Beyond the unemployment rate: three new measures of labour under utilisation for South Australia.

The official Australian unemployment rate is presently about five per cent and set to fall further. In South Australia the unemployment rate is somewhat higher, a little over six per cent. Nevertheless, it is the lowest that it has been in a generation. However, this paper re-evaluates the experiences of the South Australian labour market over the course of the last business cycle, using three new measures of labour under-utilisation. These new labour market indicators are calculated using data obtained from the monthly Labour Force Survey. The paper argues that the official unemployment rate greatly under-estimates the level of labour under-utilisation in South Australia. The real level of labour under-utilisation in South Australia has fallen little since the trough of the last recession was attained in 1993. This under-estimation of the extent of labour under-utilisation in South Australia has serious implications for industry and employment policy in South Australia.

Introduction

Australia has again recorded an annualised economic growth rate of about four per cent and so we are back at the top of the OECD league ladder. Nevertheless, the boom seems to be passing by an increasing number of Australians. The problem seems to be that the economy is not creating sufficient decent jobs to distribute adequately the benefits of continued strong economic growth (Stillwell, 2000). It is a problem that Australia shares with other, similar, developed countries such as Canada (Jackson et al., 2000; Burke and Shields, 1999). Moreover, the economic and labour force restructuring that accompanied the recovery from the recession of the early1990s has exacerbated two labour force pathologies: first, an increase in the level of hidden unemployment due to an increase in the structural unemployment that accompanied the economic restructuring of the 1990s (Watson, 2000); second, an increase in visible under-employment due to casual and part-time jobs becoming the driver of employment growth in Australia (Bell, 2000). The emergence of these two pathologies led to an increase in the level of labour under-utilisation in Australia.

The official measure of the extent of labour under-utilisation in Australia is the unemployment rate (Australian Bureau of Statistics, 2002). The method by which the unemployment rate is calculated means, however, that it does not include hidden unemployment, as the hidden unemployed are not actively seeking work and hence are precluded from participating in the Labour Force Survey. Nor is visible under-employment accurately measured, as the Labour Force Survey does not ask part-time workers who would like to work more hours how many extra hours they would they like to work. Consequently, the official picture under-estimates labour under-utilisation.

The aim of this paper is to estimate relative changes to the level of labour under-utilisation in South Australia since 1989, that is over the course of the last business cycle. For much of the following analysis 1991 becomes the comparator year and an arbitrarily chosen benchmark. This analysis will be undertaken by using three new measures of labour under-utilisation that have been developed at the Centre for Labour Research. These indicators should be seen as measures of change in the level of labour under-utilisation in South Australia, not of the absolute level. The following section is a brief re-examination of the South Australian labour force during the 1990s using the real unemployment rate, which adds an estimate of hidden unemployment to the unemployment rate. The third section is a re-examination of the South Australian labour force using the hours unemployment rate, which adds an estimate of visible under-employment to the unemployment rate. The fourth section is a re-examination of the South Australian labour force using the comprehensive unemployment rate, which adds estimates of both hidden unemployment and visible under employment to the unemployment rate. The final section concludes that there has been no improvement in the level of labour under-utilisation in South Australia since the trough of the recession in 1993. This finding has serious implications for industry and employment policy in South Australia.

Hidden Unemployment: the Real Unemployment Rate

During the last recession, a substantial amount of structural unemployment was created in Australia, in addition to the usual cyclical unemployment (Watson, 2000; Watson and Callus, 1999; Bell, 2000; Mitchell, 2000). Subsequently, many workers who became unemployed during the recession were unable to be re-employed in their old jobs, as they had been lost to the economy. Over time, many of these people have simply given up any hope of ever finding appropriate employment and have slipped into the ranks of the hidden unemployed. The Australian Bureau of Statistics collects some information about the extent of hidden unemployment during its monthly Labour Force Survey. The method by which this survey is administered means, however, that it significantly under-estimates the extent of hidden unemployment in Australia, as information about labour force participation is not gathered from people who are not actively seeking work during the reference period. Consequently, researchers at the Centre for Labour Research have developed a labour force indicator that adds an estimate of hidden unemployment to the number of people who are officially estimated to be unemployed so as to obtain the real unemployment rate for South Australia. The real unemployment rate is essentially a measure of those people who might reasonably be expected to work in a fully employed economy (Beatty and Fothergill, 1998; 138). Hence, it captures the increased labour under-utilisation that results from increases in hidden unemployment.

