Projected changes to the rate of return

3.23 The projected faster growth in the supply of graduates through the early years of the next century to 2001, as compared with employer demand, is likely to reduce rates of return below the level of seven to nine per cent estimated on the basis of pay premia from the early to mid-1990s.

3.24 How much the graduate pay premium is reduced will depend on three factors:

• the balance between supply and demand;
• the elasticity of substitution; the flexibility with which employers can profitably substitute graduate for non-graduate labour;
• whether the wider pool of students now entering higher education will be able to achieve the same value added (in given labour market conditions) as their predecessors.

3.25 Taking the last point first, it is sometimes claimed that extra graduates will be of lower ability and prior academic attainment than previous cohorts of graduates and so will not get the same pay premium as existing graduates.

3.26 Chart 3.3 shows recent trends in the entry for students to the new universities over the period of recent expansion). Chart 3.4 reports the grades achieved by GCE A level entrants. There has been some increase in the proportion of entrants without GCE A levels, although amongst entrants with GCE A levels, A level grades have not fallen.

3.27 It might be claimed, however, that, despite the stability of entry qualifications, the recent expansion of higher education has let in students of lower ability and that graduates with less ‘innate’ ability gain less from a degree. The empirical evidence does not appear to support this. The most recent and detailed study of the link between qualifications and earnings for the UK – Blundell et al (1997) – finds no variation between measures of ‘innate’ ability or parental background and the returns to education. Ashenfelter and Rouse’s (1996) review of the US literature finds no significant link for the US either.

3.28 We do not, therefore, believe that there is support for the view that the recent expansion of higher education will have brought in students achieving lower rates of return as a consequence of lower underlying ability. We do, however, expect the change in labour market conditions – and the faster growth in graduate supply – to have an impact.

3.29 The elasticity of substitution is relevant because where there is widespread scope for profitably utilising additional graduates then the impact of increased supply on the graduate pay premium will be limited. On the other hand, where there is limited flexibility then graduates will essentially be utilised to carry out tasks which would otherwise to be done by non-graduates, and the graduate pay premium will be eroded. Recent research on the utilisation of new graduates shows examples of both these patterns in different sectors of the economy (Mason; 1995).

3.30 More systematic research evidence on the flexibility of the graduate pay premium (the wage elasticity) is limited. The evidence summarised in Freeman’s survey article,¹⁶ supported by a recent article by Katz and Murphy,¹⁷ suggests an elasticity of substitution between college graduates and high school graduates in the US of around 1.5. If this is applicable to the UK it suggests a 10 per cent excess supply would reduce graduates’ earnings by around 7 per cent. Schmitt’s results¹⁸ suggest a slightly smaller effect for the UK.

3.31 It is clear, however, that this approach is quite simplified; it effectively considers a single market for graduate labour whereas in reality, the market will be segmented in a number of ways, particularly by age. A more detailed approach would seek to forecast the impact of the recent expansion upon the pay premium earned by the relevant cohort of graduates over the entire life cycle. It might be expected that the pay premium would be more significantly reduced early in the graduate’s career; while by the time people in the cohort reach their forties and fifties technical change might be expected to have significantly expanded the demand for graduate skills.

3.32 The greater weight given by rate of return estimates to earnings early on in graduates’ careers suggests that mechanically using the estimates of elasticities above might underestimate the impact on rates of return of an increase in graduate supply. Trying to forecast the profile of graduate pay would, however, be enormously uncertain. Accordingly, we have just looked at a wide range of different scenarios illustrated in the table below. The Warwick central demand forecast shows the demand to employ graduates increasing by broadly 10 per cent less then supply over the period to 2001: the projections in Chart 3.2 suggest that a greater imbalance will arise over the following decade. We have looked at the consequences for rates of return if this results in reductions in the graduate pay premium between 10 per cent (highly flexible labour market demand) and 50 per cent (inflexible labour market demand).

3.33 Whilst the calculations have wide margins of uncertainty they suggest two conclusions:

• first, unless the graduate earnings premium is reduced by more than a third the projected rates of return for today’s graduate is likely to remain at or above 6 per cent – the Treasury’s discount rate for investment;
• secondly, unless the pay premium only responds slightly to the increase in graduate supply, the rate of return is unlikely to remain substantially above 6 per cent.

3.34 There are a number of other factors which will affect the projected rate of return:

• teaching costs: further reductions in teaching costs would increase the rate of return. A reduction in teaching costs of 10 per cent will increase rates of return by
  almost 0.2 per cent;
• external economic benefits; to the extent that there are considered to be external economic benefits not captured in graduates’ salaries then this will increase rates of return. The evidence on external benefits reviewed by Professor Gemmell is not very robust and does not suggest these to be large;
• economic activity rates: changes to the relative labour market participation of graduates will affect the rates of return.