Thursday, 25 June 2020

What you study vs where you study: how FE choices affect earnings and academic achievement


By Esteban Aucejo, Claudia Hupkau and Jenifer Ruiz-Valenzuela


Following the unprecedented number of job losses and the bleak economic outlook due to the Covid-19 crisis, more people will be considering staying on or returning to education. Vocational education and training (VET) is likely to play a crucial role in providing the skills needed for economic recovery, including retraining workers who have been made redundant. In this context, it is crucial to have good information on the returns to different fields of study that can be taken at FE colleges, and whether it matters which institution one attends for earnings and employment prospects. Our new research published by the Centre for Vocational Education Research (CVER) finds that when it comes to vocational education and training (VET), what you study is very important for future earnings.  Whereas where you study can also matter for younger people but less so for adults.

We used data from more than one million students over 13 years to investigate how much value attending an FE college adds in terms of academic achievement, earnings and employment, taking into account learners’ prior achievements and their socio-economic background. Our study considers both young learners, who mostly join FE colleges shortly after compulsory education, as well as adult learners, who have often worked for many years before attending FE college.

Our value-added measure indicates that moving a student from a college ranked in the bottom 15 percent of the college value-added distribution to one ranked in the top 15% implies a fairly modest 3% higher earnings on average, measured at around seven years after leaving FE college. The difference in earnings for adult learners is smaller, at 1.5%. The fact that college quality seems to matter more for young learners is likely due to young learners spending more time in FE colleges (i.e. they enrol in and complete substantially more learning than adults). The results in terms of the likelihood of being employed show even smaller differences across FE colleges.

There is considerably more variation in FE colleges' contributions to the educational attainment of their young learners. On average, the young people in our sample enrol in just under 600 learning hours, but only achieve about 413 (or 69% of them), around 42% achieve a Level 3 qualification, and 38% progress to higher education.

But were we to move a learner from a college ranked in the bottom 15% by value-added to one ranked in the top 15%, they would, on average, achieve 6.5% more learning hours (from 69% to 73.4%). They would be almost 11% more likely to achieve a Level 3 qualification (from 42% to 46.5%) and the likelihood of attending higher education would increase by 10% (from 38% to 42%). These are large effects. As young people are likely to attend their nearest college, the variability in value-added between institutions is a source of unequal opportunity between geographic areas.

What differentiates high value-added colleges from low value-added ones? Learning characteristics seem to play an important role. Colleges that offer a larger share of their courses in the classroom (as opposed to in the workplace or at a distance) have higher value-added in earnings for young learners. This is particularly relevant in light of the current crisis, where online and distance learning is expected to remain a regular feature, at least in the medium term. We also find significant correlations between the curriculum offer and value-added measures, with colleges offering more exam-assessed qualifications (as opposed to competency-based) showing higher value-added.  

While where you study does not imply large differences in earnings after college, what you study has a much bigger effect, especially for female and young learners. We carried out a separate analysis looking at students’ earnings before and after attending FE college. In this analysis the young people were aged 18-20 and so had been working for up to two years’ prior to study. Table 1 below shows the 3 most popular fields (in terms of learners doing most of their guided learning hours in that particular field, i.e. specialising in them) by gender and age group.

The two fields of engineering and manufacturing technology, and business administration and law show large levels of enrolment among males and lead to large positive returns. For instance, the typical young male learner who chooses engineering and manufacturing technology as his main field of study will earn, on average, almost 7% more five years after finishing education when compared to earnings before attending FE college, after adjusting for inflation.  For adult male learners specialising in this field, earnings rise by 1.5% five years after leaving college. In contrast, young male learners specialising in retail and commercial enterprise do not see any increase in earnings five years after attending FE college. These results take into account that earnings increase with experience, irrespectively of which field one specialised in. Business administration and law, and health, public services and care are the two fields that show high levels of enrolment and consistent positive returns for women across age groups.

While we find consistently higher returns to fields of study for women than for men, this does not mean that overall, they have higher earnings post FE-college attendance. It means that compared to before enrolment, they experience steeper increases in earnings after completing their education at FE colleges. We also find that many specialisations present negative returns immediately after leaving college that turn positive five years after graduation, indicating that it takes time for positive returns to be reflected in wages. The fact that timing matters suggests that policy makers should be extremely cautious about evaluating colleges in terms of the labour market performance of their students.