Beatty and Fothergill (1998; 116) argue that there are essentially three ways that regional labour markets adjust to the loss of a large number of jobs due to economic restructuring: first, the increase in the demand for labour that accompanies the creation of new jobs in regions that are experiencing employment decline; second, the reduction in the supply of labour as people move to regions where job growth is strong; third, the disappearance of people into the ranks of those who are neither employed nor unemployed, becoming economically inactive and joining the hidden unemployed. How can we estimate the number of hidden unemployed? Changes in the labour force participation rate provide the basis of a method (Appendix 1).

A feature of the South Australian labour force since 1989 is the decline in the labour force participation rate. The fall in the total participation rate is driven by a downward trend in the male participation rate, which has not been fully offset by the complementary increase in the female participation rate.

The decline in the male participation rate could be due to the more rapid ageing of the South Australian population. The Labour Economics Office (2001) argues, however, that ageing has little impact on participation rates. Both the fall in the South Australian male participation rate and the slower rate of increase of the South Australian female labour force (Table 1) are evidence for the existence of hidden unemployment in South Australia.

Table 1 shows that at the peak of the last business cycle there was very little difference between the male labour force participation rates for South Australia and Australia as a whole. During the 1990s, however, the difference became quite large. Figure 1 shows clearly the male labour force participation rates peaking in Australia and South Australia in 1990 and 1991, respectively, and then falling continuously throughout the 1990s. Nationally, the male labour force participation rate peaked at 75.4 per cent in 1990 and fell to 72.3 per cent by 2003. In South Australia, the male labour force participation rate peaked at 75.0 per cent in 1991 and fell to 69.3 per cent in 2003.

[FIGURE 1 OMITTED]

It is argued here that the two maxima that occur in the male labour force participation rates at the end of the 1980s represent the male participation labour force rates that could be expected to occur if the national and South Australian economies were experiencing full employment. Therefore, the decline in the male participation rate, from its peak in 1991, is an estimate of the extent of male hidden unemployment in South Australia. Hidden unemployment is then added to the number of people who are officially estimated as unemployed to obtain the male real unemployment rate.

In contrast to the male labour force participation rate, the female labour force participation rate has been rising constantly since the late-1980s (Table 1). Figure 2 shows that in 1989 there was very little difference between the national and the South Australian female labour force participation rates. But things changed during the 1990s. Nationally, the female participation rate rose from 50.6 per cent in 1989 to 56.6 per cent in 2003. In 1989 the South Australian female labour force participation rate was 50.4 per cent, but the rate rose much more slowly to reach only 53.9 per cent in 2003. The resultant gap between the national and South Australian female participation rates is also considered to be an indicator of the growth of hidden unemployment in South Australia. Hence, the benchmark for the calculation of the extent of female hidden unemployment in South Australia is the current national female labour force participation rate, which is then added to the number of people who are officially estimated as unemployed to obtain the female real unemployment rate.

[FIGURE 2 OMITTED]

In Table 2 and Figure 3, the real unemployment rate for South Australia is compared with the official unemployment rate. Prior to the onset of the last recession in South Australia, there was some hidden unemployment, as demonstrated by the vertical distance between the two lines in Figure 3. As the South Australian economy peaked in 1991, it prompted a number of people who were not in the labour force to begin actively to seek work. This optimistic revision of the expectations of gaining appropriate employment increased the participation rate and so reduced the level of hidden unemployment. Once the recession took hold, however, many people who were marginally attached to the labour force pessimistically reconsidered their chances of gaining appropriate employment, leaving the labour force. Hence the participation rate fell. The level of hidden unemployment jumped by two percentage points early in the contractionary phase of the recession and remained at that level throughout the recession. Then, paradoxically, the participation rate continued to fall, even though the unemployment rate fell; and the level of hidden unemployment rose again during the expansionary phase of the business cycle to reach six per cent by 2003.

