Upskilling and polarisation in the Australian labour market: a simple analysis.

• Author: Esposto, Alexis
• Position: Contributed Article

Abstract

National and international studies have shown consistent upskilling trends in the labour market. While this claim is true at aggregate levels, when employment growth and total hours worked are disaggregated into permanent and casual full-time and part-time employment for men and women, upskilling trends are inconsistent. The analysis shows that permanent male and female full-time employment exhibited clear signs of upskilling both in terms of employment growth and hours worked but this was not the case in casual full-time work for men and women. Part-time casual and permanent work showed clear signs of polarisation and downskilling for men and women. These polarisation trends suggest that workers who do not possess high-level skills will face increasing levels of difficulty and uncertainty in the labour market, with an adverse impact on both household and individual inequality.

1. Introduction

   Over the last 25 years Australia experienced a rapid rise in part-time and casual employment creation (Abhayaratna et al. 2008; Ross and Whitfield 2009; Abhayaratna and Lattimore 2006; and Norris, Kelly and Giles 2005) which was accompanied by a corresponding relative fall in permanent full-time employment creation (Borland, Gregory and Sheehan 2001). Rafferty and Yu (2010) show that growth in non-standard forms of employment--where the average working week consists of over 35 hours of work that attracts standard paid-leave benefits--is outstripping standard full-time employment growth. Aungles et al. (1993) found that employment growth between 1971 and 1986 resulted in a general upskilling of the workforce. Research conducted by Cully concluded that 'there appears to be an increasing polarisation in the Australian labour market between jobs that are high-skilled (high-paid) and jobs that are unskilled (low-paid)' (1999, p. 103). Research by Wooden (2000) and Keating (2003) found that the demand for labour has favoured more skilled employees and that the labour force, in general, has become more skilled.

   Unfortunately, this body of research failed to establish whether the upskilling of the labour force has been consistent across the newer types of employment, namely, full-time and part-time 'casual' and 'permanent' work. Hence, the purposes of this paper are twofold: first, to analyse whether upskilling of the labour force has been uniform across these newer forms of employment and, second, to establish whether the labour market is experiencing a process of polarisation in these job types. In undertaking this research, a simple analysis of employment data at the aggregate level is performed using the Australian Standard Classification of Occupations (ASCO) second edition definition and its measure of skill (ABS 1997, p.9)for the periods 1971-2006 and 1989-2009. (1)

   The paper is structured as follows. The next section discusses the concept of skill bias in the demand for labour. The following two sections analyse upskilling in job types in terms of employment and hours worked. In the final section some conclusions are reviewed.

2. Skill Bias in the Demand for Labour

   Skill bias suggests that the demand for labour has become more skill-intensive and that there has been a shift in demand towards skilled workers and away from less-skilled workers. Thus, if a process of upskilling is occurring, 'changes in relative employment levels (and relative wages) have favoured more highly skilled workers at the expense of relatively less skilled' (Wooden 2000, p.191).

   In most studies about changes in the skill composition of labour there is a consensus that nearly all OECD economies are experiencing upskilling. One way of establishing this is to look at changes in the structure of employment by occupation. Most of the international evidence points to a shift in labour demand towards higher skilled workers. For example Colecchia and Papaconstantinou (1996) show that in most OECD countries in the 1980s employment grew fastest in high-skill occupations and either slowed or fell in jobs that required lower skill levels. Furthermore, they found that the share of high-skill occupations rose relative to low-skill occupations in all of the countries that they examined. Acemoglu (2002) argues that in the United States skill bias has been widespread in the last 60 years and that acceleration in skill bias took place in the last two to three decades.