Our findings also have relevant practical implications for students since they could help them to get a better understanding of the variation in FE college quality and to compare the returns to different fields of study. This information is likely to be particularly important considering the evidence suggesting that students tend to be misinformed about the labour market returns of VET qualifications.

Table 1. Top 3 Fields of study by proportion of learners who specialise in them



Mean GLH
Estimated Return
Share specialising

main field
1 year post-FE
5 years post-FE
in the field
Young male learners




Engineering and Manufacturing Technology
632
0.04
0.068
20.60%
Construction, Planning & Built Environment
621
-0.001
0.023
16.60%
Arts, Media and Publishing
942
-0.064
-0.003
10.70%
Adult male learners




Health, Public Services and Care
77
-0.006
0.004
19.00%
Engineering and Manufacturing Technology
206
-0.008
0.015
18.90%
Business Administration and Law
131
0.003
0.009
14.20%
Young female learners




Health, Public Services and Care
525
-0.002
0.045
25.20%
Retail and Commercial Enterprise
597
0.036
0.115
23.40%
Business Administration and Law
430
0.040
0.118
13.60%
Adult female learners




Health, Public Services and Care
142
-0.008
0.020
34.30%
Business Administration and Law
189
0.004
0.019
14.80%
Education and Training
143
-0.007
0.027
12.70%
Note. The estimated returns reported are the marginal effects, one and five years after leaving the college, respectively, of choosing the field as the main field. This is a summary table. The complete tables can be found in Tables 9 to 12 of CVER DP 030

This blog post appeared first on TES and is republished here with permission.

Friday, 13 March 2020

Training grants: a useful policy to address skills and productivity gaps?

As work changes, firm-provided training may become more relevant for good economic and social outcomes. However, so far there is little or no causal evidence about the effects of training on firms. Pedro Martins looks at the effects of a training grants programme in Portuguese firms.


As most academics, I am fortunate to be able to update my own skills on a regular basis. For instance, when I attend a research seminar, I learn from colleagues that are pushing the knowledge frontier in their specific fields. Some of their insights will sooner or later also feature in my own teaching and research, thus increasing my performance and that of my institution.

However, workers from other sectors typically have far fewer opportunities to increase their skills on a regular basis. Recent research by the European Investment Bank indicates that, on average, workers in Europe spend less than 0.5% of their working time on training activities. In the current context of major changes in labour markets – including artificial intelligence and automation and perhaps even coronavirus – this training figure seems too low.

Economics has long predicted some degree of under-provision of training in labour markets. First, training is expensive for firms, as it typically entails significant direct and indirect costs. Second, employers know they will lose their investments in training if employees subsequently leave. It will be even worse if workers are poached by competitors. Moreover, even leaving aside the issues above, firms may struggle to estimate the effects of training on their performance (sales, profits, etc), which will again detract them from upskilling their workers.

The context above points to an important market failure in training. This context may also explain in part the disappointing economic performance of many European countries over the last years. While labour markets have become more efficient, incentives for on-the-job training may paradoxically have declined, as workers move more easily to other firms. However, public policy may play a role in alleviating the under-provision of training. Specifically, governments can subsidise training in the workplace in order to bring its private net benefit more in line to its social value.

The new working paper featured in this blog and recently presented at a CVER seminar (‘Employee training and firm performance: Quasi-experimental evidence from the European Social Fund’) contributes empirical evidence to this question. The research evaluates the effects of a €200-million training grants scheme supported by the European Union on different dimensions of recipient firms.

The study draws on the difference-in-differences counterfactual evaluation methodology, comparing the outcomes of about 3,500 firms that applied and received a training grant (of about €30,000) and around 6,000 firms that also applied but had their application rejected. Using rich micro data from Portugal, the country where the scheme was introduced, firms can be compared over several years both before and after their participation in the training grants scheme.

The results indicate that the scheme had significant positive effects on training take up, both in terms of training hours and expenditure. For instance, training increased by about 50 hours per worker per year in the firms that received the grant, compared to firms that had their applications rejected. Deadweight – funding training that would be carried out even without the funding – appears to be very limited, in contrast to the findings of an earlier study of a programme in the UK (link).

