Scottish Government

Annex C Economic Value Note

Purpose

1. The purpose of this note is to outline a methodology and set of preliminary results that seek to place a value on the contribution that colleges make to enhance the skills of the Scottish workforce.

Limitations

2. It should be noted from the outset that this note does not attempt to place a total economic value on the college sector, rather our aim is to place a numerical value on the output of colleges where possible and where this is not possible to describe the "unvalued" output. This approach is consistent with HM Treasury's appraisal guidance ³² known as "The Green Book" which states:

"Costs and benefits that have not been valued should also be appraised; they should not be ignored simply because they cannot easily be valued."

3. We have set out in this study to attempt, as far as possible, to value the output of the sector. We have deliberately not focused on the economic impact of spending by the college sector whether on wages, consumables or buildings. This is because we wish to understand more about the difference that colleges make. If resources were not devoted to the college sector then they would be available for spending elsewhere for example in the Health Service or on a tax cut both of which would generate economic impacts through the multiplier. It is therefore of more interest to examine the output, what we are getting for the resources allocated to the sector, rather than the inputs.

4. Inherent in this approach is the assumption that the additional value ascribed to improved qualifications is the result of the learning in the college. It was beyond the scope of this exercise to control for ability. It also assumes that the students work until retirement (although we did conduct a sensitivity test on this, the results of which are given below).

Approach

5. The approach which has been adopted makes use of two main data sources; The Labour Force Survey and data returns from the college sector.

6. From the data returns from colleges we know the level of qualification that students have on entry and exit from college. Where the qualifications on exit are "higher" ³³ than those on entry we can make use of the Labour Force Survey (which contains information on qualification levels as well as wages) to attempt to place a market value on the additional learning.

7. Having placed an annual "value" on these upgrades we can then role this forward over the average working life of the student to obtain a value for the qualification. This value is then discounted using the "Social Time Preference" discount rate of 3.5% (real) given in the Green Book to obtain the "Present Value" of the qualification.

8. The relevant data set (covering 2004/05) contained 451,557 enrolment records. These are turned into a headcount using a data matching programme which produces 335,116 separate student records. This set is further reduced by imposing the condition that the students must have completed the programme, been assessed and been successful; this reduces the number that we have to work with to 139,271. The majority of these students had unknown qualifications on entry, which meant they could not be included in this stage of the analysis. This does not mean that the programme has no value as can be demonstrated by "Tracy's" experience (see, for example, the case study on support for vulnerable young people on page 42). This resulted in a sample of 44,364 students being carried forward.

9. We placed the remaining 44,364 students in a qualifications matrix and removed all who did not leave college with a higher qualification than they entered with. This final step was required as we cannot differentiate the wage levels of individuals unless their qualification level changes (by far the largest group c26,000 individuals were recorded as having no qualification fitting our matrix on exit). It does not mean that there is no value in an additional qualification at the same or a lower level. In a local labour market this might make the difference between getting and not getting employment. We are left at this point with a sample of 13,022 individuals.

10. We were concerned at this point about basing our estimates on only 13,022 students. Data limitations meant that we lacked the required entry qualification data needed to place a value on the remaining students so it was decided to separately model entry qualifications for this group based on probabilities obtained from the initial 13,022 entries where this information was present. This modelled data is presented separately as it is further removed from the original administrative data. This procedure allows us to model an additional 22,476 students (44,139-21,663 = 22,476 ³⁴).

What are the qualifications worth?

11. Having identified 13,022 individuals who have increased their qualification level between entering and leaving college the next question to address is what are the qualifications worth? Table 1 below sets out the median wages that can be expected with each qualification level adjusted for the likelihood of being employed with that qualification level. A higher value will therefore be ascribed to a qualification which increases the likelihood of employment even if the cash value of the qualifications is identical ³⁵.

Table 1:

12. The wage information contained in table 1 above can be combined with the number of students moving between each qualification to give an indication of the additional value that accrues to these students each year as a result of their college education. Table 2 below summarises the number of students in our data set moving between each qualification level.

Table 2:

13. Table 3 below combines the information from table 1 on wages with the information from table 2 on the numbers moving between each qualification level. This gives us a value for each upgrade.

