Surely there are too many graduates now?

• Author: Karmel, Tom
• Position: Invited Article - Report

Abstract

This article updates earlier work which asked the question whether Australia was producing too many graduates, noting that the increase in the proportion of the workforce with degrees had been quite dramatic. That work dated from the mid 1990s, and the quick answer was that the lack of change in relative earnings implied very extensive changes in the structure of labour demand. However, the expansion of the proportion of the workforce with qualifications has continued at a remarkable rate and therefore it is timely to ask the question again. By examining the relationship between changes in the proportions of the workforce with different levels of education and changes in relative wages, the article estimates shifts in the demand for labour across the various levels of education.

1. Introduction

   One of the most striking changes over my working life has been the increase in educational attainment of the population with the Year 12 retention rising to almost 80 per cent, from below 30 per cent in the 1960s, and university education evolving into a mass-education system from one that was relatively elitist. This apparently inexorable change led me to pose the question in the mid 1990s 'Are there too many graduates?' (Karmel 1997a). I noted that in 1971 there were 178,000 graduates representing 2 per cent of the population 15 years and over, and by 1995 the number had grown to 1,411,000 or almost 12 per cent of the population 15 years or over. The most recent data from the Survey of Education and Work (Australian Bureau of Statistics 2012, cat. no. 6227.0) put the number of individuals aged 15 to 64 years with a degree or higher qualification at 3,768,000, with a further 1,480,100 having a diploma, representing 25 per cent and 10 per cent of the population 15 years and over respectively.

   At a policy level, the push to increase education levels continues. The Australian Government has set targets such as '40 per cent of 25-34 year olds will have a degree by 2025' and 'there will be a doubling of diploma completions from 2009 to 2020' (Australian Government 2009; COAG 2012). Similarly, the Australian Workforce and Productivity Agency (2013) argues that the labour market's demand for qualifications is expected, at least in the higher-growth scenarios, to increase by between 3 and 3.9 per cent on average to 2025--well above the expected increase in the workforce. Under these types of scenario the proportion of individuals with a qualification within particular occupations will continue to increase.

   A framework is needed to answer the question that has been posed in the title of this article. The approach (Katz and Murphy 1992) is to look at what is happening to the structure of labour demand to see whether it is shifting in favour of or against graduates (and other educational groups). The intuition is that with a fixed production function, an increase in the quantity of one type of labour will result in a decline in its relative earnings. If this does not occur, then the implication is that the structure of labour demand has shifted toward that type of labour. Based on this approach, I (Karmel 1997a and b) showed that the labour market shifted very significantly between 1968-1969 to 1989-1990 in a way that favoured those with degrees; it was particularly unfriendly to early school leavers and women in the technician, trade, other category. Thus the answer in the early 1990s to the rhetorical question was 'no, there are not too many graduates'.

   This finding for Australia was consistent with the experience in the United States (Blackburn, Bloom and Freeman 1989; Katz and Murphy 1992, Murphy and Welch 1992), although the changes in labour supply were quite different, with continued increases in education levels in Australia but not in the United States. More recent work (Coelli and Wilkins 2008) suggests that the labour markets in both the United States and Australia have remained friendly to graduates. Greenstone and Looney (2012) find that the returns to college education have been largely constant over the previous 35 years. Boudarbat, Lemieux and Riddell (2010) paint a similar picture for Canada, with rising returns to education for men over the period 1980-2005. Booth by and Drewes (2010) similarly examine the returns to education over the period 1980-2005 and find that the answer to whether there are too many graduates in Canada is 'no'. O'Leary and Sloane (2011) find the premium enjoyed by university graduates remained largely stable between 1997 and 2006 while, in respect of Ireland, McGuiness, McGinnity and O'Connell (2008) find stable returns to education between 1994 and 2001. However, not all the literature points to increases in the demand for skilled labour. For example Brown, Fang and Gomes (2012) model returns to education in a life-cycle framework and find that the gains from college education in the United States were flat in the 1980s and decreased significantly in the 1991-2007 period. Beaudry, Green and Sand (2013) point to a decline in the demand for skill since 2000, even as the supply of educated workers grew.

   The data come from the Survey of Education and Training (Australian Bureau of Statistics, various years) which provides detailed educational qualification data and labour force data, including earnings. Data from three surveys are used: 1997, 2001, and 2009. I use a particularly detailed education classification, which cross classifies qualifications by whether the individual has completed Year 12 or not. This is to account for the large differences between certificates in terms of complexity and potential earnings.

   In the next section I lay out the theoretical framework. The results are presented in Section 3. I end with a short discussion.

   The analysis shows that the labour market between 1997 and 2009 has shifted in a way which is favourable for those with postgraduate qualifications, degrees, and diplomas (in that order), and unfavourable to those who fail to complete Year 12. In addition, the labour market has adapted to large increases in the number of people with certificates III and IV without a noticeable impact on wage relativities.

   The brief answer to the question posed in the title is 'not yet'.

2. Katz and Murphy's Model of Labour Demand

   As noted earlier, the intuition behind this approach is that with a fixed production function an increase in the quantity of one type of labour will result in a decline in its relative earnings. If this does not occur, then the implication is that the structure of labour demand (that is, the production function) has shifted toward that type of labour. The method quantifies this shift.

   Assume that the economy can be characterised by a CES production function.

   Y = [([[delta].sub.1][L.sup.-[rho].sub.1] + [[delta].sub.2][L.sup.-[rho].sub.2] + [[delta].sub.3][K.sup.- [rho]]).sup.1/[rho]] (1)

   where [[delta].sub.1] + [[delta].sub.2] + [[delta].sub.3] = 1

   If we assume that the factor inputs are paid their marginal...