RDP 8806: Employment, Output and Real Wages 5. Simulation Results

In this section, we use the estimates from equation 2 to estimate the effects that changes in real wages and output had in the downturns of employment in 1974/75 and 1982/83 and the recovery in employment since 1983. The questions we consider are:

  • what would have happened to employment if real wages and productivity had grown on trend (i.e., if real unit labour costs had been constant) but output had followed its actual path? and
  • what would have happened to employment if output had grown on trend but real unit labour costs had followed their actual path?

June 1974 to December 1975

Figure 2 plots the results of the simulation for the first period. The solid line on the Figure represents the actual path of employment; the dashed line represents the equation's estimate of what employment would have been if real unit labour costs had been constant; and the dotted line the estimate of employment had output grown on trend.


Consider the real wage simulation. If real unit labour costs had been unchanged in 1974/75, rather than rising sharply, then employment would have been much stronger during this period. Specifically, the estimates suggest that full-time employment in December 1975 would have been around 70,000, or 1–1/2 per cent, higher than it actually was. The estimated effect of the slowing in GDP growth was smaller. For the first part of the period, employment would have been stronger had GDP grown at trend. However, in the latter part of the period, when GDP growth exceeded trend, employment growth would have been a little weaker than actually observed.

These results suggest that the sharp rise in real unit labour costs was the major contributor to the downturn in employment in 1974/75.

December 1981 to March 1983

Figure 3 illustrates the results of the same exercise for the downturn of the early 1980s. The effect of the slowing in GDP growth in this period was relatively more important than in 1974/75. This was due to the fact that the contraction in output in 1982/83 was much sharper than in 1974/75; as well, the large rise in real unit labour costs was unwound more quickly in the latter period. On the basis of the simulations, employment would have been 126,000 higher in March 1983 if output had not slowed below its trend rate. If real unit labour costs had not risen during this period, then employment would have been 154,000 higher than was observed.


The Period Since 1983

The simulation exercises for this period suggest that much of the strength of employment can be attributed to the moderate wage outcomes. (Figure 4 provides details.) If real unit labour costs are held at a level that prevailed immediately before the wage pause, then employment in December 1987 would have been 377,000 lower than the actual level. On the other hand, employment would have only been 27,000 lower had GDP grown at trend (basically because growth in GDP was close to, or above, trend for most of that period).


These results suggest that the prolonged and large decline in real unit labour costs appears to have been the most important factor in the recent buoyancy of employment. The role of lower real wages in stimulating employment is highlighted in 1985/86. Between September 1985 and June 1986, GDP fell by 1–1/2 per cent. However, full-time employment grew by 2–3/4 per cent. This continued growth in employment in the face of a contraction in activity appears to have been due to the low and declining level of labour costs over the period.

The results of these simulations show that changes in both real wages and output have had important effects on employment in the 1970s and 1980s. The contraction in employment in 1974/75 could be labelled as largely “Classical”. A reduction in real unit labour costs in this period would have had a much larger effect on employment than a return to trend growth in GDP. The contraction in 1982/83 was a mix of both Keynesian and Classical features. Both a reduction in real unit labour costs and a return to trend GDP growth would have stimulated employment by around the same amount.