[FIGURE 3 OMITTED]

The high levels of hidden unemployment that are shown in Table 2 mean that the real unemployment rate in South Australia did not change much during the 1990s (Figure 3), despite the fall in the unemployment rate in the latter half of the 1990s. Furthermore, other than during the last couple of years, the real unemployment rate for South Australia has shown no tendency to fall despite strong falls in the unemployment rate. That is, this labour market indicator shows that there has been no change in the level of labour under-utilisation in South Australia since the trough of the recession in 1993.

Visible Under-employment: the Hours Unemployment Rate

The economic restructuring that led to the emergence of structural and hidden unemployment as important labour force problems during the 1990s exacerbated another set of problems. As the manufacturing sector continued its relative decline, and conversely the service sector increased in importance, employers sought to increase the flexibility of the their workforces in order to improve their competitiveness. Increased workplace flexibility has been a feature of the Australian labour market for at least three decades. In recent years, however, the nature of workplace flexibility has changed. In the past, flexibility has tended to mean changes in the nature of work and employment relations that benefited employees. Examples of this form of flexibility include flexi-time, job-sharing, maternity leave and the provision of work-based childcare. During the 1990s, flexibility appears to have been quietly re-defined, and turned on its head, to mean changes to the nature of work that benefit employers at the expense of employees (Adam, 2001). This quest for a more responsive, compliant and flexible workforce has led to a significant increase in non-standard forms of employment in Australia.

Indeed, non-standard forms of employment, especially casual and part-time jobs, have been the drivers of employment growth in South Australia for over a decade. This is causing significant amounts of under-employment as the aspirations of increasing numbers of workers for full-time work are being frustrated by the actions of employers who are creating part-time, rather than full-time positions (McKay, 1998; Weller and Webber, 2001, Borland, Gregory and Sheehan, 2001). The Australian Bureau of Statistics does collect some information about part-time employees who wish to work more part-time hours. This information is then used by the Bureau to estimate the extent of visible under-employment. The official labour force statistics, however, significantly under-estimate the extent of visible under employment in Australia (Mitchell and Carlson, 2000). The hours unemployment rate developed at the Centre for Labour Research is based on the argument that most part-time workers who are seeking extra hours are actually expressing their preference for full-time work and would therefore like to work the number of hours worked on average in Australia. An estimate of visible under employment is added to the number of people who are officially estimated as unemployed by the Australian Bureau of Statistics to obtain the hours unemployment rate. The estimates for the total amount of under-utilised labour and the total size of the labour force are then converted to hours to obtain an hours rate of unemployment, rather than a person rate of unemployment (Appendix 2).

Table 3 and Figure 4 show movements in the hours unemployment rate for South Australia. This labour market indicator peaked in 1993 and fell during the 1990s in concert with falls in the unemployment rate. But it has not fallen as quickly as the unemployment rate and has not yet returned to the level that prevailed prior to the recession. This reflects the fact that the growth in female part-time employment has been the driver of job growth in South Australia during the 1990s. Hence, it cannot be concluded that the South Australian labour market has fully recovered from the recession.

[FIGURE 4 OMITTED]

The Comprehensive Unemployment Rate

The second section of this paper argued that the level of labour under-utilisation in South Australia was nearly double the official rate due to the creation of substantial amounts of structural unemployment during the 1990s. The preceding section argued that the level of labour under-utilisation in South Australia was double the official rate due to the creation of large amounts of part-time work during the 1990s. The real unemployment rate and the hours unemployment rate can be combined to calculate the comprehensive unemployment rate, which adds estimates of both hidden unemployment and visible under-employment to the number of people who are officially estimated as unemployed to obtain the comprehensive unemployment rate. This section will assess the performance of the South Australian labour force since the late-1980s using the comprehensive unemployment rate, which is also an hours unemployment rate (Appendix 3).