Table 3:

14. The total gross annual benefit is therefore calculated to be of the order of £56m. This is of course not a one off benefit but recurs every year until the students leave the workforce. One of the strengths of the college sector is the heterogeneity of its students and so the average time to retirement is calculated separately using age information from the data set. The average age of students taking each qualification "step" is summarised in table 4 below. An implicit assumption of this calculation is that students work until retirement which is assumed to be 65.

Table 4:

15. Combining the information in table 4 with that in table 3 allows us to estimate the time over which these benefits persist. If we then discount the future stream of benefits using the social time preference rate of 3.5% used for appraising government spending we find that the present value of the gross benefi ts delivered by the college system is in the order of £1.2bn. Table 5 below sets out how the grand total is built up from each segment of the student population.

Table 5:

Costs

16. We cannot of course forget that there is a resource requirement to generate these benefits. There are two main sources of costs that we require to consider:

• Direct costs of the college system; and
• Opportunity costs to students of studying at college rather than working.

Direct Costs

17. There are two possible approaches to evaluating the direct costs of the college system. We can either seek to split out the costs of the 13,022 students identified above or alternatively ascribe all the costs of the college system against this flow of valued benefits. There are advantages and disadvantages with both approaches.

18. Splitting out the costs of these courses has the attraction of being "fair" as we would be equating like with like. The problem of this approach is that it leaves us with a residual set of costs to be ascribed to the "unvalued" college output. The "unvalued" output may be worth more or less than the residual costs. In addition there is the practical problem of sharing fixed costs between the two groups of courses.

19. The alternative is to allocate all of the costs against the valued benefits. When we do this we are no longer equating "like with like". This would deliver a "worst case" net value for the sector as it assumes in effect that the non valued courses are "free" so any benefit from them is a bonus over and above the final Net Present Value ( NPV).

20. Direct spend on Scotland's colleges comes from two primary sources these are the Scottish Funding Council and student support for students studying HE courses in college from Student Awards Agency for Scotland. Taking both these funding streams together gives a total cost of about £575m or less than half the estimated benefit for just the output we have been able to value.

Opportunity Costs

21. In addition to the direct costs of funding the system students also face significant opportunity costs in the form of income forgone while they study. These have been estimated in table 6 below. It has been assumed that students were earning the median wage for their pre-college skill level

Table 6:

22. It can be seen from table 6 above that college represents a serious financial investment in terms of earnings forgone for these students. This is also a cost to the wider economy in the short run as these workers are not available until they complete their studies. We know that students work while at college and the recent Student Income and Expenditure study found that on average they earned £948 per annum this reduces the opportunity cost of attending college.

23. The net opportunity cost (allowing for earnings whilst at college) is estimated at around £140 m and requires to be included in the cost benefit analysis.

Cost Benefit Result

24. Table 7 below sets out the impact of the costs associated with gaining the benefits on a step by step basis to give the range of cost benefit results. Netting off the opportunity costs faced by students results in a benefit of £1.06bn. Netting off the remaining costs of the sector gives a benefit of £483m. This should be interpreted as the lower bound benefit estimate as by implication it assumes that the other parts of the sector which have remained unvalued are worth the same as the opportunity costs of qualifying in effect have a zero net value.

Table 7:

25. The model allows us to vary the retirement assumption. Reducing the retirement age from 65 to 60 reduces the gross benefit from £1.2bn to £1.1bn. Retirement would have to occur on average at 42 in order for the net benefit to drop to zero when all costs are included.

Additional Benefits

26. It is important to note that although the calculations carried out above have yielded large and positive results for the benefits of a college education they have not valued all the activity that occurs in Scotland's colleges. The main report gives more detail on the additional benefits which we have not been able to value in this note. For example we have not valued the benefits to school pupils attending college for the simple reason that we cannot identify the separate impacts of the work of colleges from the work of the schools. (See page 25 for the case study on school and college partnership.)