Moreover, the additional training conducted by firms led to a number of important outcomes that the study can trace, including increased sales, value added, employment, productivity, and exports. These effects tend to be of at least 5% and, in some cases, 10% or more.

For instance, the figure below presents the average difference in total sales between firms that received the grant and those that did not. (Periods -9 to -1 refer to the years before the grant was awarded; while period 0, the comparison year, is when the firm applied for the grant; period 1 is when the firm conducted the training; and periods 2 to 10 refer to the years after the training was conducted.)  The results indicate that total sales are 5% higher in the firm in year 2 and 10% higher in year 5. However, there were no differences between firms before the grant was awarded, which is reassuring as to the counterfactual nature of the study.

The employment results are also interesting as they come from both fewer separations and increased hirings. Firms that increase their training activities in the context of the grant appear to want to expand their workforce but also to retain the workers they already employ. Moreover, the employment effects are stronger when the scheme ran in periods of recession, suggesting that training grants can also act as an active/passive labour market policy, with a positive ‘lock-in’ impact.

In conclusion, there is a case to be made for workplaces to become a little more similar to universities. On-the-job learning can make firms (much) more productive - but that may require a bigger role from governments. Training grants may be a promising tool in this regard.


Wednesday, 11 December 2019

Can the manifesto pledges plug the skills gap?

The parties are offering plenty of promises on improving technical and vocational skills, but, says Sandra McNally (CVER Director), there are significant gaps in their thinking.


Improving technical and vocational skills is a key aspect of improving productivity and social mobility in Britain. The relatively high number of people with poor basic skills and low number of people with high-level vocational skills are long-standing national challenges and have been highlighted in reports by the OECD, the government and academics. In light of this, key priorities of the incoming government should be:

  • To raise attainment and improve educational trajectories for “the forgotten third” who do not get good GCSEs year-on-year and many of whom never achieve a good upper secondary education.
  • To address the shortage of higher-level technical education (at Levels 4 and 5) that was highlighted by the Augar review. 
  • To increase the ability for adults to upskill or reskill later in life. 

But none of the manifestos acknowledge any problem with a “forgotten third” of young people. The source of the problem is partly structural, partly a question of resources, with spending per student between 2010–11 and 2018–19 falling by 12 per cent in real terms in 16–18 colleges, and by 23 per cent in school sixth forms. The Conservative manifesto makes no promise to increase baseline funding beyond existing commitments. Both the Labour Party and Liberal Democrats make large spending commitments to FE in general with the Labour Party making specific mention of aligning the base rate of per pupil funding in post-16 provision with Key Stage 4. The Conservative manifesto does make a significant commitment to increase capital expenditure in Further Education Colleges. While this addresses one of the issues addressed in the Augar review, investment in buildings will not improve student outcomes if there isn’t also investment in their teachers (who are paid considerably less than teachers in schools).

The manifestos do not acknowledge that there is a particular problem with the lack of high-level vocational education in England vis-à-vis higher education. In England, only 4 per cent of 25 year-olds hold a Level 4 or 5 qualification as their highest level, compared to nearly 30 per cent for both Level 3 and Level 6. In contrast, in Germany, Level 4 and 5 makes up 20 per cent of all higher education enrolments.

The main Conservative pledge relevant to this is the establishment of 20 Institutes of Technology with a focus on STEM skills. The Liberal Democrats also promise some institutional reform with the establishment of national centres of expertise for key sectors, such as renewable energy, to develop high-level vocational skills. However, they go further in explicitly acknowledging a “skills gap” and committing to address this by expanding higher vocational training, without however, stating how they would go about this.

Labour promises a free lifelong learning entitlement for everyone, including training up to Level 3 and 6 years of training at Levels 4-6. To the extent that this removes some of the distortions in the financing of the post-18 education system (as well-documented in the Augar review), this would help to address the problem of the lack of higher-level vocational education. But it would be an expensive way of doing so, with the taxpayer (most of whom does not have Level 4-6 education) having to pay the full cost. Moreover, people who are educated up to Level 4-6 have a high private return from this investment compared to people with a lower level of education.

