RDP 9215: The Evolution of Employment and Unemployment in Australia 2. Some Facts

Figure 1 shows the aggregate unemployment rate from 1950 to 1992^[6]. Three distinct regimes are apparent. The first was between 1950 and 1973, when unemployment cycled around an equilibrium of about two per cent of the workforce. The second was between 1974 and 1981, when the unemployment rate increased steadily to about 6.0 per cent. There were no apparent cycles during this time, suggesting this rise reflected an increase in the equilibrium rate of unemployment. Finally, there has been the period since 1982. Unemployment rose sharply during the recession of 1982–83, peaking at 10.0 per cent of the workforce. It then declined slowly over the remainder of the 1980s, falling to 6.0 per cent in 1989, and has since again increased rapidly, averaging 10.7 per cent in 1992.

However, the aggregate numbers mask considerable differences in the unemployment experiences of men and women, on the one hand, and of full-time and part-time workers, on the other. We show this in Figure 2, which plots the full-time and part-time unemployment rates for men and women from 1966 to 1992.^[7] Two observations can be made about full-time unemployment rates. First, the two full-time unemployment rates closely track the aggregate unemployment rate shown in Figure 1. Second, the full-time unemployment rate for women has, until recently, been considerably higher than for men. The recent downturn in the economy, however, has seen a narrowing of this difference.

The two bottom panels show the unemployment rates of those people seeking part time work. Here, the differences between men and women are quite marked. While both part-time unemployment rates have trended upwards, and show less cyclical variation than full-time rates,^[8] the slight upward trend in women's part-time unemployment rate is much less pronounced than for the other three categories.

Figure 3 puts these comparative unemployment rates into some perspective by showing the numbers of unemployed in each category. The greatest proportionate increase in the number of unemployed has occurred in the full-time male labour market. The 1992 level of unemployment is fourteen times the level in 1966, while the rate of increase in the number of unemployed females seeking either part-time or full-time employment has been roughly half of this. However, the fact that the part-time female unemployment rate was relatively stable over this period suggests that a substantial increase in both the demand and supply of part-time women's labour.

The unemployment data therefore imply that developments in the aggregate labour market have been dominated by the full-time male and part-time female markets. This supposition is confirmed in Figure 4, which plots the proportion of total unemployment accounted for by each category. Full-time male unemployment, as a proportion of the total has, since 1983, shown exactly the same pattern as the aggregate unemployment rate. Part-time female unemployment has shown the opposite pattern, falling as a proportion of aggregate unemployment during recessions and rising during periods of strong employment growth, reflecting the well-known pro-cyclical nature of women's labour supply. In trend terms, full-time male unemployment has been steadily increasing as a proportion of the total, while the fractions accounted for by full and part-time female unemployment have decreased. Part-time male unemployment has remained steady at about five per cent of total unemployment.

Further insight into female unemployment can be gained by separately considering the unemployment rates of married and unmarried women.^[9] This is done in Figure 5. The lower left quadrant shows that, since 1966, the unemployment rate for married women seeking part-time work has shown virtually no trend increase, and very little cyclical variation, implying that the part-time married women's participation rate is strongly pro cyclical, albeit around a rising trend.

The married women's full-time unemployment rate, in contrast, exhibits a slight upward trend, which clearly follows the aggregate unemployment cycle, although both the cyclical peaks and average unemployment are relatively low. On the other hand, both the full-time and part-time unemployment rates for unmarried women have shown large secular increases and a strong cyclical effect. In 1992, the unmarried women's full-time unemployment rate was a very high 16 per cent, and the part-time rate 11 per cent, compared with a rate of four per cent for married women looking for part-time work and eight per cent for those seeking full-time employment.

There are probably two reasons for these large differences. The first is that unmarried women are disproportionately young women, and so the unemployment rates for unmarried women reflect the very high rates of youth unemployment generally.^[10] The second is that unmarried women, by definition, do not have spouses who are (probably) working and so are less able to leave the labour force when jobs are scarce and re-enter it at a more favourable time in the business cycle.

In Figure 6 we plot the full-time and part-time employment of men and women. Most striking is women's part-time employment, which grew by 281 per cent over the period 1966–1992, compared with the anaemic 23 per cent growth in full-time male employment. The increase in growth in the former category, and slowdown in the latter, began in the mid 1970s.

The share of each category in total employment is shown in Figure 7. Full-time male employment as a share of the total has declined steadily, with the opposite true for part-time women. Full-time female employment has constantly been around 25 per cent of the total, while the share of part-time male employment has risen constantly, but only to about six per cent in 1992. The share of part-time employment overall increased from 10.0 per cent in 1966 to 23.5 per cent in 1992.

In Table 1 we compare the loss of employment in the two recessions, 1982–83 and 1990–92. Clearly, the bulk of the job losses in the two recessions has been borne by men working full-time. Part-time employment of both men and women has increased in the latest downturn, as it did in 1982–83. However, this possibly reflects some underemployment, with firms cutting the hours worked by some employees.^[11]

Figure 8 decomposes female employment. Here we can see that the bulk of the increase in women's employment, both full-time and part-time, went to married women. (However, full-time employment of married women actually fell slightly between 1974 and 1983, while it increased by 23 per cent for unmarried women.)

In Figures 9 and 10 we show participation rates – the proportion of the civilian population in each category aged 15 years and over that is in the labour force, either employed or unemployed. The decreasing participation of men, and increasing participation of women, is a well-documented fact (Chapman 1990). Figure 9 shows that the fall in male participation has been due to the decrease in full-time participation, as part-time participation has increased. These changes began in the mid 1970s, no doubt largely caused by the decline in the growth of full-time male employment. The increase in female participation has been largely due to an increase in part-time participation, which doubled over the 26 year period to 1992, while full-time participation remained at about 30 per cent. However, Figure 10 shows that this constancy was the result of offsetting factors: participation for full-time unmarried women fell secularly, probably due to increased retention rates in schools, while it increased for the three other categories.

The general picture that emerges from this overview is that the important changes to employment, participation and unemployment in Australia over the past 26 years or so have been driven by trends in the markets for full-time male and part-time female workers. We will elaborate this theme throughout the paper as we examine gross labour market flows, unemployment duration, and employment and unemployment by sector.

Footnotes

The 1992 data in this section are the simple average of seasonally adjusted data up to and including September 1992. Data for other years are the quarterly averages of unadjusted data up to 1978 and the monthly averages of unadjusted data from 1978 to 1991. [6]

That is, the unemployment rates for men and women seeking full-time work, and part-time work, respectively. [7]

The part-time unemployment rate for men was inexplicably high between 1972 and 1977. This probably reflects problems in the data. A new and comprehensive monthly labour force survey began in February 1978; the data from this survey are likely to be more reliable than earlier data. [8]

Despite the fact that, until recently, the participation rate for married women was less than for unmarried women (until the mid 1970s, much less), married women always made up a majority of the female labour force, over the period 1966–1992. This was because there were many more married than unmarried women of working age. (Foster and Stewart (1991), Table 4.4, p152). [9]

In 1991, 21.7 per cent of 15–19 year old women in the labour force were unemployed, while the unemployment rate for all women was 9.2 per cent. [10]

The number of people working part-time but who would prefer to work more hours increased signficantly in 1990 and 1991. [11]