The data presented in Table 4 and Figure 5 tell a story of the experience of the South Australian labour force since the late-1980s that differs markedly from the story told by the unemployment rate. Like all of the labour market indicators analysed in this paper, the comprehensive unemployment rate increased markedly during the first two years of the 1990s to peak in 1993. Like the real unemployment rate, but unlike the hours unemployment rate, the comprehensive unemployment rate did not fall during the remainder of the 1990s. Rather, it fluctuated around the level attained during the trough of the recession. On the basis of this indicator, there is no evidence to suggest that the South Australian labour market has improved at all since the trough of the last recession.

[FIGURE 5 OMITTED]

Conclusion

The official labour force statistics show that the South Australian labour market has more than fully recovered from the recession of the early-1990s. The official statistics, however, hide two serious labour market problems, hidden unemployment and visible under-employment. The two key features of the official labour force statistics for South Australia are the trends in the labour force participation rates and the rate of part-time job growth. The growth in structural unemployment and part-time employment that occurred during the 1990s has led in South Australia to the creation of significant amounts of both hidden unemployment and visible under-employment, which are not captured by the unemployment rate. This paper has sought to determine the effects of this economic restructuring and labour force restructuring by developing and applying three new measures of labour under-utilisation. The real unemployment rate is a person rate of unemployment that adds an estimate of hidden unemployment to the unemployment rate. The hours unemployment rate is an hours rate of unemployment that adds an estimate of visible under employment to the unemployment rate. The comprehensive unemployment rate is also an hours rate of unemployment that adds estimates of both hidden unemployment and visible under employment to the unemployment rate. These measures of labour under-utilisation suggest that the actual level of labour under-utilisation in South Australia is between two and three times the unemployment rate.

The official picture of the last business cycle in South Australian is well known to most Australian labour market analysts. The unemployment rate in South Australia jumped from 6.7 per cent in 1990 to 11.2 per cent in 1993. Thereafter, during the expansionary phase of the business cycle, the unemployment rate trended downwards to reach 6.0 per cent in 2003 as people who became cyclically unemployed during the recession found work. However, the alternative labour force indicators that are discussed here point to a very different picture of the labour force experience of South Australia during the 1990s. Three key points emerge from the analysis based on these indicators.

First, the real unemployment rate jumped from 8.2 per cent in 1989 to 13.7 per cent in 1994, in line with the increase in cyclical unemployment, which drives changes in the unemployment rate. However, the real unemployment rate did not follow the unemployment rate downwards during the 1990s. Rather, it plateaued during the 1990s and only began to fall in the early-2000s. This implies that the cyclical unemployment that was created during the contractionary phase of the recession was not reversed during the expansionary phase.

Second, the hours unemployment rate also jumped during the early 1990s from 10.2 per cent in 1990 to 17.4 per cent in 1993, in concert with the increase in the unemployment rate. Unlike the real unemployment rate, however, the hours unemployment rate did fall during the 1990s, to 13.6 per cent in 2003, also in concert with the fall in the unemployment rate. These trends demonstrate that part-time work, which was the driver of employment growth in South Australia during the 1990s, caused the unemployment rate to fall, but also contributed to the increase in visible underemployment in the State.

Finally, the comprehensive unemployment rate also jumped during the contractionary phase of the last recession, from 11.2 per cent in 1990 to 17.9 per cent in 1993. But unlike the other labour market indicators that are discussed in this paper, the comprehensive unemployment rate has shown no tendency to fall since 1993, despite falls in the unemployment rate. Hence, this measure suggests that little under a fifth of South Australian labour resources presently remain unemployed, at a time when the official labour force measures indicate that only six per cent of the State's labour resources are unemployed.