The party manifestos all have something to say about apprenticeships. They all acknowledge problems with how the apprenticeship levy is working. The Conservative manifesto states that they would look into the working of the Levy and see how it can be improved. Both Labour and the Liberal Democrats are far more explicit. They both commit to expand the use of the levy to other forms of training. While this seems like a sensible idea, only two per cent of employers actually pay the levy. All political parties could do with a few more ideas on how to incentivise the other 98 per cent of employers to invest in the training of their staff. The apprenticeship levy is not sufficient for this. The Conservative manifesto has ideas about how to expand R&D credits and it is a pity this does not extend to human capital.

With regard to lifelong learning, Labour and the Liberal Democrats make commitments that are universal whereas the Conservatives’ commitment is more targeted to specific groups through a National Skills Fund (which does not have much detail). Labour’s commitment is to a free lifelong learning entitlement (discussed above) whereas the key Liberal Democrat commitment is the introduction of “Skills Wallets” worth £10,000 for every adult to spend on approved education and skills courses, with the first £4,000 at age 25, £3,000 at age 40 and £3,000 at age 55. This idea has similarities to the “individual learning accounts” that were introduced in 2000 but abandoned a year later because of fraud.

Although the idea of investment throughout life is sensible (and does need to be facilitated), it would be important to ensure that similar mistakes are not repeated. But a more substantive issue is where employer investment appears in this framework. As a major beneficiary of adult training, there needs to be a mechanism for co-investment. This may also help to ensure that the training undertaken meets the needs of the labour market.

Any incoming government needs to be held to account on the extent to which their promises actually address national priorities and whether we see an improvement. The extent to which this is possible depends on the success of their overall economic strategy as well as to the success of specific measures relating to education and skills.

This article was originally published by King's College London's Policy Institute:  https://www.kcl.ac.uk/news/can-the-manifesto-pledges-plug-the-skills-gap

Friday, 27 September 2019

Is there a socio-economic gap in students’ academic match?


Young people from less well-off backgrounds are more likely to pursue lower ranked upper-secondary qualifications than their prior attainment would suggest that they can achieve.

Recent research from Konstantina Maragkou (University of Sheffield and CVER) examines whether socio-economic inequalities exist in the academic match of students in upper-secondary education.  Academic match occurs when student quality matches the quality of the qualification that they take, i.e. students are taking appropriate qualifications for their capabilities and prior attainment. The study uncovers a significant socio-economic gap in academic match among English students in upper-secondary post-compulsory education. Students from lower socio-economic backgrounds achieve less highly ranked qualifications compared to their similarly attaining but more advantaged peers. We show that this is associated with a wage penalty in the labour market.

This study makes use of detailed individual-level linked administrative data (‘Longitudinal Educational Outcomes’ data) from schools, colleges and tax authorities in England for a single cohort of students who undertook their GCSE exams in 2006. Academic match is measured using a continuous variable defining undermatched, matched and overmatched students based on the distance between each student’s prior attainment in GCSE exams and the median prior GCSE attainment of the other students who achieved the same chosen academic or vocational qualification in upper-secondary education. The students who follow upper-secondary qualifications that are studied by similarly achieving peers are then considered as matched to their qualification.  The outcome of this analysis shows that students from socially disadvantaged backgrounds are more likely to be exposed to academic undermatch (that is, having higher prior attainment than the median on their chosen upper-secondary course), even compared to others within the same school.

The study considers students’ match in Level 1 to Level 3 vocational qualifications and Level 3 academic qualifications (A-levels and AS levels) taken in upper-secondary education between ages 16 and 19. Previous literature on rates of return has shown a positive average income return to qualifications at Level 3, whether vocational or academic. However, with regard to the students who leave the education system with lower level qualifications, there is more controversy over the extent to which these qualifications offer good opportunities for future employment and earnings. Qualification choice is therefore important with potentially long-lasting consequences.

Figure 1 displays the variation of the total GCSE point score of students achieving each upper-secondary qualification with the upper and lower lines of the box representing the value at which 75% and 25% of the sample scored below that GCSE point score respectively, the middle line representing the median GCSE point score and the top and bottom extending lines the range. It is evident that there are substantial differences in the ranking of each qualification, with the median scores of students studying for the most highly-ranked qualifications being considerably higher than those of the students studying for the lowest ranked ones. In addition, the difference between the median GCSE point score of students studying for vocational and academic qualifications is also notable.

