least four years, but notmore than five years, of full-time academic study.

� Associate degree. Completion of the degree usually requires at least two years of full-timeacademic study.

� Postsecondary vocational award. Some programs last only a few weeks, while others may lastmore than a year. Programs lead to a certificate or other award not a degree.

Work-Related Training� Work experience in a related occupation. Many occupations requiring work experience are first-line

supervisor/managers of service, sales and related, production, or other occupation, or aremanagement occupations.

� Long term on-the-job training. More than 12 months of on-the-job training or combined workexperience and formal classroom instruction are needed for workers to develop the skillsnecessary to be fully qualified. This category includes formal and informal apprenticeships thatmay last up to five years. It also includes intensive occupation-specific, employer-sponsoredprograms that workers must successfully complete.

� Moderate term on-the-job training. Skills needed to be fully qualified can be acquired during 1 to 12months of combined on-the-job experience and informal training.

� Short-term on-the-job training. Skills needed to be fully qualified can be acquired during a shortdemonstration of job duties or during one month or less of on-the-job experience or instruction.

36 See Phillip Brown and Anthony Hesketh (2004) The Mismanagement of Talent: Employability and Jobsin the Knowledge Economy, Oxford: Oxford University Press.

37 See Brown and Hesketh, 2004; Briefing No.1.

38 See Fevre, Briefing No.5.

61

39 The Future Skills Wales 2003 Generic Skills Survey, undertaken on behalf of the FSW partnership by

Experian Business Strategies and MORI.

40 These are discussed by Fevre in Briefing No.5.

41 These estimates are based on both the 2001 Skill Survey and data for the Labour Force Survey. SeeBriefing No.6 for details.