From Golden Age to Golden Age: Australia’s
“Great Leap Forward”?
School of Economics, Queensland University of Technology.
Australian National University, Research School of Social Sciences
(RSSS). Address: RSSS Economics, Australian National University,
The twenty-five years after WW 2 witnessed strong labour market institutions and beneficial labour market outcomes - high wage growth and integration of low-skilled immigrants. Then came the macro shocks of the mid 1970s. Labour market outcomes deteriorated as full-time employment population ratios fell, particularly among males;
unemployment and welfare use increased; and real wages grew slowly. The golden age passed. In response, successive governments have increasingly begun to dismantle the institutional framework. We address this transition within a simple long run graphical framework to help us marshal facts and arguments and to discuss the likely impact of institutional reform.
JEL classification: J0, D6, E6, L5, O3
During the twenty-five years following World War 2, the Australian labour market produced extraordinary outcomes, probably surpassed only by Japan during this period. The labour market absorbed 3 million immigrants, many with poor English language skills; maintained full employment with unemployment levels often below one percent; and increased average real wages at around 2-3% a year, a rate similar to the US and those countries of Western Europe with similar levels of economic development. This exceptional performance took place within an institutional framework of strong unions, centralized wage fixing, a compressed
iwage structure and an extensive welfare system. Any analysis of this period would conclude
that from a labour market perspective this was a golden age; all parts of the labour market were growing along a steady growth path with advantages for all. Of course there were economic cycles, but even these seemed, in labour market terms, to be extremely benign. The most serious recession, that of 1961/62, saw a brief unemployment increase to just over 2 per cent. Then, labour market outcomes gradually began to deteriorate. The first noticeable signs began with an increase in nominal wage inflation during the later years of the 1960s, peaking at 18% in the mid 1970s. There were structural problems emerging in the manufacturing sector where increased competition from Japan, and the unwillingness in Australia to increase tariffs further, was eating into employment but the structural problem did not seem that serious at the time. The early analysis of increased unemployment in the mid 1970s focused on macro outcomes, i.e. the unemployment inflation trade-off and the increase in real wages in the middle of the decade. Consistent with the macro focus – led primarily by the Reserve Bank, Treasury and
international organizations such as the OECD – the policy responses were directed towards
monetary policy and macro wage setting rules to be followed by the centralized wage fixing authority. In terms of achieving the intermediate macro targets there was considerable success. Real wages increases ceased almost immediately after 1975 and real wages were more or less maintained at a constant level for twenty years. Correcting the rate of inflation took longer, but by the early 1990s this intermediate target had also been achieved.
Despite the success in achieving intermediate targets, the labour market outcomes such as real wage increases and full-time employment growth have been among the worst in Australia’s
long run history. For a long time many analysts were prepared to believe that the poor labour market outcomes were the cost of the macro policies of constraint and that once the intermediate inflation and real wage targets were achieved the labour market would return to the past growth trajectory and the previous golden age would be restored. This is now clearly not the case and there has been a rethink of the initial diagnosis of the source of the difficulties that emerged in the 1970-1990 period. In response to the failure of traditional macro policies to achieve good labour market outcomes, policy makers have begun to look more closely at the institutional framework within which the labour market operates and following from this the process of dismantling the previously successful institutions has begun. This leads naturally to a number of important questions.
Were the institutions associated with the golden age in any way to blame for the arrival of the end of the golden age? And, irrespective of the answer to this question, has the world changed in some way such that the successful institutions of one environment were unsuccessful in another? What are the main issues in a new institutional framework? What would be the new pattern of labour market outcomes? And finally, after twenty or so years has the labour market
adjustment period come to an end so that we can begin again along a new golden age growth path or is there still a need for considerable adjustment in labour market institutions? Over the following pages we explore some of these questions and erect a framework for thinking about all of them. Our paper is structured as follows. Part 1 provides a selection of basic labour market facts that cover the post 1970 period. These facts are reasonably well known but they provide the platform on which we build and they provide some evidence to support judgments as to the likely outcomes from previous reforms.
Part 2 and 3 presents a simple graphical framework to help focus on some of the issues that surround the likely outcomes from institutional reform. This graphical analysis, which is used to discuss various reforms, presents a stylised picture of the long run operation of the labour market. A long run focus is needed to take the analysis away from many of the short run issues
ii We end with a discussion of research issues to take upon which many analysts naturally focus.
up in the future and return to some of the questions raised above.
Part 1: the problem
We think there are three sets of key facts that matter for the long run analysis of the Australian labour market: what is happening to the average real wage and wage structure, what is happening to full-time employment, and what is happening to welfare reliance?
The solid line in Figure 1 presents average real compensation per worker. The data are taken
iiifrom the OECD and reveal a remarkable story. The golden age is clear. Real wages were
growing at about 2-3% per annum during the 1960s and early 1970s. Then, after an above average real wage increase during the mid 1970s, real wages stayed more or less constant for
the next fifteen to twenty years. The deviation from a linear projection of the golden age trend (the dotted line starting in 1970) is remarkable. By 1995 real wages are about 30% to 50% less than they might have been if the golden age had continued.