The level of labour under-utilisation revealed by these three alternative indicators have three important implications for industry and employment policy in has Australia. First, and most obviously, it demonstrates that the scale of the unemployment problem is much greater than the official story may suggest. Indeed, the unemployment rate may under estimate the extent of labour under-utilisation in South Australia by as much as fifty per cent. Second, the falls in the unemployment rate have been driven largely by increased part-time employment, which in turn has contributed to the increase in visible under-employment. Third, structurally unemployed people need to be assisted back into the labour market. This will require a much more sophisticated and supportive approach than the simplistic 'workfare' policies that are in vogue in other countries, especially the USA.

An important caveat needs to be raised at this stage. Most economic models that are based on secondary data need to be based on assumptions in order to overcome the shortcomings of data that are collected by other people for other purposes. The three models developed and applied here are no exception. Hence, changes in the assumptions may lead to changes in the values of labour-under utilisation that are obtained from the models. The key conclusion from this paper is, however, that the level of labour under-utilisation in South Australia has not improved significantly since the trough of the recession in 1993. This in turn has implications for employment policy in South Australia. Policy needs to focus on the generation of full-time employment, including through the conversion of part-time jobs into fulltime jobs, in order to reduce the extent of visible under-employment.

APPENDIX 1

The Real Unemployment Rate

The real unemployment rate is a person rate of unemployment that is intended to be a count of the number of people who might reasonably be expected to work in a fully employed economy, regardless of whether or not they are active job seekers or claimants. This measure adds to the official unemployment rate estimates of the number of hidden unemployed males and females, which are determined from changes to the labour force participation rate since the early 1990s. The estimates for male and female hidden unemployment are calculated separately due to the differences in the experiences of male and female labour force participation since the end of the 1980s boom.

Males

The estimate for the level of male hidden unemployment is based on the assumption that the labour force participation rate that prevailed in 1990-1991 represents the number of men who might reasonably be expected to work in a fully employed Australian and South Australian economy. Hence the reduction in the labour force participation rate reflects the number of males who have given up looking for work and have slipped into the ranks of the hidden unemployed. The steps in the estimation procedure are as follows.

1. The labour force participation rates that existed at the peak of the previous business cycle (75.7 per cent for Australia in June 1990 and 75.0 per cent for South Australia in February 1991) are used as the benchmarks for this calculation. These figures are multiplied by the male civilian population age 15 and over to obtain the current male labour force if the male labour force participation rates of the early 1990s still prevailed. This is called the adjusted male labour force.

2.. The adjusted male labour force (step 1) is then used as the denominator for the male real unemployment rate.

3.. The difference between the adjusted male labour force obtained in step 1 and the actual male labour force obtained from the trend labour force statistics provides the estimate of male hidden unemployment.

4.. The estimate of hidden male unemployment (step 3) is then added to the official number of unemployed males, derived from the trend labour force statistics, to obtain the numerator of the real unemployment rate equation.

5.. The numerator obtained in step 4 is divided by the denominator obtained in step 2 and multiplied by 100 to obtain the male real unemployment rate.

Females

The estimate for the level of female hidden unemployment is based on the assumption that the labour force participation rate that currently prevails nationally represents the proportion of women who might reasonably be expected to work in a fully employed South Australian economy. Hence the slower rate of increase in the South Australian female labour force participation rate reflects the number of females who either have given up looking for work or have not sought work and so have slipped into the ranks of the hidden unemployed. The estimation steps are:

1. The current national female labour force participation rate is used as the benchmark for this calculation. This figure is multiplied by the female civilian population aged 15 and over to obtain an estimate of the size of the current South Australian female labour force if the current national female labour force participation rates prevailed in South Australia. This is called the adjusted female labour force.

2. The adjusted female labour force (step 1) is then used as the denominator for the female real unemployment rate equation.

3. The difference between the adjusted female labour force obtained in step 1 and the actual female labour force obtained from the trend labour force statistics provides the estimate of female hidden unemployment in South Australia.

4. The estimate of hidden female unemployment (step 3) is then added to the official number of unemployed females, derived from the trend labour force statistics, to obtain the numerator of the real unemployment rate equation.

5. The numerator obtained in step 4 is divided by the denominator obtained in step 2 and multiplied by 100 to obtain the female real unemployment rate. This method implicitly assumes that there is no hidden female unemployment nationally. This is not seen as a problem by the researchers at the Centre for Labour Research, as their interest is the comparison between the national and that South Australian labour forces.

