RDP 1978-02: Unemployment: An Econometric Dissection Appendix A: The Data

Definitions and sources of variables for labour demand, labour supply, unemployment, award wages and earnings are set out below, together with a description of the series used in figure 2.5 (labour costs and productivity). The remaining variables are defined in Jonson et al (1977).

1. Labour demand, labour supply, unemployment

This set of variables uses data obtained from the quarterly population surveys conducted by the Australian Bureau of Statistics. These surveys provide a consistent set of data classified by age and sex for employment, unemployment, labour force and hours worked.

The definitions of the variables as used in the model, are outlined below.

Demand for labour: In the model, this is expressed in terms of “full-time equivalents” and is measured as total hours worked (number of employed persons times average hours worked) divided by the average for the sample period of average hours worked by full-time persons. The series is in effect a weighted average of full-time and part-time workers. Because part-time persons are included with a weight of less than unity, our series is lower than the published series for the number of employed persons. Figure 2.1 plots both series.

Unemployment: This is expressed in terms of the total number of unemployed persons; this series therefore includes (with equal weight) those looking for part-time work and those looking for full-time work.

Supply of labour: This is defined as the sum of demand for labour and unemployment. Reflecting the correction made to the demand for labour series, our measure of labour supply is lower than the published labour force statistics.

2. Award wages and earnings

With the exception of award wages for adult males and adult females, quarterly series for award wages and earnings are not published for each of the major age/sex groups; consequently, the remaining series have to be constructed. The method of construction is outlined below; it makes use of the following supplementary data:

• the published series for average weekly earnings per male unit;
• the implied (by published male unit ratios) series for average weekly earnings per female unit;
• average weekly earnings series for all four age/sex categories obtained from the October and May surveys of earnings and hours conducted by the Australian Bureau of Statistics.

The method for deriving the various series is as follows:

Earnings of adult males: Figures for the December quarters are those from the October surveys of earnings. Figures for the June quarters (after 1974) are from the May surveys of earnings. For the remaining quarters, figures have been calculated using quarterly dollar increases in the average weekly earnings per male unit series scaled by the ratio of the annual dollar increase in the October survey figures to the annual dollar increase in the male unit series (December quarter on December quarter).

Earnings of junior males: as for adult males.

Earnings of adult females: as for adult males (except that the quarterly increases are based on the female unit earnings series).

Earnings of junior females: as for adult females.

Awards of junior males: this series is approximated by ordinary time earnings (the difference between total earnings and overtime earnings) of junior males. The series for overtime earnings is a proportion of the male unit overtime earnings series which is estimated by an equation based on average hourly earnings and average hours of overtime worked. The proportion is calculated from the annual overtime earnings figures published in the October surveys.

Awards of junior females: as for awards of junior males.

3. Labour costs and productivity

In keeping with the rest of the model, the series shown in figure 2.5 relate to the economy as a whole; i.e. they do not distinguish between the private and government sectors. In theory, one could argue that employment and pricing decisions are made in the market sector (approximated by the private sector) and therefore the relevant comparison of labour costs and productivity is one which relates to that sector. In view of this, we have also calculated labour costs and productivity for the private sector. These calculations gave similar results to those shown in figure 2.5.

The definitions of the variables used in figure 2.5 are:

Real labour costs: a weighted average of earnings of the four age/sex groups (expressed in terms of full-time equivalents), adjusted for payroll taxes and deflated by the GDP price deflator corrected for consumer taxes;

Actual labour productivity: .7 GDP per full-time equivalent employed;

Trend labour productivity: the annual average rate of increase in actual productivity over the period 1960 to 1969.

In figure 2.5, the series are expressed in index form with a base of 1960 to 1969 equal to 100. In the model, the series enter in dollar amounts with any discrepancy between the average level of the two series being reflected in the constant.