27. Our requirement that courses lead to an increase in qualification level led to a reduction in our sample from 44,364 to 13,022. This occurred because our valuation method (based on the LFS) required a change in qualification level to identify "improvements" in qualification level (we also required that a qualification be recorded on entry). These additional individuals must gain some benefit from attending college, perhaps local labour market conditions result in a requirement for retraining at the same level but we cannot separate this effect out from the aggregate data.

28. Some individuals may also re-enrol the following year to do a higher qualification (perhaps they wish to improve their Highers before pursuing an HNC/D for example). Where this is the case (and it is their qualification on entry to the college rather than the particular course that is recorded) then their eventual upskilling should be captured in our figures assuming that the flow of such people is similar on a year by year basis.

29. A large amount of the college activity that we identified was excluded from our model because it did not lead to a certified or specifically identified qualification (see, for example, the case study on custom designed training for companies on page 49). Once again there is likely to be a benefit to doing these courses but we cannot vouch for its scale. In the absence of robust data we have taken a conscious decision to be cautious in our assumptions around the benefit calculations. It is also the case that where these courses are a first step on a longer lifelong learning journey which does lead to obtaining a recognised college qualification (see, for example, the case study on closing the opportunity gap on page 38) they will again, at least partially, be implicitly counted in our estimates.

30. Finally there is the activity that we could not value as a result of the lack of information on the qualification on entry. This had to be excluded initially as without knowing the qualification level on entry we could not identify the extent of upskilling. Our concern over the numbers being excluded on this basis led to us attempting to "model" entry qualifications for this group. Our approach to this is set out below.

Modelling

31. As mentioned above we were concerned at the extent to which we had to exclude those who did not have a valid qualification on entry recorded against them. Our data suggested that an additional 22,476 individuals fell into this group. We have sought to address this concern by trying to model the entry qualifications profile for these students.

32. The first step in assessing the value that we can ascribe to these students is to construct a matrix giving the probability that they fall into each of the classes of entrant that we have identified for this exercise. Probabilities are calculated as the percentage of the group for which we have information who fell into each of these categories. This probability table is given as table 8 overleaf.

Table 8:

33. The number of students graduating at each level is then run through the matrix. This gives us the additional numbers of individuals given in the table below

Table 9:

34. As before only those who have enhanced their earning potential are included in our benefit calculation. This reduces the modelled number upon whom the benefits calculation can rest to 11,097. The same steps outlined above for the original sample are then carried out to estimate a value for this additional supply of qualifications.

35. Assuming that these individuals retire at 65 (as assumed above) generates an estimated additional gross benefit in the order of £909m. As before opportunity costs (c£117m) are estimated. This suggests an additional net benefit in the region of £792m against prorata costs.

Conclusions

36. This piece of work has shown that there are substantial economic benefits which flow from the learning that is provided at Scotland's Colleges. We estimate that the gross benefit of a year's output is of the order of £1.2bn. If we set all of the sector's costs against this output then we are left with a net benefit of at least £483m excluding our modelled benefits. This should be considered our low estimate of net benefit.

37. Adding the original and modelled results together suggests:

A net benefit figure against all government costs of the order of £1.28bn.

38. It needs to be remembered that we are not able to value impacts of learning which is not completed, assessed and successful. Where courses are not formally assessed they may still add value although we cannot pick that up using this methodology. The main report considers other ways in which colleges add value. We have also not been able to place an explicit value on a large number of qualifications for example National Units not leading to any qualification above that level (29,575), other non-advanced certificate (20,545), any other qualification level (22,892) and programmes not leading to recognised qualification level (21,238).

39. Ascribing all costs to the output we have valued (an extremely cautious assumption) leaves us with a net benefit figure of around £1.28bn. This estimate implicitly assumes that where a course is unvalued its value is only equal to the opportunity cost a student faces in doing it. Most people would agree that this is a very restrictive and cautious assumption. Even with this assumption in place the college sector is shown to generate a benefit cost ratio against government spend of around 3.2 to 1. In effect we are saying that the college sector turns £1 into an asset worth (at least) £3.20 in a year - this represents an excellent return on investment.

40. If we accept the estimate for net benefits above then everything else the system delivers which has not been valued is in effect "free" as the costs have already been covered.