APPENDIX 2

The Hours Unemployment Rate

The hours unemployment rate is an hours based unemployment rate that analyses the extent of labour under-utilisation from the perspective of the number of hours that the labour force is prepared to work, rather than the number of people in the labour force. This labour force indicator adds an estimate of visible under-employment to the official trend unemployment statistics. It is based on the assumption that all part-time workers who are looking for extra hours are implicitly expressing a preference for full-time work and combines an estimate of the number of extra hours that part-time workers who are looking for extra work would like to work (given the assumption that such workers would like to be working full-time rather than part-time) and an estimate of the number of hours that could be worked by the people who are counted as unemployed by the official unemployment statistics. The calculations for males and females are the same:

1. The number of males in the labour force, derived from the labour force statistics, is multiplied by the average number of hours worked by males to obtain the denominator of the equation.

2. The number of unemployed males is multiplied by the average number of hours worked by males to obtain an estimate of the number of labour hours unemployed.

3. The extent of visible under employment is obtained by summing the number of part-time male workers who would like to work more hours and then multiplying this figure by the number of average hours worked by males.

4. The figures obtained in steps 2 and 3 are summed to obtained the numerator of the equation.

5. The numerator obtained in step 4 is divided by the denominator obtained in step 1 and multiplied by 100 to obtain the hours unemployment rate.

APPENDIX 3

The Comprehensive Unemployment Rate

The comprehensive unemployment rate adds to the trend unemployment rates estimates of the level of hidden unemployment and visible underemployment, expressed as an hours rate of unemployment. This measure is therefore based on the two sets of assumptions discussed in the previous two appendixes.

Males

1. The numerator for this measures is obtained by adding three separate components:

* an hours rate of unemployment obtained by multiplying the male trend unemployment rate by the average number of hours worked by males;

* an hours rate of male hidden unemployment obtained by multiplying the number of hidden unemployed males obtained in Appendix 1 by the average number of hours worked by males; and

* the hours based estimate for male labour under-utilisation obtained in Appendix 3.

2. The denominator for this measure is calculated by multiplying the denominator obtained for the male real unemployment rate in Appendix 1 by the average number of hours worked for males.

3. The male comprehensive unemployment rate is then obtained by dividing the numerator obtained in Step 1 with the denominator obtained in step 2 and multiplying by 100 to obtain a percentage.

Females

1. The numerator for this measures is obtained by adding three separate components:

* an hours rate of unemployment obtained by multiplying the official female unemployment rate by the average number of hours worked by females;

* an hours rate of female hidden unemployment obtained by multiplying the number of hidden unemployed females obtained in Appendix 1 by the average number of hours worked by females; and

* the hours based estimate for female labour under-utilisation obtained in Appendix 3.

2. The denominator for this measure is obtained by multiplying the denominator obtained for the female real unemployment rate in Appendix 1 by the average number of hours worked for females.

3. The comprehensive unemployment rate is obtained by dividing the numerator obtained in step 1 with the denominator obtained in step 2 and multiplying by 100 to obtain a percentage.

Persons

The male and female numerators and denominators are added and the total comprehensive unemployment rate obtained by calculating the former as a percentage of the latter.

Table 1: Labour Force Participation Rate (%), Australia and South
Australia, 1989-2003

              Australia              South Australia

Year   Males   Females   Total   Males   Females   Total

1989   75.3     50.6     62.7    74.5     50.2     62.1
1990   75.4     51.9     63.5    74.2     50.4     62.1
1991   75.2     52.2     63.5    75.0     51.7     63.1
1992   74.3     51.9     63.0    73.1     50.5     61.6
1993   73.6     51.5     62.4    72.7     51.1     61.7
1994   73.6     52.4     62.8    71.7     51.3     61.3
1995   73.9     53.2     63.4    71.2     51.9     61.4
1996   73.8     53.8     63.6    71.4     52.1     61.5
1997   73.3     54.0     63.5    70.7     52.9     61.6
1998   73.0     53.6     63.1    70.1     51.0     60.3
1999   72.7     53.7     63.0    70.0     51.1     60.4
2000   72.5     54.7     63.5    70.6     52.2     61.2
2001   72.3     55.1     63.6    68.7     51.7     60.0
2002   72.4     55.3     63.8    69.2     52.7     60.8
2003   72.3     56.6     64.4    69.3     53.9     61.5