This long period of real wage stagnation was completely unanticipated by those of us involved in the macro and labour market debates of the mid 1970s. No one foresaw fifteen years of real wage constancy. Some of the large changes over the period are associated with shifts in factor shares between profits and wages. The major underlying shifts, however, were the 1975-1990 slow down in productivity growth and a widening dispersion of productivity growth across different segments of the labour market, both of which had their counterparts in other countries such as the US,. Over the last decade productivity growth, as in the US, has returned to growth rates similar to the golden age and the wage dispersion has begun to narrow slightly. There are important wage issues that we have had to put aside. These include wage outcomes for women. There is a similar pattern of full time wage changes as for men, except for the large additional 30 per cent real wage increase during the equal pay period 1969 to 1975. A discussion of some of the issues can be found in Gregory and Duncan (1981), Gregory and Daly (1991), Gregory (2002) and Gregory (2004).
Figure 1 Australia, Real Total Compensation Per Employee IndexAdult Male FT Weekly Earnings, 10th, 20th and 80th Decile, 1960-20052.380th2.2
1960196519701975198019851990199520002005Source: Economic Outlook, OECD, DX Database, March 2005.YearEmployee Earnings, Benefits and Trade Union Membership, ABS Cat No 6310.0, August 1976-2004.Note - All Adult Male FT Weekly Earnings indexed at 1.7 in 1976.
It is also apparent from Figure 1 that, from about the mid 1990s real wages have returned to a growth path similar to that of the golden age. Note however, that there is no catch-up for the fifteen years of stagnation. The loss of real wage growth during the transition period appears to be lost forever. Was there any policy that could have avoided this loss? And, if so, could a policy advisor or government, against the background of the history of the golden age, and the long run success of the prevailing labour market institutions, have carried the day? We are certain that the answer to the second question is no but we do not know the answer to the first. We now turn to the wages of different groups in the labour market. During the golden age real
ivwage growth was fairly even across various groups, although, in response to data difficulties
we do not present the evidence here. There is, however, readily available data on the weekly earnings distribution for full-time workers from 1976. We include in Figure 1 the wages for the
ththvbottom 10, 20 and top 80th percentile of men employed full-time. The series are indexed to
the 1976 average real compensation per person.
There are four points to be made from this wage dispersion data. First, all parts of the full-time earnings distribution experienced the break in the real wage trend beginning in the mid 1970s and all experienced near constancy for about fifteen years. Second, within the earnings distribution, all groups have experienced earnings increases since the early 1990s but no groups have been able to recoup the real wage losses of the fifteen year stagnation period. Third, the wage dispersion has widened and for a considerable part of the period the full-time earnings of those in the lowest part of the earnings distribution fell. Fourth, full-time male workers in the bottom ten per cent of the earnings distribution receive much the same level of earnings that prevailed in 1975. For one fifth of the full-time male work force therefore there has been no real wage growth over the past thirty years. How does this pattern of earnings relate to employment outcomes?
Figure 2 Male Full-time Employment to Population Ratio, 15-641968-2004
YearsSource: The Labour Force, ABS Cat No 6203.0.
There are two sets of employment facts that are particularly important and can be used to illustrate the large changes that have occurred. First, Figure 2 presents the male full-time employment-population ratio for those aged between 15 and 64 years. The end of the golden age is clear. Since 1970 there has been a twenty-three per cent fall of male full-time employment (from 87 per cent to 67 per cent), although since the early 1990s the rate of
vi decline has slowed markedly. There has been no claw back of lost employment.If we put Figure 1 and Figure 2 together a clear picture emerges. The transition period for the male full-time employment ratio seems to have come to an end during the mid 1990s. At the same time as the decline in the male full-time employment population ratio stops real wage growth begins again. A number of largely unanswered questions are obvious. Since the male full-time employment population ratio is near its lowest level ever, why have real wages begun to increase? Why have we all the appearance of a new growth path in terms of wage increases and employment stability, but with very much lower employment levels? In many ways the responses to these questions are driving the program for institutional reform.
Males, 25-59 by Age Group and Skill Level
1981 and 2001 Australia Census
1981 Census 2001 Census
Proportion Employed Full-Time
Proportion Employed Full-Time
Definitions -Skilled is defined where education level is bachelor degree or higher, Unskilled defined as having no qualfications.Full-time is working more than 35 hours per week.Source: Census of Population and Housing, Household 1% Sample File, Australia, 1981 and 2001.
Second, who are these men who are no longer employed full-time? Table 1 lists the employment-population ratio for men by education category. The data are taken from the 1981 and 2001 Census which are years for which we have readily accessible unit record data. Full-time employment has fallen for all men but it has been primarily concentrated on the unskilled. At the time of the 2001 census the proportion of skilled men employed full-time is 77 per cent, a fall of 8 per cent over the period. By 2001 the employment-population ratio for unskilled men is a low 59 per cent, a fall of 21 per cent. Over the longer period since 1970 the falls are almost twice as large.
Has the loss of unskilled full-time jobs continued since 2001 now that the aggregate male full-time employment population ratio appears to have stabilized? The basic data are listed in Table 2 where we use occupational data to classify full-time male workers into skilled and less