Figure 1: Measure of qualification’s quality based on median standardised GCSE scores of students achieving that qualification


From the raw data, a significant socio-economic gap in the academic match of students in upper-secondary education is observed, as illustrated in Figure 2. Students from the lowest socio-economic group are less well matched to their chosen upper-secondary qualification compared to students from the highest socio-economic group (i.e. they achieve less in upper secondary education than their GCSE scores would suggest they are capable of).

Figure 2: Academic match of highest and lowest SES students


When keeping important background factors constant, including prior attainment at ages 11 and 14, demographic characteristics and secondary school attended, students from socially disadvantaged backgrounds are still more likely to be exposed to academic undermatch compared to their more advantaged peers. This gap is greater among the highest achieving students. This means that, compared to other young people from a more advantaged background with the same level of prior attainment and the same other background characteristics, those from a less well-off background are more likely to study for lower-ranked qualifications in upper-secondary education. Also, undermatched students are more likely to be found in schools with lower proportions of high achieving students and higher proportions of disadvantaged students, suggesting the importance of peer groups and school guidance on qualification choices. Among the highest achieving students, 80% of the identified socio-economic gap on academic match could be explained by such differences in the schools that those students from differing backgrounds had attended.

In addition, the study shows that a significant proportion of undermatched students are likely to be found in rural districts with higher rates of youth unemployment and higher proportions of residents with only low-level qualifications.

Does being undermatched matter?  The study also estimates wage equations with the indicators of whether young people are matched or not to their upper-secondary qualification. The results show that there is a positive relationship between being academically matched and labour market income returns. Non-university participating girls who are one standard deviation less undermatched earn 17% more at age 25 while non-university participating boys earn 5% more.

In summary, the study shows that there is a significant socio-economic gap in academic match among English students in upper-secondary post-compulsory education. Students from lower socio-economic backgrounds achieved less highly ranked qualifications compared to their similarly attaining but more advantaged peers. We show that this matters for labour market outcomes. Policy-makers should be focusing more on providing students with information related to the available upper-secondary courses that are suitable to each student’s ability credentials and future educational and occupational aspirations.

Tuesday, 25 June 2019

Changing Aspirations and Outcomes in Post-16 Education


In this latest blog post, Steven McIntosh of University of Sheffield discusses CVER contributions to the recent Augar Review of Post-18 Education, and the findings that came out of that research.

CVER have been contributing new research to the recently published Augar Review of post-18 education and funding (available here).  I supplied evidence to the Augar Review commissioners on the factors that influence aspirations and outcomes of young people in post-compulsory education. More details can be found in an accompanying CVER briefing note and full details of the research are published in a DfE research report.

The work considered two cohorts of young people who took their GCSEs almost a decade apart, in 2006 and 2015 respectively, using data from the Longitudinal Study of Young People in England. The aim was to see what influences young people’s aspirations and choices for their post-GCSE education, and whether such relationships have changed over time between the two cohorts. We might expect some such changes to be observed, given the policy initiatives enacted during this period, for example the tripling of university tuition fees to £9000 in 2012, and the promotion of apprenticeships with a target of 3 million new starts by the end of the decade.

Despite these policy changes, the results of our analysis (see Table 1 below) showed that following an academic path through A levels to university remained the most popular choice of young people, with around two-thirds in each cohort aspiring, at age 14, to follow this route post-GCSE. There was actually a small increase between cohorts in the proportion wanting to follow an academic route. There was also a small increase in vocational aspirations between cohorts. When aspirations were re-assessed in Year 11, just before taking GCSEs, the same patterns were observed, though the proportions aspiring to an academic route were lower than at age 14, perhaps as realism set in.

Table 1: Percentage Planning Type of Post-Compulsory Education, by Cohort and Sweep


Sweep 1 (Year 9, Age 14)
Sweep 3 (Year 11, Age 16)

Cohort 1
Cohort 2
Cohort 1
Cohort 2
Academic
62%
68%
55%
63%
Vocational
23%
25%
30%
29%
Neither
15%
  7%
15%
  7%


What people aspire to is often what they end up doing. This was the case for around three-quarters of young people in both cohorts, irrespective of route aspired to. Those who did not fulfil their aspirations, such as those who wanted to do academic A levels but in the end chose the vocational route, were more likely to have lower achievement at GCSE. But even holding prior attainment constant, individuals from a more advantaged family background were more likely to see their aspirations fulfilled.  This is shown in Table 2 below. It is important young people from all social backgrounds should be given equal opportunity to reach their aspirations.  Advice and guidance could be important here in guiding young people towards the best options for them.