Source: ABS unpublished trend data, figures for February of each year

Table 2: Official and Real Unemployment (%), South Australia, 1989-2003

Year   Official Unemployment        Real Unemployment rate

       Males   Females   Total      Males   Females   Total

1989   7.7     7.2       7.5        8.3     7.9       8.2
1990   6.7     6.8       6.7        7.7     9.5       8.4
1991   9.2     8.3       8.8        9.2     9.2       9.2
1992   12.1    9.4       11.0       14.4    11.8      13.3
1993   12.3    9.7       11.2       15.0    10.4      13.1
1994   11.4    9.5       10.6       15.3    11.4      13.7
1995   10.5    8.6       9.7        15.1    10.9      13.2
1996   10.0    7.9       9.1        14.3    10.8      12.9
1997   9.9     8.3       9.2        15.1    10.1      13.0
1998   10.0    9.1       9.6        15.9    13.5      14.9
1999   9.4     7.9       8.7        15.4    12.3      14.0
2000   8.3     7.6       8.0        13.6    11.9      12.9
2001   8.1     5.9       7.1        15.9    12.1      14.2
2002   7.5     6.1       6.9        14.7    10.6      12.7
2003   6.6     5.3       6.0        13.7    9.8       11.9

Source: ABS unpublished trend data, figures for February of each year.

Table 3: South Australia, Hours Unemployment Rate (%), 1989-2003

            Males             Females           Persons

       Hours   Official   Hours   Official   Hours   Official
       rate    rate       rate    rate       rate    rate

1989   9.9     7.7        12.5    7.2        11.2    7.5
1990   8.7     6.7        11.7    6.8        10.2    6.7
1991   12.1    9.2        13.5    8.3        12.7    8.8
1992   16.0    12.1       16.6    9.4        16.6    11.0
1993   17.4    12.3       16.6    9.7        17.4    11.2
1994   15.1    11.4       15.9    9.5        15.9    10.6
1995   15.6    10.5       15.7    8.6        16.0    9.7
1996   14.0    10.0       15.3    7.9        15.0    9.1
1997   15.1    9.9        16.3    8.3        16.1    9.2
1998   15.0    10.0       16.0    9.1        15.8    9.6
1999   14.6    9.4        15.7    7.9        15.4    8.7
2000   12.4    8.3        13.2    7.6        13.1    8.0
2001   12.2    8.1        12.4    5.9        12.4    7.1
2002   13.8    7.5        14.4    6.1        14.4    6.9
2003   11.5    6.6        13.2    5.3        12.6    6.0

Source: ABS unpublished trend data, figures for February of each year.

Table 4: South Australia; Comprehensive Unemployment Rate, 1989-2003,
per cent.

               Males                Females              Persons

         Compre               Compre              Compreh
        hensive   Official   hensive   Official    ensive   Official

1989       10.5        7.7      13.2        7.2      11.4        7.5
1990        9.7        6.7      14.3        6.8      11.2        6.7
1991       12.1        9.2      14.3        8.3      12.8        8.8
1992       18.1       12.1      18.9        9.4      18.3       11.0
1993       19.9       12.3      17.3        9.7      19.0       11.2
1994       18.8       11.4      17.7        9.5      18.4       10.6
1995       19.9       10.5      17.7        8.6      19.2        9.7
1996       18.1       10.0      18.0        7.9      18.1        9.1
1997       20.0        9.9      18.0        8.3      19.3        9.2
1998       20.6       10.1      20.1        9.1      20.4        9.6
1999       20.3        9.4      19.5        7.9      20.0        8.7
2000       17.6        8.3      17.2        7.6      17.4        8.0
2001       19.7        8.1      18.1        5.9      19.2        7.1
2002       20.4        7.5      18.4        6.1      19.7        6.9
2003       18.2        6.6      17.4        5.3      17.9        6.0

Source: ABS unpublished trend data, figures for February of each year.