   For Australia, the evidence is not as clear as it is elsewhere. For example, in their comparative labour market study of upskilling for the United States and Australia, Dunlop and Sheehan (1998) analysed employment change using detailed occupational data. In the case of Australia, the source data were unpublished four-digit data obtained from the ABS and aggregated to the ASCO first edition (ABS 1986a) major group level. For the United States, unit record data were obtained from the Current Population Survey, and these detailed occupation data were aggregated to the same eight occupation groups as for Australia. These data were then grouped into four skill categories, namely, white-collar high-skill, white-collar low-skill, blue-collar high-skill, and blue-collar low-skill. The skill categories were the same as those used by Colecchia and Papaconstantinou (1996). Dunlop and Sheehan (1998) arrived at the following conclusions. In contrast to the OECD findings reported by Colecchia and Papaconstantinou (1996), for neither country was there clear evidence of upskilling in aggregate over the decade to 1995. Second, for neither Australia nor the United States was the growth rate of employment in the high-skill category, in aggregate, above that for the low-skill category. Third, in spite of evidence of upskilling for women in Australia (particularly for white-collar women), there was no real evidence in either country of pronounced upskilling in the white-collar area in aggregate. For example, for the United States the employment growth rates in the white-collar area were the same for high-skill and low-skill persons; in Australia the differences were marginal. Furthermore, there was persistent deskilling in blue-collar jobs, with low-skill employment growing significantly faster than high-skill employment. In an investigation of upskilling in the Australian labour market Cully (1999) reports that over the six years from 1993 to 1999 all occupations experienced net employment growth, with the exception of advanced clerical and service workers. The most notable growth was in professionals and elementary clerical, sales, and service workers. His research also found differences in growth rates in the pattern of employment for men and women. For all employees he reported increases in the share of employment in skill categories I (highest level) and V (lowest level), and reductions in categories II, III, and IV. Cully (1999) concluded that the change in the composition of employment favoured the most skilled and the least skilled, implying that the demand for labour had polarised.

   These findings contrast with those found in the international literature, and Cully concludes 'Australia constitutes a distinct case in experiencing relative growth in both skilled and unskilled positions' (p. 103). Wooden (2000) argues that neither the findings of Dunlop and Sheehan (1998) nor those of Cully (1999) are correct. To show that these explanations are inaccurate, Wooden (2000) uses both aggregate employment and aggregate hours as measures of labour demand. Hours worked is a broader measure of labour demand 'and abstracts from changes in the full-time and part-time composition of employment' (Rozenbes and Mowbray 2009, p.29). Wooden justifies his methodology by pointing out that the records of employment are often not the same as labour demand for two reasons. First, labour demand is often not satisfied because for some types of jobs the available supply of skills and qualifications does not match the corresponding demand. Second, the total number of hours worked may differ substantially across occupations. For example, the average hours worked by managers and administrators and professionals tend to be substantially higher than those of workers who are employed in low-skill occupations. A reason for this is that the former tend to do much more unpaid overtime work, whereas the latter are employed in occupations where the incidence of part-time employment is much higher and, as a result, they work fewer hours. Applying the occupational categories and the skill hierarchy of the ASCO second edition (ABS 1997) to both employed persons and hours worked, Wooden (2000) suggests that the demand for labour in Australia has favoured the most skilled occupations:

   This widening gap in the demand for high-skill jobs compared with low-skill jobs is emphasised even further when the analysis is undertaken in terms of hours worked rather than the number of persons employed ... while the number of low-skill jobs has continued to rise, especially low-skill sales and service jobs, there has been virtually no growth in the total volume of low-skill work (p.197). According to Wooden (2000), the growth in the share of occupations that are low-skilled is due to the rapid increase in part-time and casual employment over the last three decades. The merit of Wooden's approach is that it draws attention to the importance of hours worked as a way of accounting for upskilling in the demand for labour.

   This finding is confirmed in a submission of the Commonwealth of Australia (2002) which adopted Wooden's methodology. In its submission to the Safety Net Review on wages, the Commonwealth grouped 282 occupations into three categories of employment to analyse growth in full-time employment by hours worked from 1996 to 2001. These occupations were grouped into low, middle and high-paid categories. The Commonwealth submission found that growth in hours worked in high-paid occupations accounted for 65 per cent of the total, while...