Table 2: Percentage who aspired to academic route in Year 11, who follow academic route in Year 12, by family background and prior attainment

Young person’s prior attainment
Highest Parental Education Level

No quals
Level 1/2
A Levels
Level 4
Degree
Cohort 1





7+ A*-C
85%
88%
91%
94%
95%
5-6 A*-C
56%
48%
48%
77%
70%
1-4 A*-C
*
17%
*
37%
*






Cohort 2





7+ A*-C
85%
86%
89%     
89%
92%
5-6 A*-C
41%
48%
48%
49%
53%
1-4 A*-C
*
15%
13%
14%
34%


When we looked at the factors that influence such aspirations and outcomes, then in addition to family background and prior attainment, gender and ethnicity were important. Girls were more likely than boys to aspire to an academic rather than a vocational route, and the gender gap widened between the cohorts. Similarly, those from most ethnic minority groups were more likely to aspire to academic post-compulsory qualifications, holding other factors such as attainment and background constant, with the gap becoming wider for some groups (Mixed ethnicity and Bangladeshi). With respect to region, young people living in London were more likely to aspire to undertake A levels and to apply to university, with this gap increasing in the former case but narrowing in the latter case, between the two cohorts.

Focussing on vocational study, there were more young people taking Level 3 vocational qualifications amongst the more recent cohort. This was most notable amongst those with lower GCSE attainment (some A*-C GCSEs, but fewer than 5), but was actually observed at all levels of prior attainment. For example, amongst those young people with 5 or 6 GCSEs at Grades A*-C), 32% took a vocational Level 3 qualification in Year 12 in Cohort 2, compared to just 17% in Cohort 1. At the very highest level of GCSE attainment (7+ Grade A*-C GCSEs), vocational participation was lower, though even here we saw an increase between cohorts (from 6.5% to 9.5%).

As well as increased participation in vocational Level 3 qualifications in Year 12, the results also showed that members of Cohort 2 were more likely to progress from vocational Level 2 to vocational Level 3 between Years 12 and 13, compared to the earlier cohort. 47% of Cohort 2 members initially learning at vocational Level 2 progressed to vocational Level 3 in Year 13, compared to 30% in Cohort 1.

Looking at types of vocational qualifications, there has been a clear shift between cohorts towards BTEC qualifications, and away from NVQs, particularly at Level 2. BTEC qualifications were least popular amongst whites and people outside London, particularly in the east of the country from the North-East through Yorkshire to East Anglia. Amongst those to have a chosen the vocational route, there was also some evidence that apprenticeships were becoming more popular, though this was mainly just at Level 2. For those progressing straight to Level 3 after GCSEs, Advanced (Level 3) Apprenticeships were rarely taken, with a small increase in such participation between cohorts. Nevertheless, there is potential for more growth in this area, with more young people in Cohort 2 reporting that they discussed the possibility of doing an apprenticeship at age 15 at school or with family and friends, particularly amongst those who did not go on to apply to university. In Cohort 1, 28% of those who did not go on to apply to university had talked to someone about apprenticeships, compared to just 10% of those who did go on to apply. In Cohort 2, these numbers increased to 34% and 20% respectively.

Finally, another positive for vocational education was that there was only limited evidence of ‘churn’ (cycling between low level learning programmes and periods of low-skilled employment or unemployment) amongst low level vocational learners in the two cohorts. A majority of young people starting a vocational course in Year 12 remained in education throughout the whole of that year.

In summary, the academic route remains the dominant route for 16-18 year olds, who show a preference for following A levels and then university. Nevertheless, there are signs of development of vocational education for this age group, with more interest in apprenticeships, and more learning at vocational Level 3, including increased rates of progression from lower vocational levels. The challenge remains to make such routes of broader appeal and to ensure that coming from a disadvantaged background is not a barrier to realising aspirations.