References

Adam, B, (2001), When Time is Money: Contested Rationalities of Time and Challenges to the Theory and Practice of Work, Working Paper Series Paper 16, School of Social Sciences, Cardiff University.

Australian Bureau of Statistics (2001), Labour Statistics: Concepts, Sources and Methods, Australian Bureau of Statistics, Canberra, cat. no. 6102.0.

Australian Bureau of Statistics (2002), Measuring Australia's Progress, Australian Bureau of Statistics, Canberra, cat. no. 1370.0.

Beatty, C. and Fothergill, S. (1998), 'Registered and Hidden Unemployment in the UK Coalfields', in Lawless, P., Martin, R., and Hardy, S. (eds), Unemployment and Social Exclusion: Landscapes of Labour Exclusion, Jessica Kingsley, London, pp. 116-140.

Bell, S. (2002), 'The Contours and Dynamics of Unemployment', in Saunders, P., and Taylor, R. (eds), The Price of Prosperity: The Economic and Social Costs of Unemployment, UNSW Press, Sydney, pp. 13-43.

Borland, J., Gregory, R.G. and Sheehan, P. (2001), 'Inequality and Economic Change', in Borland, J., Gregory, R.G., and Sheehan, P. (eds), Work Rich, Work Poor: Inequality and Economic Change in Australia, Centre for Strategic Economic Studies, Melbourne, pp. 1-20.

Burke, M. and Shields, J. (1999), The Job-Poor Recovery: Social Cohesion and the Canadian Labour Market, A Research Report of the Ryerson Social Reporting Network, Ryerson Polytechnic University, Toronto.

Jackson, A., Robinson, D., Baldwin, B. and Wiggins, C. (2000), Falling Behind: The State of Working Canada, 2000, Canadian Centre for Policy Alternatives, Ottawa.

Labour Economics Office (2001), Labour Force Participation Rates--a Longer Term Perspective, Department of Employment and Workplace Relations South Australia, Adelaide.

Mata Greenwood, A. (1999), 'Gender Issues in Labour Statistics', International Labour Review, 138(3), pp. 273-286.

MacKay, R.R. (1998), 'Unemployment as Exclusion; Unemployment as Choice', in Lawless, P., Martin, R., and Hardy, S. (eds), Unemployment and Social Exclusion: Landscapes of Labour Exclusion, Jessica Kingsley, London, pp. 49-68.

Mitchell, W.F. (2000), 'The Causes of Unemployment', in Saunders, P., and Taylor, R. (eds), The Price of Prosperity: The Economic and Social Costs of Unemployment, UNSW Press, Sydney, pp. 49-87.

Mitchell, W.F. and Carlson, E. (2000), Beyond the Unemployment Rate: Labour Underutilisation and Underemployment in Australia and the USA, Working Paper No. 00-06, Centre of Full Employment and Equity, University of Newcastle.

Stilwell, F. (2000), Changing Track: A New Political Economic Direction for Australia, Pluto Press, Sydney.

Watson, I. (2000), 'Beyond the Unemployment Rate: Building a Set of Indices to Measure the Health of the Labour Force', Australian Bulletin of Labour 26, 175-190.

Watson, I. and Callus, R. (1999), 'What Does the Unemployment Rate Measure?', CEDA Political and Economic Newsletter, No. 24, March, pp. 4-7.

Weller, S. and Webber, M. (2001), 'Precarious Employment and Occupational Change', in Borland, J., Gregory, R.G., and Sheehan, P. (eds), Work Rich, Work Poor: Inequality and Economic Change in Australia, Centre for Strategic Economic Studies, Melbourne, pp. 160-195.

Steven Barrett, Centre for Labour Research, University of Adelaide